IBM Laptop SG24 5131 00 User Manual

IBM Certification Study Guide  
AIX HACMP  
David Thiessen, Achim Rehor, Reinhard Zettler  
International Technical Support Organization  
http://www.redbooks.ibm.com  
SG24-5131-00  
Download from Www.Somanuals.com. All Manuals Search And Download.  
SG24-5131-00  
International Technical Support Organization  
IBM Certification Study Guide  
AIX HACMP  
May 1999  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Take Note!  
Before using this information and the product it supports, be sure to read the general information in  
First Edition (May 1999)  
This edition applies to HACMP for AIX and HACMP/Enhanced Scalability (HACMP/ES), Program  
Number 5765-D28, for use with the AIX Operating System Version 4.3.2 and later.  
Comments may be addressed to:  
IBM Corporation, International Technical Support Organization  
Dept. JN9B Building 003 Internal Zip 2834  
11400 Burnet Road  
Austin, Texas 78758-3493  
When you send information to IBM, you grant IBM a non-exclusive right to use or distribute the  
information in any way it believes appropriate without incurring any obligation to you.  
© Copyright International Business Machines Corporation 1999. All rights reserved.  
Note to U.S Government Users – Documentation related to restricted rights – Use, duplication or disclosure is  
subject to restrictions set forth in GSA ADP Schedule Contract with IBM Corp.  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Contents  
Figures. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .ix  
Tables. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xi  
Preface. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xiii  
Chapter 1. Certification Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1  
Chapter 2. Cluster Planning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7  
© Copyright IBM Corp. 1999  
iii  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Chapter 5. Cluster Customization . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117  
iv  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
5.4.5 Cross Mounted NFS File Systems and the Network Lock Manager.  
128  
Chapter 6. Cluster Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131  
Chapter 7. Cluster Troubleshooting . . . . . . . . . . . . . . . . . . . . . . . . . . 143  
v
Download from Www.Somanuals.com. All Manuals Search And Download.  
Chapter 9. Special RS/6000 SP Topics . . . . . . . . . . . . . . . . . . . . . . . . 183  
vi  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Appendix A. Special Notices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205  
Appendix B. Related Publications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209  
How to Get ITSO Redbooks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211  
List of Abbreviations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215  
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217  
ITSO Redbook Evaluation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221  
vii  
Download from Www.Somanuals.com. All Manuals Search And Download.  
viii  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Figures  
© Copyright IBM Corp. 1999  
ix  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
x
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Tables  
5. Number of Available Serial Ports in Each Model. . . . . . . . . . . . . . . . . . . 15  
© Copyright IBM Corp. 1999  
xi  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
xii  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Preface  
The AIX and RS/6000 Certifications offered through the Professional  
Certification Program from IBM are designed to validate the skills required of  
technical professionals who work in the powerful and often complex  
environments of AIX and RS/6000. A complete set of professional  
certifications is available. It includes:  
• IBM Certified AIX User  
• IBM Certified Specialist - RS/6000 Solution Sales  
• IBM Certified Specialist - AIX V4.3 System Administration  
• IBM Certified Specialist - AIX V4.3 System Support  
• IBM Certified Specialist - RS/6000 SP  
• IBM Certified Specialist - AIX HACMP  
• IBM Certified Specialist - Domino for RS/6000  
• IBM Certified Specialist - Web Server for RS/6000  
• IBM Certified Specialist - Business Intelligence for RS/6000  
• IBM Certified Advanced Technical Expert - RS/6000 AIX  
Each certification is developed by following a thorough and rigorous process  
to ensure the exam is applicable to the job role and is a meaningful and  
appropriate assessment of skill. Subject Matter Experts who successfully  
perform the job participate throughout the entire development process. These  
job incumbents bring a wealth of experience into the development process,  
thus making the exams much more meaningful than the typical test, which  
only captures classroom knowledge. These Subject Matter experts ensure  
the exams are relevant to the real world and that the test content is both  
useful and valid. The result is a certification of value that appropriately  
measures the skill required to perform the job role.  
This redbook is designed as a study guide for professionals wishing to  
prepare for the certification exam to achieve IBM Certified Specialist - AIX  
HACMP.  
The AIX HACMP certification validates the skills required to successfully  
plan, install, configure, and support an AIX HACMP cluster installation. The  
requirements for this include a working knowledge of the following:  
• Hardware options supported for use in a cluster, along with the  
considerations that affect the choices made  
xiii  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
• AIX parameters that are affected by an HACMP installation, and their  
correct settings  
• The cluster and resource configuration process, including how to choose  
the best resource configuration for a customer requirement  
• Customization of the standard HACMP facilities to satisfy special  
customer requirements  
• Diagnosis and troubleshooting knowledge and skills  
This redbook helps AIX professionals seeking a comprehensive and  
task-oriented guide for developing the knowledge and skills required for the  
certification. It is designed to provide a combination of theory and practical  
experience. It also provides sample questions that will help in the evaluation  
of personal progress and provide familiarity with the types of questions that  
will be encountered in the exam.  
This redbook will not replace the practical experience you should have, but,  
when combined with educational activities and experience, should prove to  
be a very useful preparation guide for the exam. Due to the practical nature of  
the certification content, this publication can also be used as a desk-side  
reference. So, whether you are planning to take the AIX HACMP certification  
exam, or just want to validate your HACMP skills, this book is for you.  
For additional information about certification and instructions on How to  
Register for an exam, call IBM at 1-800-426-8322 or visit our Web site at:  
http://www.ibm.com/certify  
The Team That Wrote This Redbook  
This redbook was produced by a team of specialists from around the world  
working at the International Technical Support Organization Austin Center.  
David Thiessen is an Advisory Software Engineer at the International  
Technical Support Organization, Austin Center. He writes extensively and  
teaches IBM classes worldwide on all areas of high availability and clustering.  
Before joining the ITSO six years ago, David worked in Vancouver, Canada  
as an AIX Systems Engineer.  
Achim Rehor is a Software Service Specialist in Mainz/Germany. He is Team  
Leader of the HACMP/SP Software Support Group in the European Central  
Region (Germany, Austria and Switzerland). Achim started working with AIX  
in 1990, just as AIX Version 3 and the RISC System/6000 were first being  
introduced. Since 1993, he has specialized in the 9076 RS/6000 Scalable  
xiv  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
POWERparallel Systems area, known as the SP1 at that time. In 1997 he  
began working on HACMP as the Service Groups for HACMP and RS/6000  
SP merged into one. He holds a diploma in Computer Science from the  
University of Frankfurt in Germany. This is his first redbook.  
Reinhard Zettler is an AIX Software Engineer in Munich, Germany. He has  
two years of experience working with AIX and HACMP. He has worked at IBM  
for two years. He holds a degree in Telecommunication Technology. This is  
his first redbook.  
Thanks to the following people for their invaluable contributions to this  
project:  
Marcus Brewer  
International Technical Support Organization, Austin Center  
Rebecca Gonzalez  
IBM AIX Certification Project Manager, Austin  
Milos Radosavljevic  
International Technical Support Organization, Austin Center  
Comments Welcome  
Your comments are important to us!  
We want our redbooks to be as helpful as possible. Please send us your  
comments about this or other redbooks in one of the following ways:  
• Fax the evaluation form found in “ITSO Redbook Evaluation” on page 221  
to the fax number shown on the form.  
• Use the electronic evaluation form found on the Redbooks Web sites:  
For Internet users  
For IBM Intranet users  
http://www.redbooks.ibm.com  
http://w3.itso.ibm.com  
• Send us a note at the following address:  
xv  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
xvi  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Chapter 1. Certification Overview  
This chapter provides an overview of the skill requirements for obtaining an  
IBM Certified Specialist - AIX HACMP certification. The following chapters  
are designed to provide a comprehensive review of specific topics that are  
essential for obtaining the certification.  
1.1 IBM Certified Specialist - AIX HACMP  
This certification demonstrates a proficiency in the implementation skills  
required to plan, install, and configure AIX High Availability Cluster  
Multi-Processing (HACMP) systems, and to perform the diagnostic activities  
needed to support Highly Available Clusters.  
Certification Requirement (two Tests):  
To attain the IBM Certified Specialist - AIX HACMP certification, candidates  
must first obtain the AIX System Administration or the AIX System Support  
certification. In order to obtain one of these prerequisite certifications, the  
candidate must pass one of the following two exams:  
Test 181: AIX V4.3 System Administration  
or  
Test 189: AIX V4.3 System Support.  
Following this, the candidate must pass the following exam:  
Test 167: HACMP for AIX V 4.2.  
Recommended Prerequisites  
A minimum of six to twelve months implementation experience installing,  
configuring, and testing/supporting HACMP for AIX.  
Registration for the Certification Exam  
For information about how to register for the certification exam, please visit  
the following Web site:  
http://www.ibm.com/certify  
© Copyright IBM Corp. 1999  
1
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
1.2 Certification Exam Objectives  
The following objectives were used as a basis for what is required when the  
certification exam was developed. Some of these topics have been  
regrouped to provide better organization when discussed in this publication.  
Section 1 - Preinstallation  
The following items should be considered as part of the preinstallation plan:  
• Conduct a Planning Session.  
• Set customer expectations at the beginning of the planning session.  
• Gather customer's availability requirements.  
• Articulate trade-offs of different HA configurations.  
• Assist customers in identifying HA applications.  
• Evaluate the Customer Environment and Tailorable Components.  
• Evaluate the configuration and identify Single Points of Failure (SPOF).  
• Define and analyze NFS requirements.  
• Identify components affecting HACMP.  
• Identify HACMP event logic customizations.  
• Plan for Installation.  
• Develop a disk management modification plan.  
• Understand issues regarding single adapter solutions.  
• Produce a Test Plan.  
Section 2 - HACMP Implementation  
The following items should be considered for proper implementation:  
• Configure HACMP Solutions.  
• Install HACMP Code.  
• Configure an IP Address Takeover (IPAT).  
• Configure non-IP heartbeat paths.  
• Configure a network adapter.  
• Customize/tailor AIX.  
• Set up a shared disk (SSA).  
• Set up a shared disk (SCSI).  
• Verify a cluster configuration.  
2
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
• Create an application server.  
• Set up Event Notification.  
• Set up event notification and pre/post event scripts.  
• Set up error notification.  
• Post Configuration Activities.  
• Configure a client notification and ARP update.  
• Implement a test plan.  
• Create a snapshot.  
• Create a customization document.  
• Perform Testing and Troubleshooting.  
• Troubleshoot a failed IPAT failover.  
• Troubleshoot failed shared volume groups.  
• Troubleshoot a failed network configuration.  
• Troubleshoot failed shared disk tests.  
• Troubleshoot a failed application.  
• Troubleshoot failed Pre/Post event scripts.  
• Troubleshoot failed error notifications.  
• Troubleshoot errors reported by cluster verification.  
Section 3 - System Management  
The following items should be considered for System Management:  
• Communicate with the Customer.  
• Conduct a turnover session.  
• Provide hands-on customer education.  
• Set customer expectations of their HACMP solution's capabilities.  
• Perform Systems Maintenance.  
• Perform HACMP maintenance tasks (PTFs, adding products, replacing  
disks, adapters).  
• Perform AIX maintenance tasks.  
• Dynamically update the cluster configuration.  
• Perform testing and troubleshooting as a result of changes.  
Certification Overview  
3
Download from Www.Somanuals.com. All Manuals Search And Download.  
1.3 Certification Education Courses  
Courses and publications are offered to help you prepare for the certification  
tests. These courses are recommended, but not required, before taking a  
certification test. At the printing of this guide, the following courses are  
available. For a current list, please visit the following Web site:  
http://www.ibm.com/certify  
Table 1. AIX Version 4 HACMP Installation and Implementation  
Course Number  
Course Duration Five days  
Course Abstract This course provides a detailed understanding of the  
Q1054 (USA) AU54 (Worldwide)  
High Availability Clustered Multi-Processing for AIX.  
The course is supplemented with a series of laboratory  
exercises to configure the hardware and software  
environments for HACMP. Additionally, the labs provide  
the opportunity to:  
• Install the product.  
• Define networks.  
• Create file systems.  
• Complete several modes of HACMP installations.  
4
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
The following table outlines information about the next course.  
Table 2. AIX Version 4 HACMP System Administration  
Q1150 (USA); AU50 (Worldwide)  
Five days  
Course Number  
Course Duration  
Course Abstract  
This course teaches the student the skills required to  
administer an HACMP cluster on an ongoing basis  
after it is installed. The skills that are developed in this  
course include:  
• Integrating the cluster with existing network  
services (DNS, NIS, etc.)  
• Monitoring tools for the cluster, including HAView  
for Netview  
• Maintaining user IDs and passwords across the  
cluster  
• Recovering from script failures  
• Making configuration or resource changes in the  
cluster  
• Repairing failed hardware  
• Maintaining required cluster documentation  
The course involves a significant number of hands-on  
exercises to reinforce the concepts. Students are  
expected to have completed the course AU54  
(Q1054) HACMP Installation and Implementation  
before attending this course.  
Certification Overview  
5
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
6
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Chapter 2. Cluster Planning  
The area of cluster planning is a large one. Not only does it include planning  
for the types of hardware (CPUs, networks, disks) to be used in the cluster,  
but it also includes other aspects. These include resource planning, that is,  
planning the desired behavior of the cluster in failure situations. Resource  
planning must take into account application loads and characteristics, as well  
as priorities. This chapter will cover all of these areas, as well as planning for  
event customizations and user id planning issues.  
2.1 Cluster Nodes  
One of HACMP’s key design strengths is its ability to provide support across  
the entire range of RISC System/6000 products. Because of this built-in  
flexibility and the facility to mix and match RISC System/6000 products, the  
effort required to design a highly available cluster is significantly reduced.  
In this chapter, we shall outline the various hardware options supported by  
HACMP for AIX and HACMP/ES. We realize that the rapid pace of change in  
products will almost certainly render any snapshot of the options out of date  
by the time it is published. This is true of almost all technical writing, though  
to yield to the spoils of obsolescence would probably mean nothing would  
ever make it to the printing press.  
The following sections will deal with the various:  
• CPU Options  
• Cluster Node Considerations  
available to you when you are planning your HACMP cluster.  
2.1.1 CPU Options  
HACMP is designed to execute with RISC System/6000 uniprocessors,  
Symmetric Multi-Processor (SMP) servers and the RS/6000 Scalable  
POWERparallel Systems (RS/6000 SP) in a no single point of failure server  
configuration. The minimum configuration and sizing of each system CPU is  
highly dependent on the user's application and data requirements.  
Nevertheless, systems with 32 MB of main storage and 1 GB of disk storage  
would be practical, minimum configurations.  
Almost any model of the RISC System/6000 POWERserver family can be  
included in an HACMP environment and new models continue to be added to  
the list. The following table gives you an overview of the currently supported  
© Copyright IBM Corp. 1999  
7
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
RISC System/6000 models as nodes in an HACMP 4.1 for AIX, HACMP 4.2  
for AIX, or HACMP 4.3 for AIX cluster.  
Table 3. Hardware Requirements for the Different HACMP Versions  
HACMP Version  
4.1  
4.2  
4.3  
4.2/ES  
4.3/ES  
1
7009 Mod. CXX  
yes  
yes  
yes  
yes  
yes  
yes  
yes  
yes  
yes  
yes  
yes  
yes  
yes  
yes  
yes  
yes  
yes  
yes  
yes  
yes  
yes  
yes  
yes  
yes  
yes  
yes  
yes  
yes  
yes  
yes  
yes  
yes  
yes  
no  
no  
no  
no  
no  
no  
no  
no  
no  
no  
yes  
yes  
1
7011 Mod. 2XX  
yes  
1
7012 Mod. 3XX and GXX  
7013 Mod. 5XX and JXX  
7015 Mod. 9XX and RXX  
7017 Mod. S7X  
yes  
1
yes  
1
yes  
1
yes  
1
7024 Mod. EXX  
yes  
1
7025 Mod. FXX  
yes  
1
7026 Mod. HXX  
yes  
1
7043 Mod. 43P, 260  
9076 RS/6000 SP  
yes  
1
yes  
1
AIX 4.3.2 required  
For a detailed description of system models supported by HACMP/6000 and  
HACMP/ES, you should refer to the current Announcement Letters for  
HACMP/6000 and HACMP/ES.  
HACMP/ES 4.3 further enhances cluster design flexibility even further by  
including support for the RISC System/6000 family of machines and the  
Compact Server C20. Since the introduction of HACMP 4.1 for AIX, you have  
been able to mix uniprocessor and multiprocessor machines in a single  
cluster. Even a mixture of “normal” RS/6000 machines and RS/6000 SP  
nodes is possible.  
2.1.2 Cluster Node Considerations  
It is important to understand that selecting the system components for a  
cluster requires careful consideration of factors and information that may not  
be considered in the selection of equipment for a single-system environment.  
In this section, we will offer some guidelines to assist you in choosing and  
sizing appropriate machine models to build your clusters.  
8
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
Much of the decision centers around the following areas:  
• Processor capacity  
• Application requirements  
• Anticipated growth requirements  
• I/O slot requirements  
These paradigms are certainly not new ones, and are also important  
considerations when choosing a processor for a single-system environment.  
However, when designing a cluster, you must carefully consider the  
requirements of the cluster as a total entity. This includes understanding  
system capacity requirements of other nodes in the cluster beyond the  
requirements of each system's prescribed normal load. You must consider  
the required performance of the solution during and after failover, when a  
surviving node has to add the workload of a failed node to its own workload.  
For example, in a two node cluster, where applications running on both nodes  
are critical to the business, each of the two nodes functions as a backup for  
the other, in a mutual takeover configuration. If a node is required to provide  
failover support for all the applications on the other node, then its capacity  
calculation needs to take that into account. Essentially, the choice of a model  
depends on the requirements of highly available applications, not only in  
terms of CPU cycles, but also of memory and possibly disk space.  
Approximately 50 MB of disk storage is required for full installation of the  
HACMP software.  
A major consideration in the selection of models will be the number of I/O  
expansion slots they provide. The model selected must have enough slots to  
house the components required to remove single points of failure (SPOFs)  
and provide the desired level of availability. A single point of failure is defined  
as any single component in a cluster whose failure would cause a service to  
become unavailable to end users. The more single points of failure you can  
eliminate, the higher your level of availability will be. Typically, you need to  
consider the number of slots required to support network adapters and disk  
I/O adapters. Your slot configuration must provide for at least two network  
adapters to provide adapter redundancy for one service network. If your  
system needs to be able to take over an IP address for more than one other  
system in the cluster at a time, you will want to configure more standby  
network adapters. A node can have up to seven standby adapters for each  
network it connects to. Again, if that is your requirement, you will need to  
select models as nodes where the number of slots will accomodate the  
requirement.  
Cluster Planning  
9
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
Your slot configuration must also allow for the disk I/O adapters you need to  
support the cluster’s shared disk (volume group) configuration. If you intend  
to use disk mirroring for shared volume groups, which is strongly  
recommended, then you will need to use slots for additional disk I/O  
adapters, providing I/O adapter redundancy across separate buses.  
The following table tells you the number of additional adapters you can put  
into the different RS/6000 models. Ethernet environments can sometimes  
make use of the integrated ethernet port provided by some models. No such  
feature is available for token-ring, FDDI or ATM; you must use an I/O slot to  
provide token-ring adapter redundancy.  
Table 4. Number of Adapter Slots in Each Model  
RS/6000 Model  
7006  
Number of Slots  
4 x MCA  
Integrated Ethernet Port  
yes  
no  
7009 C10, C20  
7012 Mod. 3XX and GXX  
7013 Mod. 5XX  
7013 Mod. JXX  
4x PCI  
4 x MCA  
yes  
no  
7 x MCA  
6 x MCA, 14 x MCA no  
with expansion unit J01  
7015 Mod. R10, R20, R21  
7015 Mod. R30, R40, R50  
7017 Mod. S7X  
8 x MCA  
16 x MCA  
52 x PCI  
no  
no  
no  
7024 EXX  
5 x PCI, 1 x PCI/ISA 2 x no  
ISA  
7025 F50  
6 x PCI, 2 x ISA/PCI  
6 x PCI, 2 x ISA/PCI  
3 x PCI, 2 x ISA/PCI  
4 x MCA  
yes  
7026 Mod. H50  
7043Mod.  
yes  
yes  
yes  
no  
9076 thin node  
9076 wide node  
9076 high node  
9076 thin node (silver)  
9076 wide node (silver)  
7 x MCA  
15 x MCA  
no  
1
1
2 x PCI  
yes  
10 x PCI  
yes  
1
The switch adapter is onboard and does not need an extra slot.  
10  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
2.2 Cluster Networks  
HACMP differentiates between two major types of networks: TCP/IP networks  
and non-TCP/IP networks. HACMP utilizes both of them for exchanging  
heartbeats. HACMP uses these heartbeats to diagnose failures in the cluster.  
Non-TCP/IP networks are used to distinguish an actual hardware failure from  
the failure of the TCP/IP software. If there were only TCP/IP networks being  
used, and the TCP/IP software failed, causing heartbeats to stop, HACMP  
could falsely diagnose a node failure when the node was really still  
functioning. Since a non-TCP/IP network would continue working in this  
event, the correct diagnosis could be made by HACMP. In general, all  
networks are also used for verification, synchronization, communication and  
triggering events between nodes. Of course, TCP/IP networks are used for  
communication with client machines as well.  
At the time of publication, the HACMP/ES Version 4.3 product does not use  
non-TCP/IP networks for node-to-node communications in triggering,  
synchronizing, and executing event reactions. This can be an issue if you are  
configuring a cluster with only one TCP/IP network. This limitation of  
HACMP/ES is planned to be removed in a future release. You would be  
advised to check on the status of this issue if you are planning a new  
installation, and to plan your cluster networks accordingly.  
2.2.1 TCP/IP Networks  
The following sections describe supported TCP/IP network types and network  
considerations.  
2.2.1.1 Supported TCP/IP Network Types  
Basically every adapter that is capable of running the TCP/IP protocol is a  
supported HACMP network type. There are some special considerations for  
certain types of adapters however. The following gives a brief overview on the  
supported adapters and their special considerations.  
Below is a list of TCP/IP network types as you will find them at the  
configuration time of an adapter for HACMP. You will find the non-TCP/IP  
• Generic IP  
• ATM  
• Ethernet  
• FCS  
Cluster Planning 11  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
• FDDI  
• SP Switch  
• SLIP  
• SOCC  
Token-Ring  
As an independent, layered component of AIX, the HACMP for AIX software  
works with most TCP/IP-based networks. HACMP for AIX has been tested  
with standard Ethernet interfaces (en*) but not with IEEE 802.3 Ethernet  
interfaces (et*), where * reflects the interface number. HACMP for AIX also  
has been tested with Token-Ring and Fiber Distributed Data Interchange  
(FDDI) networks, with IBM Serial Optical Channel Converter (SOCC), Serial  
Line Internet Protocol (SLIP), and Asynchronous Transfer Mode (ATM)  
point-to-point connections.  
Note  
ATM and SP Switch networks are special cases of point-to-point, private  
networks that can connect clients  
The HACMP for AIX software supports a maximum of 32 networks per cluster  
and 24 TCP/IP network adapters on each node. These numbers provide a  
great deal of flexibility in designing a network configuration. The network  
design affects the degree of system availability in that the more  
communication paths that connect clustered nodes and clients, the greater  
the degree of network availability.  
2.2.1.2 Special Network Considerations  
Each type of interface has different characteristics concerning speed, MAC  
addresses, ARP, and so on. In case there is a limitation you will have to work  
around, you need to be aware of the characteristics of the adapters you plan  
to use. In the next paragraphs, we summarize some of the considerations  
that are known.  
Hardware Address Swapping is one issue. If you enable HACMP to put one  
address on another adapter, it would need something like a boot and a  
service address for IPAT, but on the hardware layer. So, in addition to the  
manufacturers burnt-in address, there has to be an alternate address  
configured.  
The speed of the network can be another issue. Your application may have  
special network throughput requirements that must be taken into account.  
12  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
Network types also differentiate themselves in the maximum distance they  
allow between adapters, and in the maximum number of adapters allowed on  
a physical network.  
Ethernet supports 10 and 100 Mbps currently, and supports hardware  
address swapping. Alternate hardware addresses should be in the form  
xxxxxxxxxxyy, where xxxxxxxxxx is replaced with the first five pairs of digits  
of the original burned-in MAC address and yy can be chosen freely. There  
is a limit of 29 adapters on one physical network, unless a network  
repeater is used.  
Token-Ring supports 4 or 16 Mbps, but 4 Mbps is very rarely used now. It  
also supports hardware address swapping, but here the convention is to  
use 42 as the first two characters of the alternate address, since this  
indicates that it is a locally set address.  
FDDI is a 100 Mbps optical LAN interface, that supports hardware address  
takeover as well. For FDDI adapters you should leave the last six digits of  
the burned-in address as they are, and use a 4, 5, 6, or 7 as the first digit  
of the rest. FDDI can connect as many as 500 stations with a maximum  
link-to-link distance of two kilometers and a total LAN circumference of  
100 kilometers.  
• ATM is a point-to-point connection network. It currently supports the OC3  
and the OC12 standard, which is 155 Mbps or 625 Mbps. You cannot use  
hardware address swapping with ATM. ATM doesn’t support broadcasts,  
so it must be configured as a private network to HACMP. However, if you  
are using LAN Emulation on an existing ATM network, you can use the  
emulated ethernet or Token-Ring interfaces just as if they were real ones,  
except that you cannot use hardware address swapping.  
FCS is a fiber channel network, currently available as two adapters for  
either MCA or PCI technology. The Fibre Channel Adapter /1063-MCA,  
runs up to 1063 Mb/second, and the Gigabit Fibre Channel Adapter for  
PCI Bus (#6227), announced on October 5th 1998, will run with 100 MBps.  
Both of them support TCP/IP, but not hardware address swapping.  
SLIP runs at up to 38400 bps. Since it is a point-to-point connection and  
very slow, it is rarely used as an HACMP network. An HACMP cluster is  
much more likely to use the serial port as a non-TCP/IP connection. See  
below for details.  
SOCC is a fast optical connection, again point-to-point. This is an optical  
line with a serial protocol running on it. However, the SOCC Adapter  
(Feature 2860) has been withdrawn from marketing for some years now.  
Some models, like 7013 5xx, offer SOCC as an option onboard, but these  
are rarely used today.  
Cluster Planning 13  
Download from Www.Somanuals.com. All Manuals Search And Download.  
             
• SP Switch is a high-speed packet switching network, running on the  
RS/6000 SP system only. It runs bidirectionally up to 80 MBps, which adds  
up to 160 MBps of capacity per adapter. This is node-to-node  
communication and can be done in parallel between every pair of nodes  
inside an SP. The SP Switch network has to be defined as a private  
Network, and ARP must be enabled. This network is restricted to one  
adapter per node, thus, it has to be considered as a Single Point Of  
Failure. Therefore, it is strongly recommended to use AIX Error  
Notification to propagate a switch failure into a node failure when  
appropriate. As there is only one adapter per node, HACMP uses the  
ifconfig alias addresses for IPAT on the switch; so, a standby address is  
not necessary and, therefore, not used on the switch network. Hardware  
address swapping also is not supported on the SP Switch.  
For IP Address Takeover (IPAT), in general, there are two adapters per  
cluster node and network recommended in order to eliminate single points of  
failure. The only exception to this rule is the SP Switch because of hardware  
limitations.  
2.2.2 Non-TCPIP Networks  
Non-TCP/IP networks in HACMP are used as an independent path for  
exchanging messages or heartbeats between cluster nodes. In case of an IP  
subsystem failure, HACMP can still differentiate between a network failure  
and a node failure when an independent path is available and functional.  
Below is a short description of the three currently available non-TCP/IP  
network types and their characteristics. Even though HACMP works without  
one, it is strongly recommended that you use at least one non-TCP/IP  
connection between the cluster nodes.  
2.2.2.1 Supported Non-TCP/IP Network Types  
Currently HACMP supports the following types of networks for non-TCP/IP  
heartbeat exchange between cluster nodes:  
• Serial (RS232)  
Target-mode SCSI  
Target-mode SSA  
All of them must be configured as Network Type: serial in the HACMP  
definitions.  
14  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
             
2.2.2.2 Special Considerations  
As for TCP/IP networks, there are a number of restrictions on non-TCP/IP  
networks. These are explained for the three different types in more detail  
below.  
Serial (RS232)  
A serial (RS232) network needs at least one available serial port per cluster  
node. In case of a cluster consisting of more than two nodes, a ring of nodes  
is established through serial connections, which requires two serial ports per  
node. Table 3 shows a list of possible cluster nodes and the number of native  
serial ports for each:  
Table 5. Number of Available Serial Ports in Each Model.  
RS/6000 Model  
7006  
Number of Serial Ports Available  
1
1
1
7009 C10, C20  
1
7012 Mod. 3XX and GXX  
7013 Mod. 5XX  
2
2
3
3
3
7013 Mod. JXX  
7015 Mod. R10, R20, R21  
7015 Mod. R30, R40, R50  
7013,7015,7017 Mod. S7X  
7024 EXX  
2
0
1
1
7025 F50  
2
3
2
7026 Mod. H50  
7043Mod.  
3
9076 thin node  
2
3
9076 wide node  
9076 high node  
2
3
3
3
9076 thin node (silver)  
9076 wide node (silver)  
2
3
2
1
serial port can be multiplexed through a dual-port cable, thus offering two  
ports  
Cluster Planning 15  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
2
3
a PCI Multiport Async Card is required in an S7X model, no native ports  
only one serial port available for customer use, i.e. HACMP  
In case the number of native serial ports doesn’t match your HACMP cluster  
configuration needs, you can extend it by adding an eight-port asynchronous  
adapter, thus reducing the number of available MCA slots, or the  
corresponding PCI Multiport Async Card for PCI Machines, like the S7X  
model.  
Target-mode SCSI  
Another possibility for a non-TCP/IP network is a target mode SCSI  
connection. Whenever you make use of a shared SCSI device, you can also  
use the SCSI bus for exchanging heartbeats.  
Target Mode SCSI is only supported with SCSI-2 Differential or SCSI-2  
Differential Fast/Wide devices. SCSI/SE or SCSI-2/SE are not supported for  
HACMP serial networks.  
The recommendation is to not use more than 4 target mode SCSI networks in  
a cluster.  
Target-mode SSA  
If you are using shared SSA devices, target mode SSA is the third possibility  
for a serial network within HACMP. In order to use target-mode SSA, you  
must use the Enhanced RAID-5 Adapter (#6215 or #6219), since these are  
the only current adapters that support the Multi-Initiator Feature. The  
microcode level of the adapter must be 1801 or higher.  
2.3 Cluster Disks  
This section describes the various choices you have in selecting the type of  
shared disks to use in your cluster.  
2.3.1 SSA Disks  
The following is a brief description of SSA and the basic rules to follow when  
designing SSA networks. For a full description of SSA and its functionality,  
please read Monitoring and Managing IBM SSA Disk Subsystems,  
SG24-5251.  
SSA is a high-performance, serial interconnect technology used to connect  
disk devices and host adapters. SSA is an open standard, and SSA  
specifications have been approved by the SSA Industry Association and also  
as an ANSI standard through the ANSI X3T10.1 subcommittee.  
16  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
SSA subsystems are built up from loops of adapters and disks. A simple  
example is shown in Figure 1.  
SSA Architecture  
MBp  
20 M  
HOST  
M
l
l
l
l
l
High performance 80 MB/s interface  
Loop architecture with up to 127 nodes per loop  
Up to 25 m (82 ft) between SSA devices with copper cables  
Up to 2.4 km (1.5 mi) between SSA devices with optical extender  
Spatial reuse (multiple simultaneous transmissions)  
Figure 1. Basic SSA Configuration  
Here, a single adapter controls one SSA loop of eight disks. Data can be  
transferred around the loop, in either direction, at 20 MBps. Consequently,  
the peak transfer rate of the adapter is 80 MBps. The adapter contains two  
SSA nodes and can support two SSA loops. Each disk drive also contains a  
single SSA node. A node can be either an initiator or a target. An initiator  
issues commands, while a target responds with data and status information.  
The SSA nodes in the adapter are therefore initiators, while the SSA nodes in  
the disk drives are targets.  
There are two types of SSA Disk Subsystems for RISC System/6000  
available:  
• 7131 SSA Multi-Storage Tower Model 405  
Cluster Planning 17  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
• 7133 Serial Storage Architecture (SSA) Disk Subsystem Models 010, 500,  
020, 600, D40 and T40.  
The 7133 models 010 and 500 were the first SSA products announced in  
1995 with the revolutionary new Serial Storage Architecture. Some IBM  
customers still use the Models 010 and 500, but these have been replaced by  
7133 Model 020, and 7133 Model 600 respectively. More recently, in  
November 1998, the models D40 and T40 were announced.  
All 7133 Models have redundant power and cooling, which is hot-swappable.  
The following tables give you more configuration information about the  
different models:  
Table 6. 7131-Model 405 SSA Multi-Storage Tower Specifications  
Item  
Specification  
80 MB  
Transfer rate SSA  
interface  
Configuration  
2 to 5 disk drives (2.2 GB, 4.5 GB or 9.1 GB) per subsystem  
Configuration range  
4.4 to 11 GB (with 2.2 GB disk drives)  
9.0 to 22.5 GB (With 4.5 GB disk drives)  
18.2 to 45.5 GB (With 9.1 GB disk drives)  
Supported RAID  
levels  
5
Supported adapters  
Hot-swap disks  
6214, 6216, 6217, 6218  
Yes  
Table 7. 7133 Models 010, 020, 500, 600, D40, T40 Specifications  
Item  
Specification  
80 MB/s  
Transfer rate SSA  
interface  
Configuration  
4 to 16 disks  
- 1.1 GB, 2.2 GB, 4.5 GB, for Models 10, 20, 500, and 600  
- 9.1 GB for Models 20, 600, D40 and T40  
- With 1.1 GB disk drives you must have 8 to 16 disks)  
Configuration range  
8.8 to 17.6 GB (with 1.1 GB disks)  
8.8 to 35.2 GB (with 2.2 GB disks)  
18 to 72 GB (with 4.5 GB disks)  
36.4 to 145.6 GB (with 9.1 GB disks)  
72.8 to 291.2 GB (with 18.2 GB disks)  
18  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
Item  
Specification  
Supported RAID level  
Supported adapters  
Hot-swappable disk  
5
all  
Yes (and hot-swappable, redundant power and cooling)  
2.3.1.1 Disk Capacities  
Table 8 lists the different SSA disks, and provides an overview of their  
characteristics.  
Table 8. SSA Disks  
Name  
Capacities (GB)  
Buffer size (KB)  
Maximum Transfer rate  
(MBps)  
Starfire 1100  
Starfire 2200  
Starfire 4320  
Scorpion 4500  
Scorpion 9100  
Sailfin 9100  
Thresher 9100  
Ultrastar  
1.1  
0
20  
2.2  
0
20  
4.5  
512  
512  
512  
1024  
1024  
4096  
20  
4.5  
80  
9.1  
160  
160  
160  
160  
9.1  
9.1  
9.1, 18.2  
2.3.1.2 Supported and Non-Supported Adapters  
Table 9 lists the different SSA adapters and presents an overview of their  
characteristics.  
Table 9. SSA Adapters  
Feature Code Adapter  
Label  
Bus  
Adapter Description  
Number  
Adapters per Raid Types  
Loop  
of Hardware  
6214  
6215  
6216  
6217  
6218  
4-D  
4-N  
4-G  
4-I  
MCA  
PCI  
Classic  
2
8
8
1
1
n/a  
5
1
Enhanced RAID-5  
Enhanced  
RAID-5  
MCA  
MCA  
PCI  
n/a  
5
4-J  
RAID-5  
5
Cluster Planning 19  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Feature Code Adapter  
Label  
Bus  
Adapter Description  
Enhanced RAID-5  
Number  
Adapters per Raid Types  
Loop  
of Hardware  
1
6219  
4-M  
MCA  
8
5
1See 2.3.1.3, “Rules for SSA Loops” on page 20 for more information.  
The following rules apply to SSA Adapters:  
• You cannot have more than four adapters in a single system.  
• The MCA SSA 4-Port RAID Adapter (FC 6217) and PCI SSA 4-Port RAID  
Adapter (FC 6218) are not useful for HACMP, because only one can be in  
a loop.  
• Only the PCI Multi Initiator/RAID Adapter (FC 6215) and the MCA Multi  
Initiator/RAID EL Adapter (FC 6219) support target mode SSA (for more  
information about target mode SSA see 3.2.2, “Non TCP/IP Networks” on  
page 63).  
2.3.1.3 Rules for SSA Loops  
The following rules must be followed when configuring and connecting SSA  
loops:  
• Each SSA loop must be connected to a valid pair of connectors on the  
SSA adapter (that is, either Connectors A1 and A2, or Connectors B1 and  
B2).  
• Only one of the two pairs of connectors on an adapter card can be  
connected in a single SSA loop.  
• A maximum of 48 devices can be connected in a single SSA loop.  
• A maximum of two adapters can be connected in a particular loop if one  
adapter is an SSA 4-Port adapter, Feature 6214.  
• A maximum of eight adapters can be connected in a particular loop if all  
the adapters are Enhanced SSA 4-Port Adapters, Feature 6216.  
• A maximum of two SSA adapters, both connected in the same SSA loop,  
can be installed in the same system.  
For SSA loops that include an SSA Four-Port RAID adapter (Feature 6217) or  
a PCI SSA Four-Port RAID adapter (Feature 6218), the following rules apply:  
• Each SSA loop must be connected to a valid pair of connectors on the  
SSA adapter (that is, either Connectors A1 and A2, or Connectors B1 and  
B2).  
20  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
• A maximum of 48 devices can be connected in a particular SSA loop.  
• Only one pair of adapter connectors can be connected in a particular SSA  
loop.  
• Member disk drives of an array can be on either SSA loop.  
For SSA loops that include a Micro Channel Enhanced SSA  
Multi-initiator/RAID EL adapter, Feature 6215 or a PCI SSA  
Multi-initiator/RAID EL adapter, Feature 6219, the following rules apply:  
• Each SSA loop must be connected to a valid pair of connectors on the  
SSA adapter (that is, either Connectors A1 and A2, or Connectors B1 and  
B2).  
• A maximum of eight adapters can be connected in a particular loop if none  
of the disk drives in the loops are array disk drives and none of them is  
configured for fast-write operations. The adapters can be up to eight  
Micro Channel Enhanced SSA Multi-initiator/RAID EL Adapters, up to  
eight PCI Multi-initiator/RAID EL Adapters, or a mixture of the two types.  
• A maximum of two adapters can be connected in a particular loop if one or  
more of the disk drives in the loop are array disk drives that are not  
configured for fast-write operations. The adapters can be two Micro  
Channel Enhanced SSA Multi-initiator/RAID EL Adapters, two PCI  
Multi-initiator/RAID EL Adapters, or one adapter of each type.  
• Only one Micro Channel Enhanced SSA Multi-initiator/RAID EL Adapter or  
PCI SSA Multi-initiator/RAID EL Adapter can be connected in a particular  
loop if any disk drives in the loops are members of a RAID-5 array, and are  
configured for fast-write operations.  
• All member disk drives of an array must be on the same SSA loop.  
• A maximum of 48 devices can be connected in a particular SSA loop.  
• Only one pair of adapter connectors can be connected in a particular loop.  
• When an SSA adapter is connected to two SSA loops, and each loop is  
connected to a second adapter, both adapters must be connected to both  
loops.  
For the IBM 7190-100 SCSI to SSA converter, the following rules apply:  
• There can be up to 48 disk drives per loop.  
• There can be up to four IBM 7190-100 attached to any one SSA loop.  
Cluster Planning 21  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
2.3.1.4 RAID vs. Non-RAID  
RAID Technology  
RAID is an acronym for Redundant Array of Independent Disks. Disk arrays  
are groups of disk drives that work together to achieve higher data-transfer  
and I/O rates than those provided by single large drives.  
Arrays can also provide data redundancy so that no data is lost if a single  
drive (physical disk) in the array should fail. Depending on the RAID level,  
data is either mirrored or striped. The following gives you more information  
about the different RAID levels.  
RAID Level 0  
RAID 0 is also known as data striping. Conventionally, a file is written out  
sequentially to a single disk. With striping, the information is split into chunks  
(fixed amounts of data usually called blocks) and the chunks are written to (or  
read from) a series of disks in parallel. There are two main performance  
advantages to this:  
1. Data transfer rates are higher for sequential operations due to the  
overlapping of multiple I/O streams.  
2. Random access throughput is higher because access pattern skew is  
eliminated due to the distribution of the data. This means that with data  
distributed evenly across a number of disks, random accesses will most  
likely find the required information spread across multiple disks and thus  
benefit from the increased throughput of more than one drive.  
RAID 0 is only designed to increase performance. There is no redundancy; so  
any disk failures will require reloading from backups.  
RAID Level 1  
RAID 1 is also known as disk mirroring. In this implementation, identical  
copies of each chunk of data are kept on separate disks, or more commonly,  
each disk has a twin that contains an exact replica (or mirror image) of the  
information. If any disk in the array fails, then the mirrored twin can take over.  
Read performance can be enhanced because the disk with its actuator  
closest to the required data is always used, thereby minimizing seek times.  
The response time for writes can be somewhat slower than for a single disk,  
depending on the write policy; the writes can either be executed in parallel for  
speed or serially for safety.  
RAID Level 1 has data redundancy, but data should be regularly backed up  
on the array. This is the only way to recover data in the event that a file or  
directory is accidentally deleted.  
22  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
RAID Levels 2 and 3  
RAID 2 and RAID 3 are parallel process array mechanisms, where all drives  
in the array operate in unison. Similar to data striping, information to be  
written to disk is split into chunks (a fixed amount of data), and each chunk is  
written out to the same physical position on separate disks (in parallel). When  
a read occurs, simultaneous requests for the data can be sent to each disk.  
This architecture requires parity information to be written for each stripe of  
data; the difference between RAID 2 and RAID 3 is that RAID 2 can utilize  
multiple disk drives for parity, while RAID 3 can use only one. If a drive should  
fail, the system can reconstruct the missing data from the parity and  
remaining drives.  
Performance is very good for large amounts of data but poor for small  
requests since every drive is always involved, and there can be no  
overlapped or independent operation.  
RAID Level 4  
RAID 4 addresses some of the disadvantages of RAID 3 by using larger  
chunks of data and striping the data across all of the drives except the one  
reserved for parity. Using disk striping means that I/O requests need only  
reference the drive that the required data is actually on. This means that  
simultaneous, as well as independent reads, are possible. Write requests,  
however, require a read/modify/update cycle that creates a bottleneck at the  
single parity drive. Each stripe must be read, the new data inserted and the  
new parity then calculated before writing the stripe back to the disk. The  
parity disk is then updated with the new parity, but cannot be used for other  
writes until this has completed. This bottleneck means that RAID 4 is not  
used as often as RAID 5, which implements the same process but without the  
bottleneck. RAID 5 is discussed in the next section.  
RAID Level 5  
RAID 5, as has been mentioned, is very similar to RAID 4. The difference is  
that the parity information is distributed across the same disks used for the  
data, thereby eliminating the bottleneck. Parity data is never stored on the  
same drive as the chunks that it protects. This means that concurrent read  
and write operations can now be performed, and there are performance  
increases due to the availability of an extra disk (the disk previously used for  
parity). There are other enhancements possible to further increase data  
transfer rates, such as caching simultaneous reads from the disks and  
transferring that information while reading the next blocks. This can generate  
data transfer rates that approach the adapter speed.  
Cluster Planning 23  
Download from Www.Somanuals.com. All Manuals Search And Download.  
As with RAID 3, in the event of disk failure, the information can be rebuilt from  
the remaining drives. RAID level 5 array also uses parity information, though  
it is still important to make regular backups of the data in the array. RAID level  
5 stripes data across all of the drives in the array, one segment at a time (a  
segment can contain multiple blocks). In an array with n drives, a stripe  
consists of data segments written to n-1 of the drives and a parity segment  
written to the nth drive. This mechanism also means that not all of the disk  
space is available for data. For example, in an array with five 2 GB disks,  
although the total storage is 10 GB, only 8 GB are available for data.  
The advantages and disadvantages of the various RAID levels are  
summarized in the following table:  
Table 10. The Advantages and Disadvantages of the Different RAID Levels  
RAID Level  
Availability  
Mechanism  
Capacity  
Performance  
Cost  
0
1
3
5
none  
100%  
50%  
80%  
80%  
high  
medium  
high  
mirroring  
parity  
medium/high  
medium  
medium  
medium  
medium  
parity  
RAID on the 7133 Disk Subsystem  
The only RAID level supported by the 7133 SSA disk subsystem is RAID 5.  
RAID 0 and RAID 1 can be achieved with the striping and mirroring facility of  
the Logical Volume Manager (LVM).  
RAID 0 does not provide data redundancy, so it is not recommended for use  
with HACMP, because the shared disks would be a single point of failure. The  
possible configurations to use with the 7133 SSA disk subsystem are RAID 1  
(mirroring) or RAID 5. Consider the following points before you make your  
decision:  
• Mirroring is more expensive than RAID, but it provides higher data  
redundancy. Even if more than one disk fails, you may still have access to  
all of your data. In a RAID, more than one broken disk means that the data  
are lost.  
• The SSA loop can include a maximum of two SSA adapters if you use  
RAID. So, if you want to connect more than two nodes into the loop,  
mirroring is the way to go.  
• A RAID array can consist of three to 16 disks.  
24  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
• Array member drives and spares must be on same loop (cannot span A  
and B loops) on the adapter.  
• You cannot boot (ipl) from a RAID.  
2.3.1.5 Advantages  
Because SSA allows SCSI-2 mapping, all functions associated with initiators,  
targets, and logical units are translatable. Therefore, SSA can use the same  
command descriptor blocks, status codes, command queuing, and all other  
aspects of current SCSI systems. The effect of this is to make the type of disk  
subsystem transparent to the application. No porting of applications is  
required to move from traditional SCSI I/O subsystems to high-performance  
SSA. SSA and SCSI I/O systems can coexist on the same host running the  
same applications.  
The advantages of SSA are summarized as follows:  
• Dual paths to devices.  
• Simplified cabling - cheaper, smaller cables and connectors, no separate  
terminators.  
• Faster interconnect technology.  
• Not an arbitrated system.  
• Full duplex, frame multiplexed serial links.  
• 40 MBps total per port, resulting in 80 MBps total per node, and 160 MBps  
total per adapter.  
• Concurrent access to disks.  
• Hot-pluggable cables and disks.  
• Very high capacity per adapter - up to 127 devices per loop, although most  
adapter implementations limit this. For example, current IBM SSA  
adapters provide 96 disks per Micro Channel or PCI slot.  
• Distance between devices of up to 25 meters with copper cables, 10km  
with optical links.  
• Auto-configuring - no manual address allocation.  
• SSA is an open standard.  
• SSA switches can be introduced to produce even greater fan-out and  
more complex topologies.  
Cluster Planning 25  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
2.3.2 SCSI Disks  
After the announcement of the 7133 SSA Disk Subsystems, the SCSI Disk  
subsystems became less common in HACMP clusters. However, the 7135  
RAIDiant Array (Model 110 and 210) and other SCSI Subsystems are still in  
use at many customer sites. We will not describe other SCSI Subsystems  
such as 9334 External SCSI Disk Storage. See the appropriate  
documentation if you need information about these SCSI Subsystems.  
The 7135 RAIDiant Array is offered with a range of features, with a maximum  
capacity of 135 GB (RAID 0) or 108 GB (RAID-5) in a single unit, and uses  
the 4.5 GB disk drive modules. The array enclosure can be integrated into a  
RISC System/6000 system rack, or into a deskside mini-rack. It can attach to  
multiple systems through a SCSI-2 Differential 8-bit or 16-bit bus.  
2.3.2.1 Capacities  
Disks  
There are four disk sizes available for the 7135 RAIDiant Array Models 110  
and 210:  
• 1.3 GB  
• 2.0 GB  
• 2.2 GB (only supported by Dual Active Software)  
• 4.5 GB (only supported by Dual Active Software)  
Subsystems  
The 7135-110/210 can contain 15 Disks (max. 67.5 GB) in the base  
configuration and 30 Disks (max. 135 GB) in an extended configuration.You  
can for example only use the full 135 GB storage space for data if you  
configure the 7135 with RAID level 0. When using RAID level 5, only 108 GB  
of the 135 GB are available for data storage.  
2.3.2.2 How Many in a String?  
HACMP supports a maximum of two 7135s on a shared SCSI bus. This is  
because of cable length restrictions.  
2.3.2.3 Supported SCSI Adapters  
The SCSI adapters that can be used to connect RAID subsystems on a  
shared SCSI bus in an HACMP cluster are:  
• SCSI-2 Differential Controller (MCA, FC: 2420, Adapter Label: 4-2)  
• SCSI-2 Differential Fast/Wide Adapter/A (MCA, FC: 2416, Adapter Label:  
4-6)  
26  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
• Enhanced SCSI-2 Differential Fast/Wide Adapter/A (MCA, FC: 2412,  
Adapter Label: 4-C); not usable with 7135-110  
• SCSI-2 Fast/Wide Differential Adapter (PCI, FC: 6209, Adapter Label:  
4-B)  
• DE Ultra SCSI Adapter (PCI, FC: 6207, Adapter Label: 4-L); not usable  
with 7135-110  
2.3.2.4 Advantages - Disadvantages  
The 7135 RAIDiant Array incorporates the following high availability features:  
• Support for RAID-1, RAID-3 (Model 110 only) and RAID-5  
You can run any combination of RAID levels in a single 7135 subsystem.  
Each LUN can run its own RAID level.  
• Multiple Logical Unit (LUN) support  
The RAID controller takes up only one SCSI ID on the external bus. The  
internal disks are grouped into logical units (LUNs). The array will support  
up to six LUNs, each of which appears to AIX as a single hdisk device.  
Since each of these LUNs can be configured into separate volume groups,  
different parts of the subsystem can be logically attached to different  
systems at any one time.  
• Redundant Power Supply  
Redundant power supplies provide alternative sources of power. If one  
supply fails, power is automatically supplied by the other.  
• Redundant Cooling  
Extra cooling fans are built into the RAIDiant Array to safeguard against  
fan failure.  
• Concurrent Maintenance  
Power supplies, cooling fans, and failed disk drives can be replaced  
without the need to take the array offline or to power it down.  
• Optional Second Array Controller  
This allows the array subsystem to be configured with no single point of  
failure. Under the control of the system software, the machine can be  
configured in Dual Active mode, so that each controller controls the  
operation of specific sets of drives. In the event of failure of either  
controller, all I/O activity is switched to the remaining active controller.  
In the last few years, the 7133 SSA Subsystems have become more popular  
than 7135 RAIDiant Systems due to better technology. IBM decided to  
Cluster Planning 27  
Download from Www.Somanuals.com. All Manuals Search And Download.  
withdraw the 7135 RAIDiant Systems from marketing because it is equally  
possible to configure RAID on the SSA Subsystems.  
2.4 Resource Planning  
HACMP provides a highly available environment by identifying a set of  
cluster-wide resources essential to uninterrupted processing, and then  
defining relationships among nodes that ensure these resources are  
available to client processes.  
When a cluster node fails or detaches from the cluster for a scheduled  
outage, the Cluster Manager redistributes its resources among any number of  
the surviving nodes.  
HACMP considers the following as resource types:  
• Volume Groups  
• Disks  
• File Systems  
• File Systems to be NFS mounted  
• File Systems to be NFS exported  
• Service IP addresses  
• Applications  
The following paragraphs will tell you what to consider when configuring  
resources to accomplish the following:  
• IP Address Takeover  
• Shared LVM Components  
• NFS Exports  
and the options you have when combining these resources to a resource  
group.  
2.4.1 Resource Group Options  
Each resource in a cluster is defined as part of a resource group. This allows  
you to combine related resources that need to be together to provide a  
particular service. A resource group also includes the list of nodes that can  
acquire those resources and serve them to clients.  
A resource group is defined as one of three types:  
28  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
• Cascading  
• Rotating  
• Concurrent  
Each of these types describes a different set of relationships between nodes  
in the cluster, and a different set of behaviors upon nodes entering and  
leaving the cluster.  
Cascading Resource Groups: All nodes in a cascading resource group are  
assigned priorities for that resource group. These nodes are said to be part of  
that group's resource chain. In a cascading resource group, the set of  
resources cascades up or down to the highest priority node active in the  
cluster. When a node that is serving the resources fails, the surviving node  
with the highest priority takes over the resources.  
A parameter called Inactive Takeover decides which node takes the  
cascading resources when the nodes join the cluster for the first time. If this  
parameter is set to true, the first node in a group's resource chain to join the  
cluster acquires all the resources in the resource group. As successive nodes  
join the resource group, the resources cascade up to any node with a higher  
priority that joins the cluster. If this parameter is set to false, the first node in a  
group's resource chain to join the cluster acquires all the resources in the  
resource group only if it is the node with the highest priority for that group. If  
the first node to join does not acquire the resource group, the second node in  
the group's resource chain to join acquires the resource group, if it has a  
higher priority than the node already active. As successive nodes join, the  
resource group cascades to the active node with the highest priority for the  
group. The default is false.  
Member nodes of a cascading resource chain always release a resource  
group to a reintegrating node with a higher priority.  
Rotating Resource Groups: A rotating resource group is associated with a  
group of nodes, rather than a particular node. A node can be in possession of  
a maximum of one rotating resource group per network.  
As participating nodes join the cluster for the first time, they acquire the first  
available rotating resource group per network until all the groups are  
acquired. The remaining nodes maintain a standby role.  
When a node holding a rotating resource group leaves the cluster, either  
because of a failure or gracefully while specifying the takeover option, the  
node with the highest priority and available connectivity takes over. Upon  
Cluster Planning 29  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
reintegration, a node remains as a standby and does not take back any of the  
resources that it had initially served.  
Concurrent Resource Groups: A concurrent resource group may be shared  
simultaneously by multiple nodes. The resources that can be part of a  
concurrent resource group are limited to volume groups with raw logical  
volumes, raw disks, and application servers.  
When a node fails, there is no takeover involved for concurrent resources.  
Upon reintegration, a node again accesses the resources simultaneously with  
the other nodes.  
The Cluster Manager makes the following assumptions about the acquisition  
of resource groups:  
Cascading  
The active node with the highest priority controls the  
resource group.  
Concurrent  
Rotating  
All active nodes have access to the resource group.  
The node with the rotating resource group’s associated  
service IP address controls the resource group.  
2.4.2 Shared LVM Components  
The first distinction that you need to make while designing a cluster is  
whether you need a non-concurrent or a concurrent shared disk access  
environment.  
2.4.2.1 Non-Concurrent Disk Access Configurations  
The possible non-concurrent disk access configurations are:  
• Hot-Standby  
• Rotating Standby  
• Mutual Takeover  
• Third-Party Takeover  
Hot-Standby Configuration  
Figure 2 illustrates a two node cluster in a hot-standby configuration.  
30  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Figure 2. Hot-Standby Configuration  
In this configuration, there is one cascading resource group consisting of the  
four disks, hdisk1 to hdisk4, and their constituent volume groups and file  
systems. Node 1 has a priority of 1 for this resource group while node 2 has a  
priority of 2. During normal operations, node 1 provides all critical services to  
end users. Node 2 may be idle or may be providing non-critical services, and  
hence is referred to as a hot-standby node. When node 1 fails or has to leave  
the cluster for a scheduled outage, node 2 acquires the resource group and  
starts providing the critical services.  
The advantage of this type of a configuration is that you can shift from a  
single-system environment to an HACMP cluster at a low cost by adding a  
less powerful processor. Of course, this assumes that you are willing to  
accept a lower level of performance in a failover situation. This is a trade-off  
that you will have to make between availability, performance, and cost.  
Rotating Standby Configuration  
This configuration is the same as the previous configuration except that the  
resource groups used are rotating resource groups.  
In the hot-standby configuration, when node 1 reintegrates into the cluster, it  
takes back the resource group since it has the highest priority for it. This  
implies a break in service to the end users during reintegration.  
If the cluster is using rotating resource groups, reintegrating nodes do not  
reacquire any of the resource groups. A failed node that recovers and rejoins  
Cluster Planning 31  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
the cluster becomes a standby node. You must choose a rotating standby  
configuration if you do not want a break in service during reintegration.  
Since takeover nodes continue providing services until they have to leave the  
cluster, you should configure your cluster with nodes of equal power. While  
more expensive in terms of CPU hardware, a rotating standby configuration  
gives you better availability and performance than a hot-standby  
configuration.  
Mutual Takeover Configuration  
Figure 3 illustrates a two node cluster in a mutual takeover configuration.  
Figure 3. Mutual Takeover Configuration  
In this configuration, there are two cascading resource groups: A and B.  
Resource group A consists of two disks, hdisk1 and hdisk3, and one volume  
group, sharedvg. Resource group B consists of two disks, hdisk2 and hdisk4,  
and one volume group, databasevg. Node 1 has priorities of 1 and 2 for  
resource groups A and B respectively, while Node 2 has priorities of 1 and 2  
for resource groups B and A respectively.  
During normal operations, nodes 1 and 2 have control of resource groups A  
and B respectively, and both provide critical services to end users. If either  
node 1 or node 2 fails, or has to leave the cluster for a scheduled outage, the  
surviving node acquires the failed node’s resource groups and continues to  
provide the failed node’s critical services.  
32  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
When a failed node reintegrates into the cluster, it takes back the resource  
group for which it has the highest priority. Therefore, even in this  
configuration, there is a break in service during reintegration. Of course, if  
you look at it from the point of view of performance, this is the best thing to  
do, since you have one node doing the work of two when any one of the  
nodes is down.  
Third-Party Takeover Configuration  
Figure 4 illustrates a three node cluster in a third-party takeover  
configuration.  
Figure 4. Third-Party Takeover Configuration  
This configuration can avoid the performance degradation that results from a  
failover in the mutual takeover configuration.  
Cluster Planning 33  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
Here the resource groups are the same as the ones in the mutual takeover  
configuration. Also, similar to the previous configuration, nodes 1 and 2 each  
have priorities of 1 for one of the resource groups, A or B. The only thing  
different in this configuration is that there is a third node which has a priority  
of 2 for both the resource groups.  
During normal operations, node 3 is either idle or is providing non-critical  
services. In the case of either node 1 or node 2 failing, node 3 takes over the  
failed node’s resource groups and starts providing its services. When a failed  
node rejoins the cluster, it reacquires the resource group for which it has the  
highest priority.  
So, in this configuration, you are protected against the failure of two nodes  
and there is no performance degradation after the failure of one node.  
2.4.2.2 Concurrent Disk Access Configurations  
A concurrent disk access configuration usually has all its disk storage defined  
as part of one concurrent resource group. The nodes associated with a  
concurrent resource group have no priorities assigned to them.  
If a 7135 RAIDiant Array Subsystem is used for storage, you can have a  
maximum of four nodes concurrently accessing a set of storage resources. If  
you are using the 7133 SSA Disk Subsystem, you can have up to eight nodes  
concurrently accessing it.This is because of the physical characteristics of  
SCSI versus SSA.  
In the case of a node failure, a concurrent resource group is not explicitly  
taken over by any other node, since it is already active on the other nodes.  
However, in order to somewhat mask a node failure from the end users, you  
should also have cascading resource groups, each containing the service IP  
address for each node in the cluster. When a node fails, its service IP  
address will be taken over by another node and users can continue to access  
critical services at the same IP address that they were using before the node  
failed.  
2.4.3 IP Address Takeover  
The goal of IP Address Takeover is to make the server’s service address  
highly available and to give clients the possibility of always connecting to the  
same IP address. In order to achieve this, you must do the following:  
• Decide which types of networks and point-to-point connections to use in  
the cluster (see 2.2, “Cluster Networks” on page 11 for supported network  
types)  
34  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
• Design the network topology  
• Define a network mask for your site  
• Define IP addresses (adapter identifiers) for each node’s service and  
standby adapters.  
• Define a boot address for each service adapter that can be taken over, if  
you are using IP address takeover or rotating resources.  
• Define an alternate hardware address for each service adapter that can  
have its IP address taken over, if you are using hardware address  
swapping.  
2.4.3.1 Network Topology  
The following sections cover topics of network topology.  
Single Network  
In a single-network setup, each node in the cluster is connected to only one  
network and has only one service adapter available to clients. In this setup, a  
service adapter on any of the nodes may fail, and a standby adapter will  
acquire its IP address. The network itself, however, is a single point of failure.  
The following figure shows a single-network configuration:  
Client  
Network  
In the single-network setup, each node is connected to one network.  
Each node has one service adapter and can have none, one, or more  
standby adapters per public network.  
Figure 5. Single-Network Setup  
Cluster Planning 35  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
Dual Network  
A dual-network setup has two separate networks for communication. Nodes  
are connected to two networks, and each node has two service adapters  
available to clients. If one network fails, the remaining network can still  
function, connecting nodes and providing resource access to clients.  
In some recovery situations, a node connected to two networks may route  
network packets from one network to another. In normal cluster activity,  
however, each network is separate—both logically and physically.  
Keep in mind that a client, unless it is connected to more than one network, is  
susceptible to network failure.  
The following figure shows a dual-network setup:  
Figure 6. Dual-Network Setup  
Point-to-Point Connection  
A point-to-point connection links two (neighboring) cluster nodes directly.  
SOCC, SLIP, and ATM are point-to-point connection types. In HACMP  
clusters of four or more nodes, however, use an SOCC line only as a private  
network between neighboring nodes because it cannot guarantee cluster  
communications with nodes other than its neighbors.  
36  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
The following diagram shows a cluster consisting of two nodes and a client. A  
single public network connects the nodes and the client, and the nodes are  
linked point-to-point by a private high-speed SOCC connection that provides  
an alternate path for cluster and lock traffic should the public network fail.  
Figure 7. A Point-to-Point Connection  
2.4.3.2 Networks  
Networks in an HACMP cluster are identified by name and attribute.  
Network Name  
The network name is a symbolic value that identifies a network in an HACMP  
for AIX environment. Cluster processes use this information to determine  
which adapters are connected to the same physical network. In most cases,  
the network name is arbitrary, and it must be used consistently. If several  
adapters share the same physical network, make sure that you use the same  
network name when defining these adapters.  
Network Attribute  
A TCP/IP network’s attribute is either public or private.  
Public  
A public network connects from two to 32 nodes and allows clients to  
monitor or access cluster nodes. Ethernet, Token-Ring, FDDI, and  
Cluster Planning 37  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
SLIP are considered public networks. Note that a SLIP line, however,  
does not provide client access.  
Private  
Serial  
A private network provides communication between nodes only; it  
typically does not allow client access. An SOCC line or an ATM  
network are also private networks; however, an ATM network does  
allow client connections and may contain standby adapters. If an  
SP node is used as a client, the SP Switch network, although  
private, can allow client access.  
This network attribute is used for non TCP/IP networks (see 2.2.2,  
2.4.3.3 Network Adapters  
A network adapter (interface) connects a node to a network. A node typically  
is configured with at least two network interfaces for each network to which it  
connects: a service interface that handles cluster traffic, and one or more  
standby interfaces. A service adapter must also have a boot address defined  
for it if IP address takeover is enabled.  
Adapters in an HACMP cluster have a label and a function (service, standby,  
or boot). The maximum number of network interfaces per node is 24.  
Adapter Label  
A network adapter is identified by an adapter label. For TCP/IP networks, the  
adapter label is the name in the /etc/hosts file associated with a specific IP  
address. Thus, a single node can have several adapter labels and IP  
addresses assigned to it. The adapter labels, however, should not be  
confused with the “hostname”, of which there is only one per node.  
Adapter Function  
In the HACMP for AIX environment, each adapter has a specific function that  
indicates the role it performs in the cluster. An adapter’s function can be  
service, standby, or boot.  
Service Adapter The service adapter is the primary connection between  
the node and the network. A node has one service adapter  
for each physical network to which it connects. The  
service adapter is used for general TCP/IP traffic and is  
the address the Cluster Information Program (Clinfo)  
makes known to application programs that want to monitor  
or use cluster services.  
In configurations using rotating resources, the service  
adapter on the standby node remains on its boot address  
38  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
until it assumes the shared IP address. Consequently,  
Clinfo makes known the boot address for this adapter.  
In an HACMP for AIX environment on the RS/6000 SP, the  
SP Ethernet adapters can be configured as service  
adapters but should not be configured for IP address  
takeover. For the SP switch network, service addresses  
used for IP address takeover are ifconfig alias addresses  
used on the css0 network.  
Standby Adapter A standby adapter backs up a service adapter. If a service  
adapter fails, the Cluster Manager swaps the standby  
adapter’s address with the service adapter’s address.  
Using a standby adapter eliminates a network adapter as  
a single point of failure. A node can have no standby  
adapter, or it can have from one to seven standby  
adapters for each network to which it connects. Your  
software configuration and hardware slot constraints  
determine the actual number of standby adapters that a  
node can support.  
The standby adapter is configured on a different subnet  
from any service adapters on the same system, and its  
use should be reserved for HACMP only.  
In an HACMP for AIX environment on the RS/6000 SP, for  
an IP address takeover configuration using the SP switch,  
standby adapters are not required.  
Boot Adapter  
IP address takeover is an AIX facility that allows one node  
to acquire the network address of another node in the  
cluster. To enable IP address takeover, a boot adapter  
label (address) must be assigned to the service adapter  
on each cluster node. Nodes use the boot label after a  
system reboot and before the HACMP for AIX software is  
started.  
In an HACMP for AIX environment on the RS/6000 SP,  
boot addresses used in the IP address for the switch  
network takeover are ifconfig alias addresses used on  
that css0 network.  
When the HACMP for AIX software is started on a node,  
the node’s service adapter is reconfigured to use the  
Cluster Planning 39  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
service label (address) instead of the boot label. If the  
node should fail, a takeover node acquires the failed  
node’s service address on its standby adapter, thus  
making the failure transparent to clients using that specific  
service address.  
During the reintegration of the failed node, which comes  
up on its boot address, the takeover node will release the  
service address it acquired from the failed node.  
Afterwards, the reintegrating node will reconfigure its  
adapter from the boot address to its reacquired service  
address.  
Consider the following scenario: Suppose that Node A  
fails. Node B acquires Node A’s service address and  
services client requests directed to that address. Later,  
when Node A is restarted, it comes up on its boot address  
and attempts to reintegrate into the cluster on its service  
address by requesting that Node B release Node A’s  
service address. When Node B releases the requested  
address, Node A reclaims the address and reintegrates it  
into the cluster. Reintegration, however, fails if Node A has  
not been configured to boot using its boot address.  
The boot address does not use a separate physical  
adapter, but instead is a second name and IP address  
associated with a service adapter. It must be on the same  
subnetwork as the service adapter. All cluster nodes must  
have this entry in the local /etc/hosts file and, if  
applicable, in the nameserver configuration.  
2.4.3.4 Defining Hardware Addresses  
The hardware address swapping facility works in tandem with IP address  
takeover. Hardware address swapping maintains the binding between an IP  
address and a hardware address, which eliminates the need to flush the ARP  
cache of clients after an IP address takeover. This facility, however, is  
supported only for Ethernet, Token-Ring, and FDDI adapters. It does not work  
with the SP Switch or ATM LAN emulation networks.  
Note that hardware address swapping takes about 60 seconds on a  
Token-Ring network, and up to 120 seconds on an FDDI network. These  
periods are longer than the usual time it takes for the Cluster Manager to  
detect a failure and take action.  
40  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
If you do not use Hardware Address Takeover, the ARP cache of clients can  
be updated by adding the clients’ IP addresses to the PING_CLIENT_LIST  
variable in the /usr/sbin/cluster/etc/clinfo.rc file.  
2.4.4 NFS Exports and NFS Mounts  
There are two items concerning NFS when doing the configuration of a  
Resource Group:  
Filesystems to Export  
File systems listed here will be NFS exported,  
so they can be mounted by NFS client  
systems or other nodes in the cluster.  
Filesystems to NFS mount Filling in this field sets up what we call an NFS  
cross mount. Any file system defined in this  
field will be NFS mounted by all the  
participating nodes, other than the node that is  
currently holding the resource group. If the  
node holding the resource group fails, the next  
node to take over breaks its NFS mount for  
this file system, and mounts the file system  
itself as part of its takeover processing.  
2.5 Application Planning  
The central purpose for combining nodes in a cluster is to provide a highly  
available environment for mission-critical applications. These applications  
must remain available at all times in many organizations. For example, an  
HACMP cluster could run a database server program that services client  
applications. The clients send queries to the server program that responds to  
their requests by accessing a database that is stored on a shared external  
disk.  
Planning for these applications requires that you be aware of their location  
within the cluster, and that you provide a solution that enables them to be  
handled correctly, in case a node should fail. In an HACMP for AIX cluster,  
these critical applications can be a single point of failure. To ensure the  
availability of these applications, the node configured to take over the  
resources of the node leaving the cluster should also restart these  
applications so that they remain available to client processes.  
To put the application under HACMP control, you create an application server  
cluster resource that associates a user-defined name with the names of  
specially written scripts to start and stop the application. By defining an  
application server, HACMP for AIX can start another instance of the  
Cluster Planning 41  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
application on the takeover node when a fallover occurs. For more  
information about creating application server resources, see the HACMP for  
AIX, Version 4.3: Installation Guide, SC23-4278.  
2.5.1 Performance Requirements  
In order to plan your application’s needs, you must have a thorough  
understanding of it. One part of that is to have The Application Planning  
Worksheets, found in Appendix A of the HACMP for AIX Planning Guide,  
SC23-4277, filled out.  
Your applications have to be served correctly in an HACMP cluster  
environment. Therefore, you need to know not only how they run on a single  
uni- or multiprocessor machine, but also which resources are required by  
them. How much disk space is required, what is the usual and critical load the  
application puts on a server, and how users access the application are some  
critical factors that will influence your decisions on how to plan the cluster.  
Within an HACMP environment there are always a number of possible states  
in which the cluster could be. Under normal conditions, the load is serviced  
by a cluster node that was designed for this application’s needs. In case of a  
failover, another node has to handle its own work plus the application it is  
going to take over from a failing node. You can even plan one cluster node to  
be the takeover node for multiple nodes; so, when any one of its primary  
nodes fail, it has to take over its application and its load. Therefore, the  
performance requirements of any cluster application have to be understood in  
order to have the computing power available for mission-critical applications  
in all possible cluster states.  
2.5.2 Application Startup and Shutdown Routines  
Highly available applications do not only have to come up at boot time, or  
when someone is starting them up, but also when a critical resource fails and  
has to be taken over by another cluster node. In this case, there have to be  
robust scripts to both start up and shut down the application on the cluster  
nodes. The startup script especially must be able to recover the application  
from an abnormal termination, such as a power failure. You should verify that  
it runs properly in a uniprocessor environment before including the HACMP  
for AIX software.  
42  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
Note  
Application start and stop scripts have to be available on the primary as  
well as the takeover node. They are not transferred during synchronization;  
so, the administrator of a cluster has to ensure that they are found in the  
same path location, with the same permissions and in the same state, i.e.  
changes have to be transferred manually.  
2.5.3 Licensing Methods  
Some vendors require a unique license for each processor that runs an  
application, which means that you must license-protect the application by  
incorporating processor-specific information into the application when it is  
installed. As a result, it is possible that even though the HACMP for AIX  
software processes a node failure correctly, it is unable to restart the  
application on the failover node because of a restriction on the number of  
licenses available within the cluster for that application. To avoid this  
problem, make sure that you have a license for each system unit in the  
cluster that may potentially run an application.  
This can be done by “floating licenses”, where a license server is asked to  
grant the permission to run an application on request, as well as “node-locked  
licenses”, where each processor possibly running an application must have  
the licensing files installed and configured.  
2.5.4 Coexistence with other Applications  
In case of a failover, a node might have to handle several applications  
concurrently. This means the applications data or resources must not conflict  
with each other. Again, the Application Worksheets can help in deciding  
whether certain resources might conflict with others.  
2.5.5 Critical/Non-Critical Prioritizations  
Building a highly available environment for mission-critical applications also  
forces the need to differentiate between the priorities of a number of  
applications. Should a server node fail, it might be appropriate to shut down  
another application, which is not as highly prioritized, in favor of the takeover  
of the server node’s application. The applications running in a cluster have to  
be clearly ordered and prioritized in order to decide what to do under these  
circumstances.  
Cluster Planning 43  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
2.6 Customization Planning  
The Cluster Manager’s ability to recognize a specific series of events and  
subevents permits a very flexible customization scheme. The HACMP for AIX  
software provides an event customization facility that allows you to tailor  
cluster event processing to your site.  
2.6.1 Event Customization  
As part of the planning process, you need to decide whether to customize  
event processing. If the actions taken by the default scripts are sufficient for  
your purposes, you do not need to do anything further to configure events  
during the installation process.  
If you decide to tailor event processing to your environment, it is strongly  
recommended that you use the HACMP for AIX event customization facility  
described in this chapter. If you tailor event processing, you must register  
user-defined scripts with HACMP during the installation process. The HACMP  
for AIX, Version 4.3: Installation Guide, SC23-4278 describes how to  
configure event processing for a cluster.  
You cannot define additional cluster events.  
You can, however, define multiple pre- and post-events for each of the events  
defined in the HACMPevent ODM class.  
The event customization facility includes the following features:  
• Event notification  
• Pre- and post-event processing  
• Event recovery and retry  
2.6.1.1 Special Application Requirements  
Some applications may have some special requirements that have to be  
checked and ensured before or after a cluster event happens. In case of a  
failover you can customize events through the definition of pre- and post-  
events, to act according to your application’s needs. For example, an  
application might want to reset a counter or unlock a user before it can be  
started correctly on the failover node.  
2.6.1.2 Event Notification  
You can specify a notifycommand that sends mail to indicate that an event is  
about to happen (or has just occurred), and that an event script succeeded or  
failed. For example, a site may want to use a network_down notification  
44  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
event to inform system administrators that traffic may have to be rerouted.  
Afterwards, you can use a network_up notification event to inform system  
administrators that traffic can again be serviced through the restored network.  
2.6.1.3 Predictive Event Error Correction  
You can specify a command that attempts to recover from an event script  
failure. If the recovery command succeeds and the retry count for the event  
script is greater than zero, the event script is rerun. You can also specify the  
number of times to attempt to execute the recovery command.  
For example, a recovery command can include the retry of unmounting a file  
system after logging a user off and making sure no one was currently  
accessing the file system.  
If a condition that affects the processing of a given event on a cluster is  
identified, such as a timing issue, you can insert a recovery command with a  
retry count high enough to be sure to cover for the problem.  
2.6.2 Error Notification  
The AIX Error Notification facility detects errors that are logged to the AIX  
error log, such as network and disk adapter failures, and triggers a predefined  
response to the failure. It can even act on application failures, as long as they  
are logged in the error log.  
To implement error notification, you have to add an object to the Error  
Notification object class in the ODM. This object clearly identifies what sort of  
errors you are going to react to, and how.  
By specifying the following in a file:  
errnotify:  
en_name = "Failuresample"  
en_persistenceflg = 0  
en_class = "H"  
en_type = "PERM"  
en_rclass = "disk"  
en_method = "errpt -a -l $1 | mail -s ’Disk Error’ root"  
and adding this to the errnotifyclass through the odmadd <filename>  
command, the specified en_methodis executed every time the error notification  
daemon finds a matching entry in the error report. In the example above, the  
root user will get e-mail identifying the exact error report entry.  
Cluster Planning 45  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
2.6.2.1 Single Point-of-Failure Hardware Component Recovery  
HPS Switch network is one resource that has to be considered as a single  
point of failure. Since a node can support only one switch adapter, its failure  
will disable the switch network for this node. It is strongly recommended to  
promote a failure like this into a node failure, if the switch network is critical to  
your operations.  
Critical failures of the switch adapter would cause an entry in the AIX error  
log. Error labels like HPS_FAULT9_ERor HPS_FAULT3_ERare considered critical,  
and can be specified to AIX Error Notification in order to be able to act upon  
them.  
With HACMP, there is a SMIT screen to make it easier to set up an error  
notification object. This is much easier than the traditional AIX way of adding  
a template file to the ODM class. Under smit hacmp > RAS Support > Error  
Notification > Add a Notify Method, you will find the menu allowing you to  
add these objects to the ODM. An example of the SMIT panel is shown  
below:  
Add a Notify Method  
Type or select values in entry fields.  
Press Enter AFTER making all desired changes.  
[Entry Fields]  
* Notification Object Name  
* Persist across system restart?  
Process ID for use by Notify Method  
Select Error Class  
[HPS_ER9]  
Yes  
[]  
+
+#  
+
All  
Select Error Type  
Match Alertable errors?  
Select Error Label  
PERM  
All  
+
+
[HPS_FAULT9_ER] +  
Resource Name  
Resource Class  
Resource Type  
[All]  
[All]  
[All]  
+
+
+
* Notify Method  
[/usr/sbin/cluster/utilities/clstop -grsy]  
F1=Help  
F5=Reset  
F9=Shell  
F2=Refresh  
F6=Command  
F10=Exit  
F3=Cancel  
F7=Edit  
Enter=Do  
F4=List  
F8=Image  
Figure 8. Sample Screen for Add a Notification Method  
46  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
The above example screen will add a Notification Method to the ODM, so that  
upon appearance of the HPS_FAULT9_ER entry in the error log, the error  
notification daemon will trigger the execution of the  
/usr/sbin/cluster/utilities/clstop -grsy command, which shuts HACMP down  
gracefully with takeover. In this way, the switch failure is acted upon as a  
node failure.  
2.6.2.2 Notification  
The method that is triggered upon the appearance of a specified error log  
entry will be run by the error notification daemon with the command sh -c  
<en_method>. Because this a regular shell, any shell script can act as a  
method.  
So, if you want a specific notification, such as e-mail from this event, you can  
define a script that sends e-mail and then issues the appropriate commands.  
Note  
Because the Notification Method is an object in the node’s ODM, it has to  
be added to each and every node potentially facing a situation where it  
would be wise to act upon the appearance of an error log entry.  
This is NOT handled by the HACMP synchronization facility. You have to  
take care of this manually.  
Alternatively, you can always customize any cluster event to enable a Notify  
Command whenever this event is triggered through the SMIT screen for  
customizing events.  
2.6.2.3 Application Failure  
Even application failures can cause an event to happen, if you have  
configured this correctly. To do so, you have to find some method to decide  
whether an application has failed. This can be as easy as looking for a  
specific process, or much more complex, depending on the application. If you  
issue an Operator Message through the  
errlogger <message>  
command, you can act on that as you would on an error notification as  
Cluster Planning 47  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
2.7 User ID Planning  
The following sections describe various aspects of User ID Planning.  
2.7.1 Cluster User and Group IDs  
One of the basic tasks any system administrator must perform is setting up  
user accounts and groups. All users require accounts to gain access to the  
system. Every user account must belong to a group. Groups provide an  
additional level of security and allow system administrators to manipulate a  
group of users as a single entity.  
For users of an HACMP for AIX cluster, system administrators must create  
duplicate accounts on each cluster node. The user account information  
stored in the /etc/passwd file, and in other files stored in the /etc/security  
directory, should be consistent on all cluster nodes. For example, if a cluster  
node fails, users should be able to log on to the surviving nodes without  
experiencing problems caused by mismatches in the user or group IDs.  
System administrators typically keep user accounts synchronized across  
cluster nodes by copying the key system account and security files to all  
cluster nodes whenever a new account is created or an existing account is  
changed.Typically rdistor rcpis used, for that. On RS/6000 SP systems pcp  
or supperare widely used. For C-SPOC clusters, the C-SPOC utility simplifies  
the cluster-wide synchronization of user accounts by propagating the new  
account or changes to an existing account across all cluster nodes  
automatically.  
The following are some common user and group management tasks, and are  
• Listing all user accounts on all cluster nodes  
• Adding users to all cluster nodes  
• Changing characteristics of a user account on all cluster nodes  
• Removing a user account from all cluster nodes.  
• Listing all groups on all cluster nodes  
• Adding groups to all cluster nodes  
• Changing characteristics of a group on all cluster nodes  
• Removing a group from all cluster nodes  
48  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
2.7.2 Cluster Passwords  
While user and group management is very much facilitated with C-SPOC, the  
password information still has to be distributed by some other means. If the  
system is not configured to use NIS or DCE, the system administrator still has  
to distribute the password information, meaning that found in the  
/etc/security/password file, to all cluster nodes.  
As before, this can be done through rdistor rcp. On RS/6000 SP systems,  
there are tools like pcpor supperto distribute information or better files.  
2.7.3 User Home Directory Planning  
As for user IDs, the system administrator has to ensure that users have their  
home directories available and in the same position at all times. That is, they  
don’t care whether a takeover has taken place or everything is normal. They  
simply want to access their files, wherever they may reside physically, under  
the same directory path with the same permissions, as they would on a single  
machine.  
There are different approaches to that. You could either put them on a shared  
volume and handle them within a resource group, or you could use NFS  
mounts.  
2.7.3.1 Home Directories on Shared Volumes  
Within an HACMP cluster, this approach is quite obvious, however, it restricts  
you to only one machine where a home directory can be active at any given  
time. If you have only one application that the user needs to access, or all of  
the applications are running on one machine, where the second node serves  
as a standby machine only, this would be sufficient.  
2.7.3.2 NFS-Mounted Home Directories  
The NFS mounted home directory approach is much more flexible. Because  
the directory can be mounted on several machines at the same time, a user  
can work with it in several applications on several nodes at the same time.  
However, if one cluster node provides NFS service of home directories to  
other nodes, in case of a failure of the NFS server node, the access to the  
home directories is barred. Placing them onto a machine outside the cluster  
doesn’t help either, since this again introduces a single point of failure, and  
machines outside the cluster are not any less likely to fail than machines  
within.  
Cluster Planning 49  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
2.7.3.3 NFS-Mounted Home Directories on Shared Volumes  
So, a combined approach is used in most cases. In order to make home  
directories a highly available resource, they have to be part of a resource  
group and placed on a shared volume. That way, all cluster nodes can access  
them in case they need to.  
To make the home directories accessible on nodes that currently do not own  
the resource where they are physically residing, they have to be NFS  
exported from the resource group and imported on all the other nodes in case  
any application is running there, needing access to the users files.  
In order to make the directory available to users again, when a failover  
happens, the takeover node that previously had the directory NFS mounted  
from the failed node has to break locks on NFS files, if there are any. Next, it  
must unmount the NFS directory, acquire the shared volume (varyon the  
shared volume group) and mount the shared file system. Only after that can  
users access the application on the takeover node again.  
50  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Chapter 3. Cluster Hardware and Software Preparation  
This chapter covers the steps that are required to prepare the RS/6000  
hardware and AIX software for the installation of HACMP and the  
configuration of the cluster. This includes configuring adapters for TCP/IP,  
setting up shared volume groups, and mirroring and editing AIX configuration  
files.  
3.1 Cluster Node Setup  
The following sections describe important details of cluster node setup.  
3.1.1 Adapter Slot Placement  
For information regarding proper adapter placement, see the following  
documentation:  
PCI Adapter Placement Reference Guide, SA38-0538  
Adapters, Devices, and Cable Information for Micro Channel Bus  
Systems, SA38-0533  
Adapters, Devices, and Cable Information for Multiple Bus Systems,  
SA38-0516  
3.1.2 Rootvg Mirroring  
Of all the components used to build a computer system, physical disk devices  
are usually the most susceptible to failure. Because of this, disk mirroring is a  
frequently used technique for increasing system availability.  
File system mirroring and disk mirroring are easily configured using the AIX  
Logical Volume Manager. However, conventional file system and disk  
mirroring offer no protection against operating system failure or against a  
failure of the disk from which the operating system normally boots.  
Operating system failure does not always occur instantaneously, as  
demonstrated by a system that gradually loses access to operating system  
services. This happens as code and data that were previously being  
accessed from memory gradually disappear in response to normal paging.  
Normally, in an HACMP environment, it is not necessary to think about  
mirroring the root volume group, because the node failure facilities of HACMP  
can cover for the loss of any of the rootvg physical volumes. However, it is  
possible that a customer with business-critical applications will justify  
© Copyright IBM Corp. 1999  
51  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
mirroring rootvg in order to avoid the impact of the failover time involved in a  
node failure. In terms of maximizing availability, this technique is just as valid  
for increasing the availability of a cluster as it is for increasing single-system  
availability.  
The following procedure contains information that will enable you to mirror  
the root volume group (rootvg), using the advanced functions of the Logical  
Volume Manager (LVM). It contains the steps required to:  
• Mirror all the file systems in rootvg.  
• Create an additional boot logical volume (blv).  
• Modify the bootlist to contain all boot devices.  
You may mirror logical volumes in the rootvg in the same way as any AIX  
logical volume may be mirrored, either once (two copies), or twice (three  
copies). The following procedure is designed for mirroring rootvg to a second  
disk only. Upon completion of these steps, your system will remain available if  
one of the disks in rootvg fails, and will even automatically boot from an  
alternate disk drive, if necessary.  
If the dump device is mirrored, you may not be able to capture the dump  
image from a crash or the dump image may be corrupted. The design of LVM  
prevents mirrored writes of the dump device. Only one of the mirrors will  
receive the dump image. Depending on the boot sequence and disk  
availability after a crash, the dump will be in one of the following three states:  
1. Not available  
2. Available and not corrupted  
3. Available and corrupted  
State (1) will always be a possibility. If the user prefers to prevent the risk of  
encountering State (3), then the user must create a non-mirrored logical  
volume (that is not hd6) and set the dump device to this non-mirrored logical  
volume.  
In AIX 4.2.1, two new LVM commands were introduced: mirrorvgand  
unmirrorvg. These two commands where introduced to simplify mirroring or  
unmirroring of the entire contents of a volume group. The commands will  
detect if the entity to be mirrored or unmirrored is rootvg, and will give slightly  
different completion messages based on the type of volume group.The  
mirrorvg command does the equivalent of Procedure steps (2), (3), and (4).  
The mirrorvgcommand takes dump devices and paging devices into account.  
If the dump devices are also the paging device, the logical volume will be  
52  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
mirrored. If the dump devices are NOT the paging device, that dump logical  
volume will not be mirrored.  
3.1.2.1 Procedure  
The following steps assume the user has rootvg contained on hdisk0 and is  
attempting to mirror the rootvg to a new disk: hdisk1.  
1. Extend rootvg to hdisk1 by executing the following:  
extendvg rootvg hdisk1  
2. Disable QUORUM, by executing the following:  
chvg -Qn rootvg  
3. Mirror the logical volumes that make up the AIX operating system by  
executing the following:  
mklvcopy hd1 2 hdisk1 # /home file system  
mklvcopy hd2 2 hdisk1 # /usr file system  
mklvcopy hd3 2 hdisk1 # /tmp file system  
mklvcopy hd4 2 hdisk1 # / (root) file system  
mklvcopy hd5 2 hdisk1 # blv, boot logical volume  
mklvcopy hd6 2 hdisk1 # paging space  
mklvcopy hd8 2 hdisk1 # file system log  
mklvcopy hd9var 2 hdisk1 # /var file system  
If you have other paging devices, rootvg and non-rootvg, it is  
recommended that you also mirror those logical volumes in addition to  
hd6.  
If hd5 consists of more than one logical partition, then, after mirroring hd5  
you must verify that the mirrored copy of hd5 resides on contiguous  
physical partitions. This can be verified with the following command:  
lslv -m hd5  
If the mirrored hd5 partitions are not contiguous, you must delete the  
mirror copy of hd5 (on hdisk1) and rerun the mklvcopyfor hd5, using the  
Cluster Hardware and Software Preparation 53  
Download from Www.Somanuals.com. All Manuals Search And Download.  
“-m” option. You should consult documentation on the usage of the “-m”  
option for mklvcopy.  
4. Synchronize the newly created mirrors with the following command:  
syncvg -v rootvg  
5. Bosboot to initialize all boot records and devices by executing the  
following command:  
bosboot -a -d /dev/hdisk?  
where hdisk?is the first hdisk listed under the “PV” heading after the  
command lslv -l hd5 has executed.  
6. Initialize the boot list by executing the following:  
bootlist -m normal hdisk0 hdisk1  
Note  
Even though this command identifies the list of possible boot disks, it  
does not guarantee that the system will boot from the alternate disk in all  
cases involving failures of the first disk. In such situations, it may be  
necessary for the user to boot from the installation/maintenance media.  
Select maintenance, reissue the bootlistcommand leaving out the  
failing disk, and then reboot. On some models, firmware provides a  
utility for selecting the boot device at boot time. This may also be used  
to force the system to boot from the alternate disk.  
7. Shutdown and reboot the system by executing the following command:  
shutdown -Fr  
This is so that the “Quorum OFF” functionality takes effect.  
54  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
3.1.2.2 Necessary APAR Fixes  
Table 11. Necessary APAR Fixes  
AIX Version  
APARs needed  
4.1  
IX56564  
IX61184  
IX60521  
4.2  
IX62417  
IX68483  
IX70884  
IX72058  
4.3  
IX72550  
To determine if either fix is installed on a machine, execute the following:  
instfix -i -k <apar number>  
3.1.3 AIX Prerequisite LPPs  
In order to install HACMP and HACMP/ES the AIX setup must be in a proper  
state. The following table gives you the prerequisite AIX levels for the  
different HACMP versions:  
Table 12. AIX Prerequisite LPPs  
HACMP Version  
Prerequisite AIX and PSSP Version  
HACMP 4.1 for AIX  
AIX 4.1.5  
PSSP 2.2, if installed on an SP  
HACMP 4.2 for AIX  
HACMP 4.3 for AIX  
HACMP/ES 4.2 for AIX  
HACMP/ES 4.3 for AIX  
AIX 4.1.5  
PSSP 2.2, if installed on an SP  
AIX 4.3.2  
PSSP 2.2, if installed on an SP  
AIX 4.2.1  
PSSP 2.2, if installed on an SP  
AIX 4.3.2  
PSSP 3.1, if installed on an SP  
The Prerequisites for the HACMP component HAView 4.2 are  
• xlC.rte 3.1.3.0  
• nv6000.base.obj 4.1.0.0  
Cluster Hardware and Software Preparation 55  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
• nv6000.database.obj 4.1.0.0  
• nv6000.Features.obj 4.1.2.0  
• nv6000.client.obj 4.1.0.0  
and for HAView 4.3  
• xlC.rte 3.1.4.0  
• nv6000.base.obj 4.1.2.0  
• nv6000.database.obj 4.1.2.0  
• nv6000.Features.obj 4.1.2.0  
• nv6000.client.obj 4.1.2.0  
3.1.4 AIX Parameter Settings  
This section discusses several general tasks necessary to ensure that your  
HACMP for AIX cluster environment works as planned. Consider or check the  
following issues to ensure that AIX works as expected in an HACMP cluster.  
• I/O pacing  
• Network option settings  
• /etc/hosts file and nameserver edits  
• /.rhosts file edits  
3.1.4.1 I/O Pacing  
AIX users have occasionally seen poor interactive performance from some  
applications when another application on the system is doing heavy  
input/output. Under certain conditions, I/O can take several seconds to  
complete. While the heavy I/O is occurring, an interactive process can be  
severely affected if its I/O is blocked, or, if it needs resources held by a  
blocked process.  
Under these conditions, the HACMP for AIX software may be unable to send  
keepalive packets from the affected node. The Cluster Managers on other  
cluster nodes interpret the lack of keepalives as node failure, and the  
I/O-bound node is “failed” by the other nodes. When the I/O finishes, the node  
resumes sending keepalives. Its packets, however, are now out of sync with  
the other nodes, which then kill the I/O-bound node with a RESET packet.  
You can use I/O pacing to tune the system so that system resources are  
distributed more equitably during high disk I/O. You do this by setting high-  
56  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
and low-water marks. If a process tries to write to a file at the high-water  
mark, it must wait until enough I/O operations have finished to make the  
low-water mark.  
Use the smit chgsysfastpath to set high- and low-water marks on the  
Change/Show Characteristics of the Operating System screen.  
By default, AIX is installed with high- and low-water marks set to zero, which  
disables I/O pacing.  
While enabling I/O pacing may have a slight performance effect on very  
I/O-intensive processes, it is required for an HACMP cluster to behave  
correctly during large disk writes. If you anticipate heavy I/O on your HACMP  
cluster, you should enable I/O pacing.  
Although the most efficient high- and low-water marks vary from system to  
system, an initial high-water mark of 33 and a low-water mark of 24 provides  
a good starting point. These settings only slightly reduce write times and  
consistently generate correct fallover behavior from the HACMP for AIX  
software.  
See the AIX Performance Monitoring & Tuning Guide, SC23-2365, for more  
information on I/O pacing.  
3.1.4.2 Checking Network Option Settings  
To ensure that HACMP for AIX requests for memory are handled correctly,  
you can set (on every cluster node) thewall network option to be higher than  
its default value. The suggested value for this option is shown below:  
thewall = 5120  
To change this default value, add the following line to the end of the  
/etc/rc.net file:  
no -o thewall=5120  
After making this change, monitor mbuf usage using the netstat -m command  
and increase or decrease “thewall” option as needed.  
To list the values of other network options (not configurable) that are currently  
set on a node, enter:  
no -a  
Cluster Hardware and Software Preparation 57  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
3.1.4.3 Editing the /etc/hosts File and Nameserver Configuration  
Make sure all nodes can resolve all cluster addresses. See the chapter on  
planning TCP/IP networks (the section Using HACMP with NIS and DNS) in  
the HACMP for AIX, Version 4.3: Planning Guide, SC23-4277 for more  
information on name serving and HACMP.  
Edit the /etc/hosts file (and the /etc/resolv.conf file, if using the nameserver  
configuration) on each node in the cluster to make sure the IP addresses of  
all clustered interfaces are listed.  
For each boot address, make an entry similar to the following:  
100.100.50.200 crab_boot  
Also, make sure that the /etc/hosts file on each node has the following entry:  
127.0.0.1 loopback localhost  
3.1.4.4 cron and NIS Considerations  
If your HACMP cluster nodes use NIS services, which include the mapping of  
the /etc/passwd file, and IPAT is enabled, users that are known only in the  
NIS-managed version of the /etc/passwd file will not be able to create  
crontabs. This is because cron is started with the /etc/inittab file with run level  
2 (for example, when the system is booted), but ypbind is started in the  
course of starting HACMP with the rcnfsentry in /etc/inittab. When IPAT is  
enabled in HACMP, the run level of the rcnfsentry is changed to -aand run  
with the telinit -acommand by HACMP.  
In order to let those NIS-managed users create crontabs, you can do one of  
the following:  
• Change the run level of the cronentry in /etc/inittabto -aand make sure  
it is positioned after the rcnfsentry in /etc/inittab. This solution is  
recommended if it is acceptable to start cron after HACMP has started.  
• Add an entry to the /etc/inittab file like the following script with run level -a.  
Make sure it is positioned after the rcnfsentry in /etc/inittab. The important  
thing is to kill the cron process, which will respawn and know about all of  
the NIS-managed users. Whether or not you log the fact that cron has  
been refreshed is optional.  
58  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
#! /bin/sh  
# This script checks for a ypbind and a cron process. If both  
# exist and cron was started before ypbind, cron is killed so  
# it will respawn and know about any new users that are found  
# in the passwd file managed as an NIS map.  
echo "Entering $0 at ‘date‘" >> /tmp/refr_cron.out  
cronPid=‘ps -ef |grep "/etc/cron" |grep -v grep |awk \  
’{ print $2 }’‘  
ypbindPid=‘ps -ef | grep "/usr/etc/ypbind" | grep -v grep | \  
if [ ! -z "${ypbindPid}" ]  
then  
if [ ! -z "${cronPid}" ]  
then  
echo "ypbind pid is ${ypbindPid}" >> /tmp/refr_cron.out  
echo "cron pid is ${cronPid}" >> /tmp/refr_cron.out  
echo "Killing cron(pid ${cronPid}) to refresh user \  
list" >> /tmp/refr_cron.out  
kill -9 ${cronPid}  
if [ $? -ne 0 ]  
then  
echo "$PROGNAME: Unable to refresh cron." \  
>>/tmp/refr_cron.out  
exit 1  
fi  
fi  
fi  
echo "Exiting $0 at ‘date‘" >> /tmp/refr_cron.out  
exit 0  
3.1.4.5 Editing the /.rhosts File  
Make sure that each node’s service adapters and boot addresses are listed in  
the /.rhosts file on each cluster node. Doing so allows the  
/usr/sbin/cluster/utilities/clruncmdcommand and the  
/usr/sbin/cluster/godm daemon to run. The /usr/sbin/cluster/godm daemon is  
used when nodes are configured from a central location.  
For security reasons, IP label entries that you add to the /.rhosts file to  
identify cluster nodes should be deleted when you no longer need to log on to  
a remote node from these nodes. The cluster synchronization and verification  
functions use rcmdand rshand thus require these /.rhostsentries. These  
entries are also required to use C-SPOC commands in a cluster environment.  
The /usr/sbin/cluster/clstrmgr daemon, however, does not depend on /.rhosts  
file entries.  
The /.rhosts file is not required on SP systems running the HACMP Enhanced  
Security. This feature removes the requirement of TCP/IP access control lists  
(for example, the /.rhosts file) on remote nodes during HACMP configuration.  
Cluster Hardware and Software Preparation 59  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
3.2 Network Connection and Testing  
The following sections describe important aspects of network connection and  
testing.  
3.2.1 TCP/IP Networks  
Since there are several types of TCP/IP Networks available within HACMP,  
there are several different characteristics and some restrictions on them.  
Characteristics like maximum distance between nodes have to be  
considered. You don’t want to put two cluster nodes running a mission-critical  
application in the same room for example.  
3.2.1.1 Cabling Considerations  
Characteristics of the different types of cable, their maximum length, and the  
like are beyond the scope of this book. However, for actual planning of your  
clusters, you have to check whether your network cabling allows you to put  
two cluster nodes away from each other, or even in different buildings.  
There’s one additional point with cabling, that should be taken care of.  
Cabling of networks often involves hubs or switches. If not carefully planned,  
this sometimes introduces another single point of failure into your cluster. To  
eliminate this you should have at least two hubs.  
As shown in Figure 9, failure of a hub would not result in one machine being  
disconnected from the network. In that case, a hub failure would cause either  
both service adapters to fail, which would cause a swap_adapter event, and  
the standby adapters would take over the network, or both standby adapters  
would fail, which would cause fail_standby events. Configuring a notify  
method for these events can alert the network administrator to check and fix  
the broken hub.  
60  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
.
Figure 9. Connecting Networks to a Hub  
3.2.1.2 IP Addresses and Subnets  
The design of the HACMP for AIX software specifies that:  
• All client traffic be carried over the service adapter  
• Standby adapters be hidden from client applications and carry only  
internal Cluster Manager traffic  
Cluster Hardware and Software Preparation 61  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
To comply with these rules, pay careful attention to the IP addresses you  
assign to standby adapters. Standby adapters must be on a separate subnet  
from the service adapters, even though they are on the same physical  
network. Placing standby adapters on a different subnet from the service  
adapter allows HACMP for AIX to determine which adapter TCP/IP will use to  
send a packet to a network.  
If there is more than one adapter with the same network address, there is no  
way to guarantee which of these adapters will be chosen by IP as the  
transmission route. All choices will be correct, since each choice will deliver  
the packet to the correct network. To guarantee that only the service adapter  
handles critical traffic, you must limit IP’s choice of a transmission route to  
one adapter. This keeps all traffic off the standby adapter so that it is  
available for adapter swapping and IP address takeover (IPAT). Limiting the  
IP’s choice of a transmission route also facilitates identifying an adapter  
failure.  
Note  
The netmask for all adapters in an HACMP network must be the same even  
though the service and standby adapters are on different logical subnets.  
See the HACMP for AIX, Version 4.3: Concepts and Facilities, SC23-4276  
guide for more information about using the same netmask for all adapters.  
information.  
3.2.1.3 Testing  
After setting up all adapters with AIX, you can do several things to check  
whether TCP/IP is working correctly. Note, that without HACMP being started,  
the service adapters defined to HACMP will remain on their boot address.  
After startup these adapters change to their service addresses.  
Use the following AIX commands to investigate the TCP/IP subsystem:  
• Use the netstatcommand to make sure that the adapters are initialized  
and that a communication path exists between the local node and the  
target node.  
• Use the pingcommand to check the point-to-point connectivity between  
nodes.  
• Use the ifconfigcommand on all adapters to detect bad IP addresses,  
incorrect subnet masks, and improper broadcast addresses.  
62  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
• Scan the /tmp/hacmp.out file to confirm that the /etc/rc.net script has run  
successfully. Look for a zero exit status.  
• If IP address takeover is enabled, confirm that the /etc/rc.net script has  
run and that the service adapter is on its service address and not on its  
boot address.  
• Use the lssrc -g tcpipcommand to make sure that the inetd daemon is  
running.  
• Use the lssrc -g portmapcommand to make sure that the portmapper  
daemon is running.  
• Use the arpcommand to make sure that the cluster nodes are not using  
the same IP or hardware address.  
3.2.2 Non TCP/IP Networks  
Currently three types of non-TCP/IP networks are supported:  
• Serial (RS232)  
Target-mode SCSI  
Target-mode SSA  
While we use the word serial here to refer to RS232 only, in HACMP  
definitions, a “serial” network means a non-TCP/IP network of any kind.  
Therefore, when we are talking about HACMP network definitions, a serial  
network could also be a target-mode SCSI or target-mode SSA network.  
The following describes some cabling issues on each type of non-TCP/IP  
network, how they are to be configured, and how you can test if they are  
operational.  
3.2.2.1 Cabling Considerations  
RS232  
Cabling a serial connection requires a null-modem cable. As often  
cluster nodes are further apart than 60 m (181 ft.), sometimes  
modem eliminators or converters to fiber channel are used.  
TMSCSI If your cluster uses SCSI disks as shared devices, you can use  
that line for TMSCSI as well. TMSCSI requires Differential SCSI  
page 26). Because the SCSI bus has to be terminated on both  
ends, and not anywhere else in between, resistors on the  
adapters should be removed, and cabling should be done as  
shown in Figure 11 on page 77, that is, with Y-cables that are  
terminated at one end connected to the adapters where the other  
end connects to the shared disk device.  
Cluster Hardware and Software Preparation 63  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
TMSSA  
Target-mode SSA is only supported with the SSA Multi-Initiator  
RAID Adapters (Feature #6215 and #6219), Microcode Level  
1801 or later. You need at least HACMP Version 4.2.2 with APAR  
IX75718.  
3.2.2.2 Configuring RS232  
Use the smit ttyfastpath to create a tty device on the nodes. On the resulting  
panel, you can add an RS232 tty by selecting a native serial port, or a port on  
an asynchronous adapter. Make sure that the Enable Login field is set to  
disable. You do not want a getty process being spawned on this interface.  
3.2.2.3 Configuring Target Mode SCSI  
To configure a target-mode SCSI network on the Differential SCSI adapters,  
you have to enable the SCSI adapter’s feature TARGET MODE by setting the  
enabled characteristics to yes. Since disks on the SCSI bus are normally  
configured at boot time, and the characteristics of the parent device cannot  
be changed as long as there are child devices present and active, you have  
to set all the disks on that bus to Definedwith the  
rmdev -l hdiskx  
command, before you can enable that feature. Alternatively you can make  
these changes to the database (ODM) only, and they will be activated at the  
time of the next reboot.  
If you choose not to reboot, instead setting all the child devices to Defined,  
you have to run cfgmgr, to get the tmscsi device created, as well as all the  
child devices of the adapter back to the available state.  
Note  
The target mode device created is a logical new device on the bus.  
Because it is created by scanning the bus for possible initiator devices, a  
tmscsix device is created on a node for each SCSI adapter on the same  
bus that has the target mode flag enabled, therefore representing this  
adapter’s unique SCSI ID. In that way, the initiator can address packets to  
exactly one target device.  
This procedure has to be done for all the cluster nodes that are going to use a  
serial network of type tmscsi as defined in your planning sheets.  
64  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
3.2.2.4 Configuring Target Mode SSA  
The node number on each system needs to be changed from the default of  
zero to a number. All systems on the SSA loop must have a unique node  
number.  
To change the node number use the following command:  
chdev -l ssar -a node_number=#  
To show the system’s node number use the following command:  
lsattr -El ssar  
Having the node numbers set to non-zero values enables the target mode  
devices to be configured. Run the cfgmgrcommand to configure the tmssa#  
devices on each system. Check that the tmssa devices are available on each  
system using the following command:  
lsdev -C | grep tmssa  
The Target Mode SCSI or SSA serial network can now be configured into an  
HACMP cluster.  
3.2.2.5 Testing RS232 and Target Mode Networks  
Testing of the serial networks functionality is similar. Basically you just write  
to one side’s device and read from the other.  
Serial (RS323): After configuring the serial adapter and cabling it correctly,  
you can check the functionality of the connection by entering the command  
cat < /dev/ttyx  
on one node for reading from that device and  
cat /etc/environment > /dev/ttyy  
on the corresponding node for writing. You should see the first command  
hanging until the second command is issued, and then showing the output of  
it.  
Target Mode SSA: After configuration of Target Mode SSA, you can check  
the functionality of the connection by entering the command:  
cat < /dev/tmssax.tm  
on one node for reading from that device and:  
Cluster Hardware and Software Preparation 65  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
cat /etc/environment > /dev/tmssay.im  
on the corresponding node for writing. x and y correspond to the appropriate  
opposite nodenumber. You should see the first command hanging until the  
second command is issued, and then showing its output.  
Target Mode SCSI: After configuration of Target Mode SCSI, you can check  
the functionality of the connection by entering the command:  
cat < /dev/tmscsix.tm  
on one node for reading from that device and:  
cat /etc/environment > /dev/tmscsiy.im  
on the corresponding node for writing. You should see the first command  
hanging until the second command is issued, and then showing the output of  
that second command.  
3.3 Cluster Disk Setup  
The following sections relate important information about cluster disk setup.  
3.3.1 SSA  
The following sections describe cabling, AIX configuration, microcode  
loading, and configuring a RAID on SSA disks.  
3.3.1.1 Cabling  
The following rules must be followed when connecting a 7133 SSA  
Subsystem:  
• Each SSA loop must be connected to a valid pair of connectors on the  
SSA adapter card (A1 and A2 to form one loop, or B1 and B2 to form one  
loop).  
• Only one pair of connectors of an SSA adapter can be connected in a  
particular SSA loop (A1 or A2, with B1 or B2 cannot be in the same SSA  
loop).  
• A maximum of 48 disks can be connected in an SSA loop.  
• A maximum of three dummy disk drive modules can be connected next to  
each other.  
• The maximum length of an SSA cable is 25 m. With Fiber-Optic  
Extenders, the connection length can be up to 2.4 km.  
66  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
For more information regarding adapters and cabling rules see 2.3.1, “SSA  
Disks” on page 16 or the following documents:  
• 7133 SSA Disk Subsystems: Service Guide, SY33-0185-02  
• 7133 SSA Disk Subsystem: Operator Guide, GA33-3259-01  
• 7133 Models 010 and 020 SSA Disk Subsystems: Installation Guide,  
GA33-3260-02  
• 7133 Models 500 and 600 SSA Disk Subsystems: Installation Guide,  
GA33-3263-02  
• 7133 SSA Disk Subsystems for Open Attachment: Service Guide,  
SY33-0191-00  
• 7133 SSA Disk Subsystems for Open Attachment: Installation and User's  
Guide, SA33-3273-00  
3.3.1.2 AIX Configuration  
During boot time, the configuration manager of AIX configures all the device  
drivers needed to have the SSA disks available for usage. The configuration  
manager can’t do this configuration if the SSA Subsystem is not properly  
connected or if the SSA Software is not installed. If the SSA Software is not  
already installed, the configuration manager will tell you the missing filesets.  
You can either install the missing filesets with smit, or call the configuration  
manager with the -i flag.  
The configuration manager configures the following devices:  
• SSA Adapter Router  
• SSA Adapter  
• SSA Disks  
Adapter Router  
The adapter Router (ssar) is only a conceptual configuration aid and is  
always in a “Defined” state. It cannot be made “Available.” You can list the  
ssar with the following command:  
#lsdev -C | grep ssar  
ssar  
Defined  
SSA Adapter Router  
Cluster Hardware and Software Preparation 67  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
Adapter Definitions  
By issuing the following command, you can check the correct adapter  
configuration. In order to work correctly, the adapter must be in the  
“Available” state:  
#lsdev -C | grep ssa  
ssa0  
ssar  
Available 00-07  
Defined  
SSA Enhanced Adapter  
SSA Adapter Router  
The third column in the adapter device line shows the location of the adapter.  
Disk Definitions  
SSA disk drives are represented in AIX as SSA logical disks (hdisk0,  
hdisk1,...,hdiskN) and SSA physical disks (pdisk0, pdisk1,...,pdiskN). SSA  
RAID arrays are represented as SSA logical disks (hdisk0, hdisk1,...,hdiskN).  
SSA logical disks represent the logical properties of the disk drive or array,  
and can have volume groups and file systems mounted on them. SSA  
physical disks represent the physical properties of the disk drive. By default,  
one pdisk is always configured for each physical disk drive. One hdisk is  
configured for each disk drive that is connected to the using system, or for  
each array. By default, all disk drives are configured as system (AIX) disk  
drives. The array management software can be used to change the disks  
from hdisks to array candidate disks or hot spares.  
SSA logical disks:  
• Are configured as hdisk0, hdisk1,...,hdiskN.  
• Support a character special file (/dev/rhdisk0,  
/dev/rhdisk1,...,/dev/rhdiskN).  
• Support a block special file (/dev/hdisk0, /dev/hdisk1,...,/dev/hdiskN).  
• Support the I/O Control (IOCTL) subroutine call for non service and  
diagnostic functions only.  
• Accept the read and write subroutine calls to the special files.  
• Can be members of volume groups and have file systems mounted on  
them.  
In order to list the logical disk definitions, use the following command:  
68  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
#lsdev -Cc disk| grep SSA  
hdisk3 Available 00-07-L  
hdisk4 Available 00-07-L  
hdisk5 Available 00-07-L  
hdisk6 Available 00-07-L  
hdisk7 Available 00-07-L  
hdisk8 Available 00-07-L  
SSA Logical Disk Drive  
SSA Logical Disk Drive  
SSA Logical Disk Drive  
SSA Logical Disk Drive  
SSA Logical Disk Drive  
SSA Logical Disk Drive  
SSA physical disks:  
• Are configured as pdisk0, pdisk1,...,pdiskN.  
• Have errors logged against them in the system error log.  
• Support a character special file (/dev/pdisk0, /dev/pdisk1,...,/dev/p.diskN).  
• Support the IOCTLl subroutine for servicing and diagnostic functions.  
• Do not accept read or write subroutine calls for the character special file.  
In order to list the physical disk definitions use the following command:  
#lsdev -Cc pdisk| grep SSA  
pdisk0 Available 00-07-P 1GB SSA C Physical Disk Drive  
pdisk1 Available 00-07-P 1GB SSA C Physical Disk Drive  
pdisk2 Available 00-07-P 1GB SSA C Physical Disk Drive  
pdisk3 Available 00-07-P 1GB SSA C Physical Disk Drive  
pdisk4 Available 00-07-P 1GB SSA C Physical Disk Drive  
pdisk5 Available 00-07-P 1GB SSA C Physical Disk Drive  
Diagnostics  
A good tool to get rid of SSA problems are the SSA service aids in the AIX  
diagnostic program diag.The SSA diagnostic routines are fully documented  
in A Practical Guide to SSA for AIX, SG24-4599. The following is a brief  
overview:  
The SSA service aids are accessed from the main menu of the diag program.  
Select Task Selection -> SSA Service Aids. This will give you the following  
options:  
Set Service Mode  
This option enables you to determine the location  
of a specific SSA disk drive within a loop and to  
remove the drive from the configuration, if  
required.  
Link Verification  
This option enables you to determine the  
operational status of a link  
Cluster Hardware and Software Preparation 69  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Configuration Verification This option enables you to display the  
relationships between physical (pdisk) and logical  
(hdisk) disks.  
Format Disk  
Certify Disk  
This option enables you to format SSA disk drives.  
This option enables you to test whether data on an  
SSA disk drive can be read correctly.  
Display/Download...  
This option enables you to display the microcode  
level of the SSA disk drives and to download new  
microcode to individual or all SSA disk drives  
connected to the system.  
Note  
When an SSA loop is attached to multiple host systems, do not invoke the  
diagnostic routines from more than one host simultaneously, to avoid  
unpredictable results that may result in data corruption.  
3.3.1.3 Microcode Loading  
To ensure that everything works correctly, install the latest filesets, fixes and  
microcode for your SSA disk subsystem. The latest information and  
downloadable files can be found under http://www.hursley.ibm.com/~ssa.  
Upgrade Instructions  
Follow these steps to perform an upgrade:  
1. Login as root  
2. Download the appropriate microcode file for your AIX version from the  
web-site mentioned above  
3. Save the file upgrade.tar in your /tmp directory  
4. Type tar -xvf upgrade.tar  
5. Run smitty install  
6. Select install & update software  
7. Select install & update from ALL available software  
8. Use the directory /usr/sys/inst.images as the install device  
9. Select all filesets in this directory for install  
10.Execute the command  
11.Exit Smit  
70  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Note  
You must ensure that:  
• You do not attempt to perform this adapter microcode download  
concurrently on systems that are in the same SSA loop. This may cause  
a portion of the loop to be isolated and could prevent access to these  
disks from elsewhere in the loop.  
• You do not run advanced diagnostics while downloads are in progress.  
Advanced diagnostics causes the SSA adapter to be reset temporarily,  
thereby introducing a break in the loop, portions of the loop may  
become temporarily isolated and inaccessible.  
• You have complete SSA loops. Check this by using diagnostics in  
System Verification mode. If you have incomplete loops (such as  
strings) action must be taken to resolve this before you can continue.  
• All of your loops are valid, in this case with one or two adapters in each  
loop. This is also done by using Diagnostics in System Verification  
mode.  
12.Run cfgmgr to install the microcode to adapters.  
13.To complete the device driver upgrade, you must now reboot your system.  
14.To confirm that the upgrade was a success, type lscfg -vl ssaXwhere X is  
0,1... for all SSA adapters. Check the ROS Level line to see that each  
adapter has the appropriate microcode level (for the correct microcode  
level, see the above mentioned web-site).  
15.Run lslpp -l|grep SSAand check that the fileset levels you have match, or  
are above the levels shown in the list on the above mentioned web-site. If  
any of the SSA filesets are at a lower level than those shown in the above  
link, please repeat the whole upgrade procedure again. If, after repeating  
the procedure, the code levels do not match the latest ones, place a call  
with your local IBM Service Center.  
16.If the adapters are in SSA loops which contain other adapters in other  
systems, please repeat this procedure on all systems as soon as possible.  
17.In order to install the disk microcode, run ssadload -ufrom each system in  
turn.  
Note  
Allow ssadload to complete on one system before running it on another.  
Cluster Hardware and Software Preparation 71  
Download from Www.Somanuals.com. All Manuals Search And Download.  
18.To confirm that the upgrade was a success, type lscfg -vl pdiskXwhere X  
is 0,1... for all SSA disks. Check the ROS Level line to see that each disk  
has the appropriate microcode level (for the correct microcode level see  
the above mentioned web-site).  
3.3.1.4 Configuring a RAID on SSA Disks  
Disk arrays are groups of disk drives that act like one disk as far as the  
operating system is concerned, and which provide better availability or  
performance characteristics than the individual drives operating alone.  
Depending on the particular type of array that is used, it is possible to  
optimize availability or performance, or to select a compromise between both.  
The SSA Enhanced Raid adapters only support RAID level 5 (RAID5). RAID0  
(Striping) and RAID1 (Mirroring) is not directly supported by the SSA  
Enhanced Raid adapters, but with the Logical Volume Manager (LVM), RAID0  
and RAID1 can be configured on non-RAID disks.  
In order to create a RAID5 on SSA Disks, use the command smitty ssaraid.  
This will show you the following menu:  
SSA RAID Arrays  
Move cursor to desired item and press Enter.  
List All Defined SSA RAID Arrays  
List All Supported SSA RAID Arrays  
List All SSA RAID Arrays Connected to a RAID Manager  
List Status Of All Defined SSA RAID Arrays  
List/Identify SSA Physical Disks  
List/Delete Old RAID Arrays Recorded in an SSA RAID Manager  
Add an SSA RAID Array  
Delete an SSA RAID Array  
Change/Show Attributes of an SSA RAID Array  
Change Member Disks in an SSA RAID Array  
Change/Show Use of an SSA Physical Disk  
Change Use of Multiple SSA Physical Disks  
F1=Help  
F9=Shell  
F2=Refresh  
F10=Exit  
F3=Cancel  
Enter=Do  
F8=Image  
Select Add an SSA RAID Array to do the definitions.  
3.3.2 SCSI  
The following sections contain important information about SCSI: cabling,  
connecting RAID subsystems, and adapter SCSI ID and termination change.  
72  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
3.3.2.1 Cabling  
The following sections describe important information about cabling.  
SCSI Adapters  
A overview of SCSI adapters that can be used on a shared SCSI bus is given  
RAID Enclosures  
The 7135 RAIDiant Array can hold a maximum of 30 single-ended disks in  
two units (one base and one expansion). It has one controller by default, and  
another controller can be added for improved performance and availability.  
Each controller takes up one SCSI ID. The disks sit on internal single-ended  
buses and hence do not take up IDs on the external bus. In an HACMP  
cluster, each 7135 should have two controllers, each of which is connected to  
a separate shared SCSI bus. This configuration protects you against any  
failure (SCSI adapter, cables, or RAID controller) on either SCSI bus.  
Because of cable length restrictions, a maximum of two 7135s on a shared  
SCSI bus are supported by HACMP.  
3.3.2.2 Connecting RAID Subsystems  
In this section, we will list the different components required to connect RAID  
subsystems on a shared bus. We will also show you how to connect these  
components together.  
The 7135-110 RAIDiant Array can be connected to multiple systems on either  
an 8-bit or a 16-bit SCSI-2 differential bus. The Model 210 can only be  
connected to a 16-bit SCSI-2 Fast/Wide differential bus, using the Enhanced  
SCSI-2 Differential Fast/Wide Adapter/A.  
To connect a set of 7135-110s to SCSI-2 Differential Controllers on a shared  
8-bit SCSI bus, you need the following:  
• SCSI-2 Differential Y-Cable  
FC: 2422 (0.765m), PN: 52G7348  
• SCSI-2 Differential System-to-System Cable  
FC: 2423 (2.5m), PN: 52G7349  
This cable is used only if there are more than two nodes attached to the  
same shared bus.  
• Differential SCSI Cable (RAID Cable)  
FC: 2901 or 9201 (0.6m), PN: 67G1259 - OR -  
Cluster Hardware and Software Preparation 73  
Download from Www.Somanuals.com. All Manuals Search And Download.  
FC: 2902 or 9202 (2.4m), PN: 67G1260 - OR -  
FC: 2905 or 9205 (4.5m), PN: 67G1261 - OR -  
FC: 2912 or 9212 (12m), PN: 67G1262 - OR -  
FC: 2914 or 9214 (14m), PN: 67G1263 - OR -  
FC: 2918 or 9218 (18m), PN: 67G1264  
Terminator (T)  
Included in FC 2422 (Y-Cable), PN: 52G7350  
• Cable Interposer (I)  
FC: 2919, PN: 61G8323  
One of these is required for each connection between an SCSI-2  
Differential Y-Cable and a Differential SCSI Cable going to the 7135  
unit, as shown in Figure 10.  
Figure 10 shows four RS/6000s, each represented by two SCSI-2 Differential  
Controllers, connected on two 8-bit buses to two 7135-110s, each with two  
controllers.  
Figure 10. 7135-110 RAIDiant Arrays Connected on Two Shared 8-Bit SCSI Buses  
To connect a set of 7135s to SCSI-2 Differential Fast/Wide Adapter/As or  
Enhanced SCSI-2 Differential Fast/Wide Adapter/As on a shared 16-bit SCSI  
bus, you need the following:  
• 16-Bit SCSI-2 Differential Y-Cable  
74  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
FC: 2426 (0.94m), PN: 52G4234  
• 16-Bit SCSI-2 Differential System-to-System Cable  
FC: 2424 (0.6m), PN: 52G4291 - OR -  
FC: 2425 (2.5m), PN: 52G4233  
This cable is used only if there are more than two nodes attached to the  
same shared bus.  
• 16-Bit Differential SCSI Cable (RAID Cable)  
FC: 2901 or 9201 (0.6m), PN: 67G1259 - OR -  
FC: 2902 or 9202 (2.4m), PN: 67G1260 - OR -  
FC: 2905 or 9205 (4.5m), PN: 67G1261 - OR -  
FC: 2912 or 9212 (12m), PN: 67G1262 - OR -  
FC: 2914 or 9214 (14m), PN: 67G1263 - OR -  
FC: 2918 or 9218 (18m), PN: 67G1264  
• 16-Bit Terminator (T)  
Included in FC 2426 (Y-Cable), PN: 61G8324  
Figure 11 shows four RS/6000s, each represented by two SCSI-2 Differential  
Fast/Wide Adapter/As connected on two 16-bit buses to two 7135-110s, each  
with two controllers.  
The 7135-210 requires the Enhanced SCSI-2 Differential Fast/Wide  
Adapter/A adapter for connection. Other than that, the cabling is exactly the  
same as shown in Figure 11, if you just substitute the Enhanced SCSI-2  
Differential Fast/Wide Adapter/A (FC: 2412) for the SCSI-2 Differential  
Fast/Wide Adapter/A (FC: 2416) in the picture.  
Cluster Hardware and Software Preparation 75  
Download from Www.Somanuals.com. All Manuals Search And Download.  
#2416 (16-  
6 (16-bit)  
#2424  
#2426  
#2426  
#2416 (16-b  
6-bit)  
T
T
#2416 (16-bit)  
6 (16-bit )  
T
T
Maximum total cable length: 25m  
76  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
7135-110  
7135-110  
#2902  
RAID  
cable  
Controller 1  
Controller 1  
#2902  
#2902  
#2902  
#2426  
#2416 (16-bit)  
#2902  
RAID  
cable  
#2426  
Controller 2  
Controller 2  
#2902  
#2416 (16-bit)  
#2416 (16-bit)  
#2416 (16-bit)  
#2424  
#2426  
#2426  
#2416 (16-bit)  
#2416 (16-bit)  
#2416 (16-bit )  
T
T
#2416 (16-bit)  
T
T
Maximum total cable length: 25m  
Figure 11. 7135-110 RAIDiant Arrays Connected on Two Shared 16-Bit SCSI Buses  
3.3.2.3 Adapter SCSI ID and Termination change  
The SCSI-2 Differential Controller is used to connect to 8-bit disk devices on  
a shared bus. The SCSI-2 Differential Fast/Wide Adapter/A or Enhanced  
SCSI-2 Differential Fast/Wide Adapter/A is usually used to connect to 16-bit  
devices but can also be used with 8-bit devices.  
In a dual head-of-chain configuration of shared disks, there should be no  
termination anywhere on the bus except at the extremities. Therefore, you  
should remove the termination resistor blocks from the SCSI-2 Differential  
Controller and the SCSI-2 Differential Fast/Wide Adapter/A or Enhanced  
SCSI-2 Differential Fast/Wide Adapter/A. The positions of these blocks (U8  
and U26 on the SCSI-2 Differential Controller, and RN1, RN2 and RN3 on the  
Cluster Hardware and Software Preparation 77  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
SCSI-2 Differential Fast/Wide Adapter/A and Enhanced SCSI-2 Differential  
Fast/Wide Adapter/A) are shown in Figure 12 and Figure 13 respectively.  
Termination  
Resistor  
Blocks  
P/N 43G0176  
4-2  
Figure 12. Termination on the SCSI-2 Differential Controller  
Termination  
Resistor  
Internal 16-bit SE Internal 8-bit SE Blocks  
P/N 56G7315  
4-6  
Figure 13. Termination on the SCSI-2 Differential Fast/Wide Adapters  
78  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
The ID of an SCSI adapter, by default, is 7. Since each device on an SCSI  
bus must have a unique ID, the ID of at least one of the adapters on a shared  
SCSI bus has to be changed.  
The procedure to change the ID of an SCSI-2 Differential Controller is:  
1. At the command prompt, enter smit chgscsi.  
2. Select the adapter whose ID you want to change from the list presented to  
you.  
SCSI Adapter  
Move cursor to desired item and press Enter.  
scsi0 Available 00-02 SCSI I/O Controller  
scsi1 Available 06-02 SCSI I/O Controller  
scsi2 Available 08-02 SCSI I/O Controller  
scsi3 Available 07-02 SCSI I/O Controller  
F1=Help  
F8=Image  
/=Find  
F2=Refresh  
F10=Exit  
n=Find Next  
F3=Cancel  
Enter=Do  
3. Enter the new ID (any integer from 0 to 7) for this adapter in the Adapter  
card SCSI ID field. Since the device with the highest SCSI ID on a bus  
gets control of the bus, set the adapter’s ID to the highest available ID. Set  
the Apply change to DATABASE only field to yes.  
Cluster Hardware and Software Preparation 79  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Change / Show Characteristics of a SCSI Adapter  
Type or select values in entry fields.  
Press Enter AFTER making all desired changes.  
[Entry Fields]  
SCSI Adapter  
scsi0  
Description  
Status  
Location  
SCSI I/O Controller  
Available  
00-08  
Adapter card SCSI ID  
BATTERY backed adapter  
DMA bus memory LENGTH  
Enable TARGET MODE interface  
Target Mode interface enabled  
PERCENTAGE of bus memory DMA area  
for target mode  
[6]  
no  
[0x202000]  
no  
no  
+#  
+
+
+
[50]  
+#  
Name of adapter code download file  
Apply change to DATABASE only  
/etc/microcode/8d>  
yes  
+
F1=Help  
F5=Reset  
F9=Shell  
F2=Refresh  
F6=Command  
F10=Exit  
F3=Cancel  
F7=Edit  
Enter=Do  
F4=List  
F8=Image  
4. Reboot the machine to bring the change into effect.  
The same task can be executed from the command line by entering:  
# chdev -l scsi1 -a id=6 -P  
Also with this method, a reboot is required to bring the change into effect.  
The procedure to change the ID of an SCSI-2 Differential Fast/Wide  
Adapter/A or Enhanced SCSI-2 Differential Fast/Wide Adapter/A is almost the  
same as the one described above. Here, the adapter that you choose from  
the list you get after executing the smit chgsyscommand should be an ascsi  
device. Also, as shown below, you need to change the external SCSI ID only.  
80  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Change/Show Characteristics of a SCSI Adapter  
SCSI adapter  
ascsi1  
Description  
Wide SCSI I/O Control>  
Status  
Location  
Internal SCSI ID  
External SCSI ID  
WIDE bus enabled  
...  
Available  
00-06  
7
[6]  
yes  
+#  
+#  
+
Apply change to DATABASE only  
yes  
The command line version of this is:  
# chdev -l ascsi1 -a id=6 -P  
As in the case of the SCSI-2 Differential Controller, a system reboot is  
required to bring the change into effect.  
The maximum length of the bus, including any internal cabling in disk  
subsystems, is limited to 19 meters for buses connected to the SCSI-2  
Differential Controller, and 25 meters for those connected to the SCSI-2  
Differential Fast/Wide Adapter/A or Enhanced SCSI-2 Differential Fast/Wide  
Adapter/A.  
3.4 Shared LVM Component Configuration  
This section describes how to define the LVM components shared by cluster  
nodes in an HACMP for AIX cluster environment.  
Creating the volume groups, logical volumes, and file systems shared by the  
nodes in an HACMP cluster requires that you perform steps on all nodes in  
the cluster. In general, you define the components on one node (referred to in  
the text as the source node) and then import the volume group on the other  
nodes in the cluster (referred to as destination nodes). This ensures that the  
ODM definitions of the shared components are the same on all nodes in the  
cluster.  
Non-concurrent access environments typically use journaled file systems to  
manage data, while concurrent access environments use raw logical  
volumes. This chapter provides different instructions for defining shared LVM  
components in non-concurrent access and concurrent access environments.  
Cluster Hardware and Software Preparation 81  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
3.4.1 Creating Shared VGs  
The following sections contain information about creating non-concurrent  
VGs and VGs for concurrent access.  
3.4.1.1 Creating Non-Concurrent VGs  
This section covers how to create a shared volume group on the source node  
using the SMIT interface. Use the smit mkvgfastpath to create a shared  
volume group. Use the default field values unless your site has other  
requirements, or unless you are specifically instructed otherwise here.  
Table 13. smit mkvg Options (Non-Concurrent)  
Options  
Description  
VOLUME GROUP name  
The name of the shared volume group should be  
unique within the cluster.  
Activate  
volume  
group Set to no so that the volume group can be  
AUTOMATICALLY at system activated as appropriate by the cluster event  
restart? scripts.  
ACTIVATE volume group after it Set to yes.  
is created?  
Volume  
NUMBER  
Group  
MAJOR If you are not using NFS, you can use the default  
(which is the next available number in the valid  
range). If you are using NFS, you must make sure  
to use the same major number on all nodes. Use  
the lvlstmajor command on each node to  
determine a free major number common to all  
nodes.  
3.4.1.2 Creating VGs for Concurrent Access  
The procedure used to create a concurrent access volume group varies  
depending on which type of device you are using: serial disk subsystem  
(7133) or RAID disk subsystem (7135).  
Note  
If you are creating (or plan to create) concurrent volume groups on SSA  
devices, be sure to assign unique non-zero node numbers through the  
SSAR on each cluster node. If you plan to specify SSA disk fencing in your  
concurrent resource group, the node numbers are assigned when you  
synchronize resources. If you do not specify SSA disk fencing, assign node  
numbers using the following command: chdev -l ssar -a node_number=x,  
where x is the number to assign to that node. You must reboot the system  
to effect the change.  
82  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
Creating a Concurrent Access Volume Group on Serial Disk  
Subsystems  
To use a concurrent access volume group, defined on a serial disk subsystem  
such as an IBM 7133 disk subsystem, you must create it as a  
concurrent-capable volume group. A concurrent-capable volume group can  
be activated (varied on) in either non-concurrent mode or concurrent access  
mode. To define logical volumes on a concurrent-capable volume group, it  
must be varied on in non-concurrent mode.  
You can use smit mkvgwith the following options to build the volume group:  
Table 14. smit mkvg Options (Concurrent, Non-RAID)  
Options  
Description  
VOLUME GROUP name  
Specify name of volume group.  
Physical partition SIZE in Accept the default.  
megabytes  
PHYSICAL VOLUME NAMES  
Specify the names of the physical volumes you want  
included in the volume group.  
Activate  
volume  
group Set this field to no so that the volume group can be  
AUTOMATICALLY at system activated as appropriate by the cluster event scripts.  
restart?  
ACTIVATE volume group after it Set this field to no.  
is created?  
Volume  
Group  
MAJOR Accept the default.  
NUMBER  
Create VG concurrent capable? Set this field to yes so that the volume group can be  
activated in concurrent access mode by the HACMP  
for AIX event scripts.  
Auto-varyon concurrent mode?  
Set this field to no so that the volume group can be  
activated as appropriate by the cluster event scripts.  
Creating a Concurrent Access Volume Group on RAID Disk  
Subsystems  
To create a concurrent access volume group on a RAID disk subsystem, such  
as an IBM 7135 disk subsystem, follow the same procedure as you would to  
create a non-concurrent access volume group. A concurrent access volume  
group can be activated (varied on) in either non-concurrent mode or  
concurrent access mode. To define logical volumes on a concurrent access  
volume group, it must be varied on in non-concurrent mode.  
Cluster Hardware and Software Preparation 83  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
Use the smit mkvg fastpath to create a shared volume group. Use the default  
field values unless your site has other requirements, or unless you are  
specifically instructed otherwise.  
Table 15. smit mkvg Options (Concurrent, RAID)  
Options  
Description  
VOLUME GROUP name  
The name of the shared volume group should be  
unique within the cluster.  
Activate  
volume  
group Set to no so that the volume group can be  
AUTOMATICALLY at system activated as appropriate by the cluster event  
restart? scripts.  
ACTIVATE volume group after it Set to yes.  
is created?  
Volume  
NUMBER  
Group  
MAJOR While it is only really required when you are using  
NFS, it is always good practice in an HACMP  
cluster to have a shared volume group have the  
same major number on all the nodes that serve it.  
Use the lvlstmajor command on each node to  
determine a free major number common to all  
nodes.  
Create VG concurrent capable?  
Set this field to no.  
3.4.2 Creating Shared LVs and File Systems  
Use the smit crjfsfast path to create the shared file system on the source  
node. When you create a journaled file system, AIX creates the  
corresponding logical volume. Therefore, you do not need to define a logical  
volume. You do, however, need to later rename both the logical volume and  
the log logical volume for the file system and volume group.  
Table 16. smit crjfs Options  
Options  
Description  
Mount AUTOMATICALLY at Make sure this field is set to no.  
system restart?  
Start Disk Accounting  
Make sure this field is set to no.  
Renaming a jfslog and Logical Volumes on the Source Node  
AIX assigns a logical volume name to each logical volume it creates.  
Examples of logical volume names are /dev/lv00 and /dev/lv01. Within an  
HACMP cluster, the name of any shared logical volume must be unique. Also,  
84  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
the journaled file system log (jfslog) is a logical volume that requires a unique  
name in the cluster.  
To make sure that logical volumes have unique names, rename the logical  
volume associated with the file system and the corresponding jfslog logical  
volume. Use a naming scheme that indicates the logical volume is associated  
with a certain file system. For example, lvsharefs could name a logical  
volume for the /sharefs file system.  
1. Use the lsvg -l volume_group_namecommand to determine the name of the  
logical volume and the log logical volume (jfslog) associated with the  
shared volume groups. In the resulting display, look for the logical volume  
name that has type jfs. This is the logical volume. Then look for the logical  
volume name that has type jfslog. This is the log logical volume.  
2. Use the smit chlvfastpath to rename the logical volume and the log  
logical volume.  
3. After renaming the jfslog or a logical volume, check the /etc/filesystems  
file to make sure the dev and log attributes reflect the change. Check the  
log attribute for each file system in the volume group, and make sure that  
it has the new jfslog name. Check the dev attribute for the logical volume  
that you renamed, and make sure that it has the new logical volume name.  
Adding Copies to Logical Volume on the Source Node  
Note  
These steps do not apply to RAID devices, which provide their own  
mirroring of logical volumes.  
1. Use the smit mklvcopy fastpath to add copies to a logical volume. Add  
copies to both the jfslog log logical volume and the logical volumes in the  
shared file systems. To avoid space problems, first mirror the jfslog log  
logical volume and then the shared logical volumes.  
The copies should reside on separate disks that are controlled by different  
disk adapters and are located in separate drawers or units, if possible.  
2. Verify the number of logical volume copies by entering: lsvg -l  
volume_group_name. In the resulting display, locate the line for the logical  
volume for which you just added copies. Notice that the number in the  
physical partitions column is x times the number in the logical partitions  
column, where x is the number of copies.  
3. To verify the placement of logical volume copies, enter: lspv -l hdiskx,  
where hdiskx is the name of each disk to which you assigned copies.  
Cluster Hardware and Software Preparation 85  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
That is, you enter this command for each disk. In the resulting display,  
locate the line for the logical volume for which you just added copies. For  
copies placed on separate disks, the numbers in the logical partitions  
column and the physical partitions column should be equal. Otherwise, the  
copies were placed on the same disk and the mirrored copies will not  
protect against disk failure.  
Testing a File System  
To run a consistency check on each file system’s information:  
1. Enter: fsck /filesystem_name  
2. Verify that you can mount the file system by entering:  
mount /filesystem_name  
3. Verify that you can unmount the file system by entering:  
umount /filesystem_name  
3.4.3 Mirroring Strategies  
Shared logical volumes residing on non-RAID disk devices should be  
mirrored in AIX to eliminate the disk as a single point of failure. Shared  
volume groups residing on a RAID device should not be AIX mirrored; the  
disk array provides its own data redundancy.  
The copies should reside on separate disks that are controlled by different  
disk adapters and are located in separate drawers or units, if possible.  
3.4.4 Importing to Other Nodes  
The following sections cover: varying off a volume group on the source node,  
importing it onto the destination node, changing its startup status, and varying  
it off on the destination nodes.  
3.4.4.1 Varying Off a Volume Group on the Source Node  
After completing the previous tasks, use the varyoffvgcommand to  
deactivate the shared volume group. You vary off the volume group so that it  
can be properly imported onto a destination node and activated as  
appropriate by the cluster event scripts. Enter the following command:  
varyoffvg volume_group_name. Make sure that all the file systems of the  
volume group have been unmounted, otherwise the varyoffvg command will  
not work.  
3.4.4.2 Importing a Volume Group onto the Destination Node  
This section covers how to import a volume group onto destination nodes  
using the SMIT interface. You can also use the TaskGuide utility for this task.  
86  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
The TaskGuide uses a graphical interface to guide you through the steps of  
adding nodes to an existing volume group. For more information on the  
Importing the volume group onto the destination nodes synchronizes the  
ODM definition of the volume group on each node on which it is imported.  
You can use the smit importvgfastpath to import the volume group.  
Table 17. smit importvg Options  
Options  
Description  
VOLUME GROUP name  
Enter the name of the volume group that you are  
importing. Make sure the volume group name is the  
same name that you used on the source node.  
PHYSICAL VOLUME name  
Enter the name of a physical volume that resides in  
the volume group. Note that a disk may have a  
different logical name on different nodes. Make sure  
that you use the disk name as it is defined on the  
destination node.  
ACTIVATE volume group after it Set the field to yes.  
is imported?  
Volume  
NUMBER  
Group  
MAJOR If you are not using NFS, you may use the default  
(which is the next available number in the valid  
range). If you are using NFS, you must make sure to  
use the same major number on all nodes. Use the  
lvlstmajorcommand on each node to determine a  
free major number common to all nodes.  
3.4.4.3 Changing a Volume Group’s Startup Status  
By default, a volume group that has just been imported is configured to  
automatically become active at system restart. In an HACMP for AIX  
environment, a volume group should be varied on as appropriate by the  
cluster event scripts. Therefore, after importing a volume group, use the SMIT  
Change a Volume Group screen to reconfigure the volume group so that it is  
not activated automatically at system restart.  
Use the smit chvg fastpath to change the characteristics of a volume group.  
Table 18. smit crjfs Options  
Options  
Description  
Activate  
volume  
group Set this field to no.  
automatically at system restart?  
Cluster Hardware and Software Preparation 87  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
Options  
Description  
A QUORUM of disks required to This field is site-dependent. See 3.4.5, “Quorum” on  
keep the volume group online?  
page 88 for a discussion of quorum in an HACMP  
cluster.  
3.4.4.4 Varying Off the Volume Group on the Destination Nodes  
Use the varyoffvg command to deactivate the shared volume group so that it  
can be imported onto another destination node or activated as appropriate by  
the cluster event scripts. Enter: varyoffvg volume_group_name.  
3.4.5 Quorum  
Note  
This section does not apply to the IBM 7135-110 or 7135-210 RAIDiant  
Disk Array, which provides its own data redundancy.  
Quorum is a feature of the AIX LVM that determines whether or not a volume  
group can be placed online using the varyonvgcommand, and whether or not  
it can remain online after a failure of one or more of the physical volumes in  
the volume group.  
Each physical volume in a volume group has a Volume Group Descriptor  
Area (VGDA) and a Volume Group Status Area (VGSA).  
VGDA  
VGSA  
Describes the physical volumes (PVs) and logical volumes (LVs)  
that make up a volume group and maps logical partitions to  
physical partitions. The varyonvgcommand reads information from  
this area.  
Maintains the status of all physical volumes and physical partitions  
in the volume group. It stores information regarding whether a  
physical partition is potentially inconsistent (stale) with mirror  
copies on other physical partitions, or is consistent or  
synchronized with its mirror copies. Proper functioning of LVM  
mirroring relies upon the availability and accuracy of the VGSA  
data.  
3.4.5.1 Quorum at Vary On  
When a volume group is brought online using the varyonvg command, VGDA  
and VGSA data structures are examined. If more than half of the copies are  
readable and identical in content, quorum is achieved and the varyonvg  
88  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
command succeeds. If exactly half the copies are available, as with two of  
four, quorum is not achieved and the varyonvgcommand fails.  
3.4.5.2 Quorum after Vary On  
If a write to a physical volume fails, the VGSAs on the other physical volumes  
within the volume group are updated to indicate that one physical volume has  
failed. As long as more than half of all VGDAs and VGSAs can be written,  
quorum is maintained and the volume group remains varied on. If exactly half  
or less than half of the VGDAs and VGSAs are inaccessible, quorum is lost,  
the volume group is varied off, and its data becomes unavailable.  
Keep in mind that a volume group can be varied on or remain varied on with  
one or more of the physical volumes unavailable. However, data contained on  
the missing physical volume will not be accessible unless the data is  
replicated using LVM mirroring, and a mirror copy of the data is still available  
on another physical volume. Maintaining quorum without mirroring does not  
guarantee that all data contained in a volume group is available.  
Quorum has nothing to do with the availability of mirrored data. It is possible  
to have failures that result in loss of all copies of a logical volume, yet the  
volume group remains varied on because a quorum of VGDAs/VGSAs are  
still accessible.  
3.4.5.3 Disabling and Enabling Quorum  
Quorum checking is enabled by default. Quorum checking can be disabled  
using the chvg -Qn vgnamecommand, or by using the smit chvg fastpath.  
Quorum Enabled  
With quorum enabled, more than half of the physical volumes must be  
available and the VGDA and VGSA data structures must be identical before a  
volume group can be varied on with the varyonvgcommand.  
With quorum enabled, a volume group will be forced offline if one or more  
disk failures cause a majority of the physical volumes to be unavailable.  
Having three or more disks in a volume group avoids a loss of quorum in the  
event of a single disk failure.  
Quorum Disabled  
With quorum disabled, all the physical volumes in the volume group must be  
available and the VGDA data structures must be identical for the varyonvg  
command to succeed. With quorum disabled, a volume group will remain  
varied on until the last physical volume in the volume group becomes  
unavailable. This section summarizes the effect quorum has on the  
availability of a volume group.  
Cluster Hardware and Software Preparation 89  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Forcing a Varyon  
A volume group with quorum disabled and one or more physical volumes  
unavailable can be “forced” to vary on by using the -f flag with the varyonvg  
command. Forcing a varyon with missing disk resources can cause  
unpredictable results, including a reducevg of the physical volume from the  
volume group. Forcing a varyon should be an overt (manual) action and  
should only be performed with a complete understanding of the risks  
involved.  
The HACMP for AIX software assumes that a volume group is not degraded  
and all physical volumes are available when the varyonvg command is issued  
at startup or when a volume group resource is taken over during a fallover.  
The cluster event scripts provided with the HACMP for AIX software do not  
“force” varyon with the -f flag, which could cause unpredictable results. For  
this reason, modifying the cluster event scripts to use the -f flag is strongly  
discouraged.  
Quorum in Non-Concurrent Access Configurations  
While specific scenarios can be constructed where quorum protection does  
provide some level of protection against data corruption and loss of  
availability, quorum provides very little actual protection in non-concurrent  
access configurations. In fact, enabling quorum may mask failures by  
allowing a volume group to varyon with missing resources. Also, designing  
logical volume configuration for no single point of failure with quorum enabled  
may require the purchase of additional hardware. Although these facts are  
true, you must keep in mind that disabling quorum can result in subsequent  
loss of disks—after varying on the volume group—that go undetected.  
Quorum in Concurrent Access Configurations  
Quorum must be enabled for an HACMP for AIX concurrent access  
configuration. Disabling quorum could result in data corruption. Any  
concurrent access configuration where multiple failures could result in no  
common shared disk between cluster nodes has the potential for data  
corruption or inconsistency.  
3.4.6 Alternate Method - TaskGuide  
The TaskGuide is a graphical interface that simplifies the task of creating a  
shared volume group within an HACMP cluster configuration. The TaskGuide  
presents a series of panels that guide the user through the steps of specifying  
initial and sharing nodes, disks, concurrent or non-concurrent access, volume  
group name and physical partition size, and cluster settings. The TaskGuide  
can reduce errors, as it does not allow a user to proceed with steps that  
90  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
conflict with the cluster’s configuration. Online help panels give additional  
information to aid in each step.  
3.4.6.1 TaskGuide Requirements  
Before starting the TaskGuide, make sure:  
• You have a configured HACMP cluster in place.  
• You are on a graphics capable terminal.  
3.4.6.2 Starting the TaskGuide  
You can start the TaskGuide from the command line by typing:  
/usr/sbin/cluster/tguides/bin/cl_ccvg or you can use the SMIT interface as  
follows:  
1. Type smit hacmp.  
2. From the SMIT main menu, choose Cluster System Management ->  
Cluster Logical Volume Manager ->Taskguide for Creating a Shared  
Volume Group. After a pause, the TaskGuide Welcome panel appears.  
3. Proceed through the panels to create or share a volume group.  
Cluster Hardware and Software Preparation 91  
Download from Www.Somanuals.com. All Manuals Search And Download.  
92  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Chapter 4. HACMP Installation and Cluster Definition  
This chapter describes issues concerning the actual installation of HACMP  
Version 4.3 and the definition of a cluster and its resources. It concentrates  
on the HACMP part of the installation, so, we will assume AIX is already at  
the 4.3.2 level. Please refer to the AIX Version 4.3: Migration Guide,  
SG24-5116, for details on installation or migration to that level.  
This chapter is meant to give an overview of the steps to be taken, and not to  
be a complete handbook for performing these tasks. When actually  
performing the HACMP install, the HACMP for AIX, Version 4.3: Installation  
Guide, SC23-4278 should be consulted.  
4.1 Installing HACMP  
Before installing, you need to ensure that all the prerequisites are met.  
Chapter 8 of the HACMP for AIX, Version 4.3: Installation Guide, SC23-4278,  
gives a detailed list. The AIX Level of the server nodes has to be at AIX 4.3.2,  
for example, and the required free space in /usrmust be confirmed. For parts  
of the product, like HAView, there are prerequisites for other lpps, nv6000 in  
this case, that have to be ensured.  
You can install either from the installation media, from an installation server  
through Network Installation Management (NIM), or from a hard disk to which  
the software has been copied.  
You will either be installing the HACMP for AIX software for the first time, or  
upgrading from an earlier version. Both of those situations are discussed in  
the following sections.  
4.1.1 First Time Installs  
There are a number of filesets involved in an HACMP Installation. Here is a  
short overview of them, and what their purpose is.  
• cluster.base  
This is the basic component that has to be installed on all server nodes in  
the cluster, and it contains the following:  
cluster.base.client.lib  
cluster.base.client.rte  
cluster.base.client.utils  
cluster.base.server.diag  
cluster.base.server.events  
cluster.base.server.rte  
HACMP Base Client Libraries  
HACMP Base Client Runtime  
HACMP Base Client Utilities  
HACMP Base Server Diags  
HACMP Base Server Events  
HACMP Base Server Runtime  
© Copyright IBM Corp. 1999  
93  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
cluster.base.server.utils  
HACMP Base Server Utilities  
• cluster.cspoc  
This component includes all of the commands and environment for the  
C-SPOC utility, the Cluster-Single Point Of Control feature. These routines  
are responsible for centralized administration of the cluster. There is no  
restriction on the node from which you run the C-SPOC utility commands,  
so it should also be installed on all the server nodes. It consists of the  
following:  
cluster.cspoc.rte  
cluster.cspoc.cmds  
cluster.cspoc.dsh  
HACMP CSPOC Runtime Commands  
HACMP CSPOC commands  
HACMP CSPOC dsh and perl  
• cluster.adt  
This component contains demo clients and their include files, for example,  
for building a clinfo client on a non-AIX machine. Since these are sample  
files and demos, you might want to install this on a dedicated machine  
only. This machine can further be used for development of server or client  
code:  
cluster.adt.client.demos  
HACMP Client Demos  
cluster.adt.client.samples.demos HACMP Client Demos Samples  
cluster.adt.client.samples.clinfo HACMP Client clinfo Samples  
cluster.adt.client.samples.clstat HACMP Client clstat Samples  
cluster.adt.client.include  
HACMP Client includes  
cluster.adt.client.samples.libcl HACMP Client libcl Samples  
cluster.adt.server.samples.images HACMP Sample Images  
cluster.adt.server.demos  
HACMP Server Demos  
cluster.adt.server.samples.demos HACMP Server Sample Demos  
• cluster.man.en_US.data  
This component contains the man pages in US English. You may like to  
exchange this with your own language:  
cluster.man.en_US.cspoc.data  
cluster.man.en_US.client.data  
cluster.man.en_US.server.data  
cluster.man.en_US.haview.data  
HACMP CSPOC Man pages  
HACMP Client Man pages  
HACMP Server Man pages  
HACMP HAView Man pages  
• cluster.msg.en_US  
These filesets contain the messages in US English. In contrast to the man  
pages, the en_US version must be installed. You might add your  
language’s messages if you want:  
cluster.msg.en_US.cspoc  
cluster.msg.en_US.client  
cluster.man.en_US.haview.data  
HACMP CSPOC Messages  
HACMP Client Messages  
HACMP HAView Messages  
94  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
• cluster.vsm  
The Visual Systems Management Fileset contains Icons and bitmaps for  
the graphical Management of HACMP Resources, as well as the xhacmpm  
command:  
cluster.vsm  
HACMP X11 Dependent  
• cluster.haview  
This fileset contains the files for including HACMP cluster views into a  
TME 10 Netview Environment. It is installed on a Netview network  
management machine, and not on a cluster node:  
cluster.haview  
HACMP HAView  
• cluster.man.en_US.haview.data  
This fileset contains man pages and data for the HAView component:  
cluster.man.en_US.haview.data  
HACMP HAView Manpages  
• cluster.msg.en_US.haview  
This fileset contains the US English messages for the HAView component:  
cluster.msg.en_US.haview  
HACMP HAView Messages  
Note  
TME 10 NetView for AIX must be installed on any system where you will  
install HAView. If NetView is installed using a client/server configuration,  
HAView should be installed on the NetView client; otherwise, install it on  
the NetView server node. Also, be aware that the NetView client should not  
be configured as a cluster node to avoid NetView's failure after a failover.  
• cluster.taskguides  
This is the fileset that contains the taskguide for easy creation of shared  
volume groups:  
cluster.taskguides.shrvolgrp  
HAES Shr Vol Grp Task Guides  
• cluster.clvm  
This fileset contains the Concurrent Resource Manager (CRM) option:  
cluster.clvm  
HACMP for AIX Concurrent Access  
• cluster.hc  
HACMP Installation and Cluster Definition 95  
Download from Www.Somanuals.com. All Manuals Search And Download.  
This fileset contains the Application Heart Beat Daemon, Oracle Parallel  
Server is an application that makes use of it:  
cluster.hc.rte  
Application Heart Beat Daemon  
The installation of CRM requires the following software:  
bos.rte.lvm.usr.4.3.2.0  
AIX Run-time Executable  
Install Server Nodes  
From whatever medium you are going to use, install the needed filesets on  
each node. Refer to Chapter 8 of the HACMP for AIX, Version 4.3: Installation  
Guide, SC23-4278 for details.  
Rebooting Servers  
The final step in installing the HACMP for AIX software is to reboot each  
server in your HACMP for AIX environment.  
4.1.2 Upgrading From a Previous Version  
If you are upgrading your cluster nodes from a previous version, there are  
some things you have to take care of in order to get your existing cluster back  
the way you want it after the upgrade is through.  
• Ensure that all the prerequisites are met. For details, look into Chapter 8  
of the HACMP for AIX Version 4.3: Installation Guide, SC23-4278.  
• Archive any localized script and configuration files to prevent losing them  
during an upgrade.  
• Its always a good idea to have a mksysb of a working system, so take one  
of the cluster nodes to be upgraded.  
• Commit your current HACMP Version 4.* software (if it is applied but not  
committed) so that the HACMP 4.3 software can be installed over the  
existing version.  
• Save the current configuration using the cluster snapshot utility, and save  
any customized event scripts in a directory of your own.  
(To review cluster snapshot instructions, see the chapter on saving and  
restoring cluster configurations in the HACMP for AIX, Version 4.3:  
Administration Guide, SC23-4279.)  
As the version from where you start may differ, there is more than one way to  
get to the required level.  
If your site is currently running an earlier version of the HACMP for AIX  
software in its cluster environment, except for Version 4.2.2 already running  
on AIX 4.3, the following procedures describe how to upgrade your existing  
96  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
HACMP software to HACMP for AIX, Version 4.3. The comments on  
upgrading the Operating System are not included.  
If you are already running AIX 4.3, see the special note at the end of this  
section.  
Note  
Although your objective in performing a migration installation is to keep the  
cluster operational and to preserve essential configuration information, do  
not run your cluster with mixed versions of the HACMP for AIX software for  
an extended period of time.  
4.1.2.1 Upgrading from Version 4.1.0 through 4.2.2 to Version 4.3  
The following procedure applies to upgrading a two-node or multi-node  
cluster running HACMP Version 4.1.0 through 4.2.2 to Version 4.3 when the  
installed AIX version is earlier than 4.3.  
To perform a rolling AIX migration installation and HACMP upgrade from  
Version 4.1.0 through Version 4.2.2 to Version 4.3, complete the following  
steps:  
Upgrade AIX on One Node  
The following steps describe how to upgrade AIX on one node:  
1. If you wish to save your cluster configuration, see the chapter “Saving and  
Restoring Cluster Configurations” in the HACMP for AIX, Version 4.3:  
Administration Guide, SC23-4279.  
2. Shut down the first node (gracefully with takeover) using the smit clstop  
fastpath. For this example, shut down Node A. Node B will take over Node  
A's resources and make them available to clients.  
See the chapter “Starting and Stopping Cluster Services” in the HACMP  
for AIX, Version 4.3: Administration Guide, SC23-4279 for more  
information about stopping cluster services.  
3. Perform a Migration Installation as described in your AIX Installation  
Guide, SBOF-1803 on Node A.  
The Migration Installation option preserves the current version of the  
HACMP for AIX software and upgrades the existing base operating  
system to AIX 4.3.2. Product (application) files and configuration data are  
also saved.  
4. Check the Migration Installation. Verify that all the disks are available. Run  
lppchk -vand oslevel to ensure that the system is in a stable state.  
HACMP Installation and Cluster Definition 97  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Install HACMP 4.3 for AIX on Node A  
5. After upgrading AIX and verifying that the disks are correctly configured,  
install the HACMP 4.3 for AIX software on Node A. For a short description  
of the filesets, please refer to 4.1.1, “First Time Installs” on page 93 or to  
Chapter 8 of the HACMP for AIX, Version 4.3: Installation Guide,  
SC23-4278.  
6. The installation process automatically runs the cl_convertprogram. It  
removes the current HACMP objects from /etc/objreposand saves them  
to HACMP.old. It creates new HACMP ODM object classes for Version 4.3  
in /etc/objrepos.  
Note  
If upgrading from HACMP 4.1 for AIX, you must run the cl_convertutility  
manually. Enter the following command:  
/usr/sbin/cluster/conversion/cl_convert -v 4.1  
7. Start the Version 4.3 software on Node A using the smit clstartfastpath.  
After HACMP is running, start the previous version HACMP software on  
Node B, if it is not still running. Check to ensure that the nodes  
successfully join the cluster.  
Important  
If the node running Version 4.3 fails while the cluster is in this state, the  
surviving node running the previous version may not successfully mount  
the file systems that were not properly unmounted due to Node A’s  
failure.  
8. Repeat Steps 2 through 7 on Node B on remaining cluster nodes, one at a  
time.  
Important  
In a multi-node cluster, do not synchronize the node configuration or the  
cluster topology until the last node has been upgraded.  
9. When the last node has been upgraded to both AIX 4.3.2 and HACMP 4.3,  
the cluster install/upgrade process is complete.  
Check Upgraded Configuration  
10.If using tty devices, check that the tty device is configured as a serial  
network using the smit chgttyfastpath.  
11.In order to verify and synchronize the configuration (if desired), you must  
have /.rhosts files on cluster nodes. If they do not exist, create the /.rhosts  
98  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
file on Node A using the following command:  
/usr/sbin/cluster/utilities/cllsif -x >> /.rhosts  
This command will append information to the /.rhosts file instead of  
overwriting it. Then, you can ftp this file to the other nodes as necessary.  
12.Verify the cluster topology on all nodes using the clverifyutility.  
13.Check that custom event scripts are properly installed.  
14.Synchronize the node configuration and the cluster topology from Node A  
to all nodes (this step is optional).  
15.It is recommended that you test the upgraded cluster to ensure proper  
behavior.  
Client-only Migration  
If you are migrating from an HACMP 4.1 for AIX through HACMP 4.2 for AIX  
server node to a client-only node running Version 4.3, first remove the  
existing server portion of HACMP. If, after upgrading AIX, you install the  
cluster.base.client.* filesets on a node running an earlier version of HACMP  
for AIX without de-installing the server, the results are unpredictable.  
To determine if there is a mismatch between the HACMP client and server  
software installed on a node, issue the following command to list the installed  
software:  
lslpp -L "cluster*"  
Examine the list and make sure that all cluster filesets are at 4.3.0.  
If you determine that there is a mismatch between the client and server,  
de-install the server and then repeat the installation of the client software.  
In case the node was just a client before, you only have to install the  
cluster.base.client.* filesets after it has been migrated to AIX Version 4.3.2.  
Again, please check whether the installation succeeded by issuing the  
command: lslpp -L "cluster*"  
4.1.2.2 Upgrading from Version 4.2.2 on AIX 4.3.2 to HACMP Version 4.3  
The following steps describe upgrading from Version 4.2.2 on AIX 4.3.2 to  
HACMP Version 4.3:  
1. If your cluster is currently running Version 4.2.2 of the HACMP software on  
AIX Version 4.3.2, you should upgrade your cluster configuration to  
HACMP 4.3 for AIX.  
HACMP Installation and Cluster Definition 99  
Download from Www.Somanuals.com. All Manuals Search And Download.  
2. If you wish to save your cluster configuration, see the chapter Saving and  
Restoring Cluster Configurations in the HACMP for AIX, Version 4.3:  
Administration Guide, SC23-4279.  
3. Commit your current HACMP for AIX software on all nodes.  
4. Shut down one node (gracefully with takeover) using the smit clstop  
fastpath. For this example, shut down Node A. Node B will take over Node  
A’s resources and make them available to clients.  
See the chapter “Starting and Stopping Cluster Services” in the HACMP  
for AIX, Version 4.3: Administration Guide, SC23-4279, for more  
information on stopping cluster services.  
5. Install HACMP for AIX Version 4.3. See Chapter 8 of the HACMP for AIX,  
Version 4.3: Installation Guide, SC23-4278, starting with the section  
“Installation Choices”, for instructions.  
The cl_convertutility automatically updates the HACMP ODM object  
classes to the 4.3 version.  
Note  
If IP address swapping is being used on this node, that is, a boot  
address is defined for this node, check to ensure that the HACMP  
changes to /etc/inittaband /etc/rc.netexist as specified in Appendix  
A of the HACMP for AIX, Version 4.3: Installation Guide, SC23-4278,  
before rebooting the node.  
6. Reboot Node A.  
7. Start the HACMP for AIX software on Node A using the smit clstart  
fastpath and verify that Node A successfully joins the cluster.  
8. Repeat Steps 3 through 7 on remaining cluster nodes, one at a time.  
9. After all nodes have been upgraded to HACMP Version 4.3, synchronize  
the node configuration and the cluster topology from Node A to all nodes.  
10.Verify the cluster topology on all nodes using the clverifyutility.  
11.Complete a test phase on the cluster before putting it into production.  
4.2 Defining Cluster Topology  
The cluster topology is comprised of the following components:  
• The cluster definition  
• The cluster nodes  
• The network adapters  
100  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
• The network modules  
You define the cluster topology by entering information about each  
component into HACMP-specific ODM classes. You enter the HACMP ODM  
data by using the HACMP SMIT interface or the VSM utility xhacmpm. The  
xhacmpmutility is an X Windows tool for creating cluster configurations using  
icons to represent cluster components. For more information about the  
xhacmpmutility, see the administrative facilities chapter of the HACMP for AIX,  
Version 4.3: Concepts and Facilities, SC23-4276.  
Note  
The SP Switch network module can support multiple clusters; therefore, its  
settings should remain at their default values to avoid affecting HACMP  
event scripts. If you must change these settings, see the chapter on  
changing the cluster topology in the HACMP for AIX, Version 4.3:  
Administration Guide, SC23-4279 for more information.  
4.2.1 Defining the Cluster  
The cluster ID and name identifies a cluster in an HACMP environment. The  
cluster ID and name must be unique for each cluster defined.  
Cluster IDs have to be a positive integer in the range from 1 through 99999,  
and the cluster name is a text string of up to 31 alphanumeric characters,  
including underscores. It doesn’t necessarily need to match the hostname.  
The HACMP software uses this information to create the cluster entries for  
the ODM.  
4.2.2 Defining Nodes  
After defining the cluster name and ID, cluster nodes have to be defined. As  
above, this is usually done through smit hacmp. Each of the cluster nodes  
needs a unique name, so the cluster manager can address them.  
Again, a node name is a text string of up to 31 alphanumeric characters that  
can contain underscores.  
You can add more than one node at a time by separating them with  
whitespace characters.  
HACMP Installation and Cluster Definition 101  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Note  
The node names are logically sorted in their ascii order within HACMP in  
order to decide which nodes are considered to be neighbors for heartbeat  
purposes.  
In order to build a logical ring, a node always talks to its up- and  
downstream neighbor in their node name’s ascii order. The uppermost and  
the lowest node are also considered neighbors.  
Adding or Changing a Node Name after the Initial Configuration  
If you want to add or change a node name after the initial configuration, use  
the Change/Show Cluster Node Name screen. See the chapter on changing  
the cluster topology of the HACMP for AIX, Version 4.3:Administration Guide,  
SC23-4279 for more information.  
4.2.3 Defining Adapters  
To define the adapters after defining the node names, first consult your  
planning worksheets for both TCP/IP and serial networks.  
There are a number of attributes associated with an Adapter in the HACMP  
configuration which need to be specified:  
Adapter IP Label  
Enter the IP label (the name) of the adapter you have  
chosen as the service address for this adapter. Adapter  
labels can be any ASCII text string consisting of  
alphabetical and numeric characters, underscores, and  
hyphens, up to 31 characters.  
If IP address takeover is defined for that adapter, a boot  
adapter (address) label has to be defined for it. Use a  
consistent naming convention for boot adapter labels.  
(You will choose the Add an Adapter option again to  
define the boot adapter when you finish defining the  
service adapter.)  
You can use hyphens in adapter labels. However,  
currently it might not be a good idea since the  
/usr/sbin/cluster/diag/clverifyutility flags adapter  
labels that contain hyphens each time it runs.  
Network Type  
Indicate the type of network to which this adapter is  
connected. Pre-installed network modules are listed on  
the pop-up pick list.  
102  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
Network Name  
Enter an ASCII text string that identifies the network.  
The network name can include alphabetic and numeric  
characters and underscores. Use no more than 31  
characters. The network name is arbitrary, but must be  
used consistently for adapters on the same physical  
network.  
If several adapters share the same physical network,  
make sure you use the same network name for each of  
these adapters.  
Network Attribute Indicate whether the network is public, private, or  
serial. Press Tab to toggle the values. In the context of  
HACMP, serial networks means “non-TCP/IP”; public  
and private networks are TCP/IP networks. Ethernet,  
Token-Ring, FDDI, and SLIP are public networks.  
SOCC, ATM, and an SP Switch are private networks.  
RS232 lines, target mode SSA loops, and target mode  
SCSI-2 buses are serial networks.  
Adapter Function Indicate whether the adapter's function is service,  
standby, or boot. Press Tab to toggle the values. A node  
has a single service adapter for each public or private  
network. A serial network has only a single service  
adapter.  
A node can have none, one, or more standby adapters  
for each public network. Serial and private networks do  
not have standby adapters, with the exception of ATM  
networks. ATM networks must be defined as private,  
and therefore standby adapters are supported.  
In an HACMP environment on the RS/6000 SP, the  
ethernet adapters can be configured as service  
adapters but should not be configured for IP address  
takeover. Regarding the SP Switch, network, boot, and  
service addresses used for IP address takeover are  
ifconfig alias addresses used on the css0 network. See  
the appendix entitled HACMP for the RS/6000 SP in the  
HACMP Installation Guide Version 4.3 for more  
information on adapter functions in an SP Switch  
environment.  
In an ATM network, the adapter function should be listed  
as svc_s to indicate that the interface is used by  
HACMP servers. Keep in mind that the netmask for all  
adapters in an HACMP network must be the same.  
HACMP Installation and Cluster Definition 103  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Adapter Identifier Enter the IP address in dotted decimal format or a  
device file name. IP address information is required for  
non-serial network adapters only if the node’s address  
cannot be obtained from the domain name server or the  
local /etc/hosts file (using the adapter IP label given).  
You must enter device filenames for serial network  
adapters. RS232 serial adapters must have the device  
filename /dev/ttyN. Target mode SCSI serial adapters  
must have the device file name /dev/tmscsiN. Target  
mode SSA adapters must have the device file name  
/dev/tmssaN.im or /dev/tmssaN.tm.  
Adapter Hardware Address(optional) Enter a hardware address for the  
adapter. The hardware address must be unique within  
the physical network. Enter a value in this field only if:  
You are currently defining a service adapter, and the  
adapter has a boot address, and you want to use  
hardware address swapping. See the chapter on  
planning TCP/IP networks in the HACMP for AIX,  
Version 4.3:Planning Guide, SC23-4277 for more  
information on hardware address swapping. This facility  
is supported only for Ethernet, Token Ring, and FDDI  
adapters. It does not work with the SP Switch.  
Node Name  
Define a node name for all adapters except for those  
service adapters whose addresses may be shared by  
nodes participating in the resource chain for a rotating  
resource configuration. These adapters are rotating  
resources. The event scripts use the user-defined  
configuration to associate these service addresses with  
the proper node. In all other cases, addresses are  
associated with a particular node (service, boot, and  
standby)  
Note  
Although it is possible to have only one physical network adapter (no  
standby adapters), this constitutes a potential single point of failure  
condition and is not recommended for an HACMP configuration.  
104  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
Note  
When IPAT is configured, the run level of the IP-related entries (e. g.  
rctcpip, rcnfs...) of the /etc/inittabare changed to “a”. This has the result  
that these services are not started at boot time, but with HACMP.  
Adding or Changing Adapters after the Initial Configuration  
If you want to change the information about an adapter after the initial  
configuration, use the Change/Show an Adapter screen. See the chapter on  
changing the cluster topology in the HACMP for AIX, Version 4.3:  
Administration Guide, SC23-4279, for more information.  
4.2.4 Configuring Network Modules  
Each supported cluster network in a configured HACMP cluster has a  
corresponding cluster network module. Each network module monitors all I/O  
to its cluster network.  
The Network Modules are pre-loaded when you install the HACMP software.  
You do not need to enter information in the Network Module SMIT screens  
unless you want to change some field associated with a network module,  
such as the failure detection rate.  
Each network module maintains a connection to other network modules in the  
cluster. The Cluster Managers on cluster nodes send messages to each other  
through these connections. Each network module is responsible for  
maintaining a working set of service adapters and for verifying connectivity to  
cluster peers. The network module is also responsible for reporting when a  
given link actually fails. It does this by sending and receiving periodic  
heartbeat messages to or from other network modules in the cluster, and  
reporting back to the Cluster Manager when it misses a threshold number of  
heartbeats.  
Currently, network modules support communication over the following types  
of networks:  
• Serial (RS232)  
Target-mode SCSI  
Target-mode SSA  
• IP (Generic IP)  
• Ethernet  
Token-Ring  
• FDDI  
• SOCC  
HACMP Installation and Cluster Definition 105  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
• SLIP  
• SP Switch  
• ATM  
It is highly unlikely that you will add or remove a network module. For  
information about changing a characteristic of a Network Module, such as the  
failure detection rate, see the chapter on changing the cluster topology in the  
HACMP for AIX, Version 4.3: Administration Guide, SC23-4279. Changing  
the network module allows the user to influence the rate of heartbeats being  
sent and received by a Network Module, thereby changing the sensitivity of  
the detection of a network failure.  
In HACMP/ES, topology services and group services are used instead of  
Network Interface Modules (NIMs) in order to keep track of the status of  
nodes, adapters or resources.  
In HACMP/ES, the tuning of network sensitivity is a little different.  
Customizable attributes are the interval between heartbeats in seconds and  
the Fibrillate Count, which is the acceptable number of missed heartbeats  
before some event is triggered. You will find the Change / Show Topology  
and Group Services Configuration in the Cluster Topology screen, just  
like the NIM tuning options.  
4.2.5 Synchronizing the Cluster Definition Across Nodes  
Synchronization of the cluster topology ensures that the ODM data on all  
cluster nodes is in sync. The HACMP ODM entries must be the same on each  
node in the cluster. If the definitions are not synchronized across nodes, the  
HACMP for AIX software generates a run-time error at cluster startup.  
Even if you have a cluster defined with only one node, you must still  
synchronize the cluster.  
The processing performed in synchronization varies depending on whether  
the cluster manager is active on the local node. If the cluster manager is not  
active on the local node when you select this option, the ODM data in the  
system default configuration directory (DCD) on the local node is copied to  
the ODMs stored in the DCDs on all cluster nodes. The cluster manager is  
typically not running when you synchronize the initial cluster configuration.  
If the cluster manager is active on the local node, the ODM data stored in the  
DCDs on all cluster nodes are synchronized. In addition, the configuration  
data stored in the active configuration directory (ACD) on each cluster node is  
overwritten with the new configuration data, which becomes the new active  
106  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
configuration. If the cluster manager is active on some other cluster nodes  
but not on the local node, the synchronization operation is aborted.  
Before attempting to synchronize a cluster configuration, ensure that all  
nodes are powered on, that the HACMP software is installed, and that the  
/etc/hosts and /.rhosts files on all nodes include all HACMP boot and service  
IP labels.  
The /.rhosts file may not be required if you are running HACMP on the SP  
system.The SP system uses kerberos as its security infrastructure. If you are  
running HACMP on a node with kerberos enabled (usually an SP node, but  
could also be a standalone RS/6000 that has been configured with kerberos),  
you can set a parameter in HACMP to use “Enhanced Security”. This feature  
removes the requirement of TCP/IP access control lists (for example, the  
/.rhosts file) on remote nodes during HACMP configuration. Instead, it uses a  
kerberized version of remote commands to accomplish the synchronization.  
It should be noted that Kerberos support is not included in standard AIX  
4.3.2. However, Kerberos is available in the public domain, and it is possible  
to get it and configure it on a non-SP RS/6000 node. This is not very common  
though, so you will almost always see HACMP Enhanced Security used on  
the SP system.  
When you synchronize the cluster topology, there are two options that control  
the behavior of this process as follows:  
Table 19. Options for Synchronization of the Cluster Topology  
Ignore Cluster  
Verification Errors  
If you specify yes, the result of the cluster verification  
performed as part of synchronization is ignored and the  
configuration is synchronized even if verification fails.  
If you specify no, the changes are not synchronized if  
verification fails. View the error messages in the system  
error log to determine the configuration problem.  
For information about the  
/usr/sbin/cluster/diag/clverifyutility, see the  
chapter on verifying a cluster configuration in the HACMP  
for AIX, Version 4.3: Administration Guide, SC23-4279.  
Emulate or Actual  
If you set this field to Emulate, the synchronization will be  
an emulation and will not affect the Cluster Manager. If you  
set this field to Actual, the synchronization will actually  
occur, and any subsequent changes will be made to the  
Cluster Manager. Emulate is the default value.  
The cluster topology definition (including all node, adapter, and network  
module information) is copied to the other nodes in the cluster.  
HACMP Installation and Cluster Definition 107  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
4.3 Defining Resources  
The HACMP for AIX software provides a highly available environment by  
identifying a set of cluster-wide resources essential to uninterrupted  
processing, and then by defining relationships among nodes that ensure  
these resources are available to client processes. Resources include the  
following hardware and software:  
• Disks  
• Volume groups  
• File systems  
• Network addresses  
• Application servers  
In the HACMP for AIX software, you define each resource as part of a  
resource group. This allows you to combine related resources into a single  
logical entity for easier configuration and management. You then configure  
each resource group to have a particular kind of relationship with a set of  
nodes. Depending on this relationship, resources can be defined as one of  
three types: cascading, concurrent access, or rotating. See 2.4.1, “Resource  
After configuring the cluster topology, you must configure resources and set  
up the cluster node. This involves:  
• Configuring resource groups and node relationships to behave as desired  
• Adding individual resources to each resource group  
• Setting up run-time parameters for each node  
• Synchronizing cluster nodes  
4.3.1 Configuring Resource Groups  
Resource Groups are initialized by telling the HACMP ODM their names, the  
participating nodes, and their relationship. The order in which the  
participating nodes are defined is taken as the priority of the resource chain,  
that is, priority is decreasing from left to right.  
The relationship can be one of Cascading, Rotating or Concurrent. See 2.4.1,  
108  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
4.3.1.1 Configuring Resources for Resource Groups  
Once you have defined resource groups, you further configure them by  
assigning cluster resources to one resource group or another. You can  
configure resource groups even if a node is powered down. However, SMIT  
cannot list possible shared resources for the node (making configuration  
errors likely).  
Note  
You cannot configure a resource group until you have completed the  
information on the Add a Resource Group screen.  
Note  
If you configure a cascading resource group with an NFS mount point, you  
must also configure the resource to use IP Address Takeover. If you do not  
do this, takeover results are unpredictable.You should also set the field  
value Filesystems Mounted Before IP Configured to true so that the  
takeover process proceeds correctly.  
Note  
When setting up a cascading resource with an IP Address takeover  
configuration, each cluster node should be configured in no more than  
(N+1) resource groups on a particular network. Here, N is the number of  
standby adapters on a particular node and network.  
The following describes the different possibilities for resources that might be  
added to a resource group:  
Table 20. Options Configuring Resources for a Resource Group  
Service IP Label  
If IP address takeover is being used, list the IP label to be  
moved when this resource group is taken over. Press F4  
to see a list of valid IP labels. These include addresses  
which rotate or may be taken over.  
HTY Service IP Label  
File Systems  
NTX adapters are not supported by HACMP for AIX 4.3  
Identify the file systems to include in this resource group.  
Press F4 to see a list of the file systems. When you enter  
a file system in this field, the HACMP for AIX software  
determines the correct values for the Volume Groups and  
Raw Disk PVIDs fields.  
HACMP Installation and Cluster Definition 109  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
Service IP Label  
If IP address takeover is being used, list the IP label to be  
moved when this resource group is taken over. Press F4  
to see a list of valid IP labels. These include addresses  
which rotate or may be taken over.  
File Systems  
Identify the method for checking consistency of file  
Consistency Check  
systems, fsck(default) or logredo(for fast recovery).  
File Systems Recovery Identify the recovery method for the file systems, parallel  
Method  
(for fast recovery) or sequential (default).  
Do not set this field to parallel if you have shared, nested  
file systems. These must be recovered sequentially. (Note  
that the cluster verification utility, clverify, does not report  
file system and fast recovery inconsistencies.)  
File Systems to Export  
Identify the file systems to be exported to include in this  
resource group. These should be a subset of the file  
systems listed above. Press F4 for a list.  
File Systems to  
NFS Mount  
Identify the subset of file systems to NFS mount. All nodes  
in the resource chain that do not currently hold the  
resource will attempt to NFS mount these file systems  
while the owner node is active in the cluster.  
These settings also have to be synchronized throughout the cluster.  
Therefore Synchronize Cluster Resources has to be chosen from the  
corresponding SMIT Menu.  
If the Cluster Manager is running on the local node, synchronizing cluster  
resources triggers a dynamic reconfiguration event (DARE, see 8.5.3, “DARE  
4.3.1.2 Configuring Run-Time Parameters  
There are two types of Run-Time Parameters for a node that can be chosen.  
One of them is the debug level, which can be switched from high to low,  
meaning all cluster manager actions are logged, or only errors are logged,  
respectively. The other is the differentiation whether the node uses NIS or  
DNS nameservice or not, to enable the cluster manager to turn that off in  
case it would interfere with its actions. Both of these parameters can be  
changed while the cluster is running.  
4.3.1.3 Defining Application Servers  
Application servers are another resource that can be configured into a  
Resource Group. They consist of a (hopefully meaningful) name, in order to  
enable the cluster manager to identify the application server uniquely, as well  
110  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
as the path locations for start and stop scripts for the application. These  
scripts have to be in the same location on every service node.  
Just as for pre- and post-events, these scripts can be adapted to specific  
nodes. They don’t need to be equal in content. The system administrator has  
to ensure, however, that they are in the same location, use the same name,  
and are executable for the root user.  
4.3.1.4 Synchronizing Cluster Resources  
After defining these resources and their relationship with the resource  
group, the act of synchronizing cluster resources sends the information  
contained on the current node to all defined cluster nodes.  
Note  
All configured nodes must be on their boot addresses when a cluster has  
been configured and the nodes are synchronized for the first time. Any  
node not on its boot address will not have its /etc/rc.net file updated with  
the HACMP entry; this causes problems for the reintegration of this node  
into the cluster.  
If a node attempts to join the cluster when its configuration is out-of-sync with  
other active cluster nodes, it will be denied. You must ensure that other nodes  
are synchronized to the joining member.  
4.4 Initial Testing  
After installing and configuring your cluster, it is recommended that you do  
some initial testing in order to verify that the cluster is acting as it should.  
4.4.1 Clverify  
Running /usr/sbin/cluster/diag/clverifyis probably a good start to the  
testing. It allows you to check the software and the cluster.  
Software checking is reduced to lpp checking, which is basically checking  
whether HACMP-specific modifications to AIX files are correct. For  
correctness of the installation itself, use the lppcheck -v command.  
Cluster verification is divided into topology and configuration checking. These  
two parts do basically the same as smit clverify, i.e. verifying that the  
clusters topology as well as the resource configurations are in sync on the  
cluster nodes.  
HACMP Installation and Cluster Definition 111  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
4.4.2 Initial Startup  
At this point in time, the cluster is not yet started. So the cluster manager has  
to be started first. To check whether the cluster manager is up, you can either  
look for the process with the pscommand:  
ps -ef | grep clstr  
or look for the status of the cluster group subsystems:  
lssrc -g cluster  
or look for the status of the network interfaces. If you have IP Address  
Takeover (IPAT) configured you should see that the network interface is on its  
boot address with the netstat -i command.  
Then start HACMP through smit clstart. In the panel that appears, choose  
the following parameters and press Enter:  
1. start now  
2. broadcast message true  
3. start cluster lock services false  
4. start cluster information daemon true  
Reissue either the pscommand (see above) or look for the interface state  
with the netstat -i command. Now, you should see that the boot interface is  
gone in favor of the service-interface.  
You also would like to check whether a takeover will work, so, you have to  
bring up HACMP on all cluster nodes through smitty clstartand check  
whether the cluster gets into a stable state. Use clstatfor this purpose.  
4.4.3 Takeover and Reintegration  
When the cluster is up and running, stop one of the node’s cluster managers  
with smitty clstopand choose graceful with takeover. One possibility to  
check whether the takeover went through smoothly is to look at the  
/tmp/hacmp.out file during the takeover, preferably on the takeover node. You  
can use the tail -f /tmp/hacmp.outcommand for this.  
After the cluster has become stable, you might check the netstat -ioutput  
again to verify that the takeover node has acquired the IP address of the  
“failed” node.  
112  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
For cascading resource groups the failed node is going to reaquire its  
resources, once it is up and running again. So, you have to restart HACMP on  
it through smitty clstartand check again for the logfile, as well as the  
clusters status.  
Further and more intensive debugging issues are covered in Chapter 7,  
4.5 Cluster Snapshot  
Now that the actual installation is finished, the cluster is well documented in  
the planning sheets, all information from there has been implemented in the  
HACMP ODM, and the cluster is verified and synchronized; provided the  
initial testing didn’t bring up any curiosities, you should save this working  
configuration in a cluster snapshot.  
The cluster snapshot utility allows you to save, in a file, a record of all the  
data that defines a particular cluster configuration. This facility gives you the  
ability to recreate a particular cluster configuration, a process called applying  
a snapshot, provided the cluster is configured with the requisite hardware and  
software to support the configuration.  
You can perform many of the cluster snapshot utility operations, such as  
saving a configuration and applying a saved configuration, using the HACMP  
for AIX VSM application (xhacmpm). For more information, see the  
administrative facilities chapter in the HACMP for AIX, Version 4.3: Concepts  
and Facilities, SC23-4276, and the online help information available with the  
application.  
In addition, a snapshot can provide useful information for troubleshooting  
cluster problems. Because the snapshots are simple ASCII files that can be  
sent via e-mail, they can make remote problem determination easier.  
You can also add your own custom snapshot methods to store additional  
user-specified cluster and system information in your snapshots. The output  
from these user-defined custom methods is reported along with the  
conventional snapshot information.  
Note  
You cannot use the cluster snapshot facility in a cluster with nodes  
concurrently running different versions of HACMP for AIX.  
HACMP Installation and Cluster Definition  
113  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
Essentially, a snapshot saves all the ODM classes HACMP has generated  
during its configuration. It does not save user customized scripts, such as  
start or stop scripts for an application server. However, the location and  
names of these scripts are in an HACMP ODM class, and are therefore  
saved. It is very helpful to put all the customized data in one defined place, in  
order to make saving these customizations easier. You can then use a  
custom snapshot method to save this data as well, by including a  
user-defined script in the custom snapshot.  
4.5.1 Applying a Cluster Snapshot  
Applying a cluster snapshot overwrites the data in the existing HACMP for  
AIX ODM classes on all nodes in the cluster with the new ODM data  
contained in the snapshot. You can apply a cluster snapshot from any cluster  
node. However, you have to differentiate between two possible states the  
cluster could be in when applying the snapshot.  
If cluster services are inactive on all cluster nodes, applying the snapshot  
changes the ODM data stored in the system default configuration directory  
(DCD). If cluster services are active on the local node, applying a snapshot  
triggers a cluster-wide dynamic reconfiguration event. In dynamic  
reconfiguration, in addition to synchronizing the ODM data stored in the  
DCDs on each node, HACMP for AIX replaces the current configuration data  
stored in the active configuration directory (ACD) with the changed  
configuration data in the DCD. The snapshot becomes the currently active  
configuration.  
Note  
A cluster snapshot used for dynamic reconfiguration may contain changes  
to either the cluster topology OR to cluster resources, but not both. You  
cannot change both the cluster topology and cluster resources in a single  
dynamic reconfiguration event.  
Note  
Applying a cluster snapshot may affect both AIX and HACMP for AIX ODM  
objects and system files as well as user-defined files.  
More detailed Information about Cluster Snapshot can be found in the  
HACMP for AIX, Version 4.3: Administration Guide, SC23-4279, Chapter 11  
as well as in the HACMP for AIX, Version 4.3: Troubleshooting Guide,  
SC23-4280.  
114  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
HACMP Installation and Cluster Definition  
115  
Download from Www.Somanuals.com. All Manuals Search And Download.  
116  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Chapter 5. Cluster Customization  
Within an HACMP for AIX cluster, there are several things that are  
customizable. The following paragraphs explain the customizing features for  
events, error notification, network modules and topology services.  
5.1 Event Customization  
An HACMP for AIX cluster environment acts upon a state change with a set  
of predefined cluster events (see 5.1.1, “Predefined Cluster Events” on page  
117). Whenever a state change is detected by the cluster manager, it decides  
which event will be started. It then executes the script for that event in a shell,  
as well as the subevents associated with it. These predefined events can be  
found under /usr/sbin/cluster/events.  
The HACMP for AIX software provides an event customization facility that  
allows you to tailor event processing to your site. This facility can be used to  
include the following types of customization:  
• Adding, changing, and removing custom cluster events  
• Pre- and post-event processing  
• Event notification  
• Event recovery and retry  
5.1.1 Predefined Cluster Events  
HACMP has the following predefined cluster events:  
5.1.1.1 Node Events  
This is the sequence of node_up events:  
node_up  
This event occurs when a node joins the cluster.  
Depending on whether the node is local or  
remote, this event initiates either a  
node_up_local or node_up_remote event.  
node_up_local  
This script acquires the service address (or  
shared address), gets all its owned (or shared)  
resources, and takes the resources. This  
includes making disks available, varying on  
volume groups, mounting file systems, exporting  
file systems, NFS-mounting file systems, and  
varying on concurrent access volumes groups.  
© Copyright IBM Corp. 1999  
117  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
acquire_service_addr  
acquire_takeover_addr  
(If configured for IP address takeover.)  
Configures boot addresses to the corresponding  
service address, and starts TCP/IP servers and  
network daemons by running the telinit -a  
command.  
The script checks to see if a configured standby  
address exists, then swaps the standby address  
with the takeover address.  
get_disk_vg_fs  
Acquires disk, volume group, and file system  
resources.  
node_up_remote  
Causes the local node to release all resources  
taken from the remote node and to place any  
concurrent volume groups in concurrent mode.  
Some of the scripts called by node_up_remote  
include the following:  
release_takeover_addr  
(If configured for IP address takeover.) Identifies  
a takeover address to be released because a  
standby adapter on the local node is  
masquerading as the service address of the  
remote node. Reconfigures the local standby  
adapter to its original address (and hardware  
address, if necessary).  
stop_server  
Stops application servers belonging to the  
reintegrating node.  
release_vg_fs  
Releases volume groups and file systems  
belonging to a resource group that the remote  
node will be taking over.  
cl_deactivate_nfs  
node_up_complete  
Unmounts NFS file systems.  
This event occurs only after a node_up event  
has successfully completed. Depending on  
whether the node is local or remote, this event  
initiates either a node_up_local_complete or  
node_up_remote_complete event.  
node_up_local_complete Calls the start_server script to start application  
servers. This event occurs only after a  
node_up_local event has successfully  
completed.  
node_up_remote_completeAllows the local node to do an NFS mount only  
after the remote node is completely up. This  
118  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
event occurs only after a node_up_remote event  
has successfully completed.  
Sequence of node_down Events  
node_down  
This event occurs when a node intentionally  
leaves the cluster or fails. Depending on  
whether the exiting node is local or remote, this  
event initiates either the node_down_local or  
node_down_remote event, which in turn initiates  
a series of subevents.  
node_down_local  
stop_server  
Processes the following events:  
Stops application servers.  
release_takeover_addr  
(If configured for IP address takeover.) Identifies  
a takeover address to be released because a  
standby adapter on the local node is  
masquerading as the service address of the  
remote node. Reconfigures the local standby  
with its original IP address (and hardware  
address, if necessary).  
release_vg_fs  
Releases volume groups and file systems that  
are part of a resource group the local node is  
serving.  
release_service_addr  
(If configured for IP address takeover.)  
Detaches the service address and reconfigures  
the service adapter to its boot address.  
node_down_remote  
Processes the following events:  
acquire_takeover_addr  
(If configured for IP address takeover.) Checks  
for a configured standby address currently seen  
as up by the Cluster Manager, and then does a  
standby_address to takeover_address swap  
(and hardware address, if necessary.  
get_disk_vg_fs  
Acquires disk, volume group, and file system  
resources as part of a takeover.  
node_down_complete  
This event occurs only after a node_down event  
has successfully completed. Depending on  
whether the node is local or remote, this event  
initiates either a node_down_local_complete or  
node_down_remote_complete event.  
Cluster Customization  
119  
Download from Www.Somanuals.com. All Manuals Search And Download.  
node_down_local_completeInstructs the Cluster Manager to exit when the  
local node has left the cluster. This event occurs  
only after a node_down_local event has  
successfully completed.  
node_down_remote_completeStarts takeover application servers. This  
event runs only after a node_down_remote  
event has successfully completed.  
start_server  
Starts application servers.  
5.1.1.2 Network Events  
network_down  
This event occurs when the Cluster Manager  
determines a network has failed. A  
network_down event can take one of two forms:  
Local network_down, where only a particular  
node has lost contact with a network.  
Global network_down, where all of the nodes  
connected to a network have lost contact with a  
network. It is assumed in this case that a  
network-related failure has occurred rather than  
a node-related failure.  
The network_down event mails a notification to  
the system administrator, but takes no further  
action since appropriate actions depend on the  
local network configuration.  
network_down_complete This event occurs only after a network_down  
event has successfully completed. The default  
network_down_complete event processing  
takes no actions since appropriate actions  
depend on the local network configuration.  
network_up  
This event occurs when the Cluster Manager  
determines a network has become available for  
use. The default network_up event processing  
takes no actions since appropriate actions  
depend on the local network configuration.  
network_up_complete  
This event occurs only after a network_up event  
has successfully completed. The default  
network_up_complete event processing takes  
120  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
no actions since appropriate actions depend on  
the local network configuration.  
5.1.1.3 Network Adapter Events  
swap_adapter  
This event occurs when the service adapter on a  
node fails. The swap_adapter event exchanges  
or swaps the IP addresses of the service and a  
standby adapter on the same HACMP network  
and then reconstructs the routing table.  
swap_adapter_complete This event occurs only after a swap_adapter  
event has successfully completed. The  
swap_adapter_complete event ensures that the  
local ARP cache is updated by deleting entries  
and pinging cluster IP addresses.  
fail_standby  
join_standby  
This event occurs if a standby adapter fails or  
becomes unavailable as the result of an IP  
address takeover. The fail_standby event  
displays a console message indicating that a  
standby adapter has failed or is no longer  
available.  
This event occurs if a standby adapter becomes  
available. The join_standby event displays a  
console message indicating that a standby  
adapter has become available.  
5.1.1.4 Cluster Status Events  
config_too_long  
This event occurs when a node has been in  
reconfiguration for more than six minutes. The  
event periodically displays a console  
message.  
reconfig_topology_start  
This event marks the beginning of a dynamic  
reconfiguration of the cluster topology.  
reconfig_topology_completeThis event indicates that a cluster topology  
dynamic reconfiguration has completed.  
reconfig_resource_acquire This event indicates that cluster resources  
that are affected by dynamic reconfiguration  
are being acquired by appropriate nodes.  
reconfig_resource_release This event indicates that cluster resources  
affected by dynamic reconfiguration are being  
released by appropriate nodes.  
Cluster Customization  
121  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
reconfig_resource_completeThis event indicates that a cluster resource  
dynamic reconfiguration has completed.  
5.1.2 Pre- and Post-Event Processing  
To tailor event processing to your environment, specify commands or  
user-defined scripts that should execute before and/or after a specific event  
is generated by the Cluster Manager. You specify them by selecting the  
HACMP event to be customized on the smit hacmp -> Cluster  
Configuration -> Resources -> Cluster Events -> Change/Show Cluster  
Events screen, and then, choosing the one to be tailored. Now, you can  
enter the location of your pre- or post-event to be executed before or after the  
chosen event has been processed.  
For preprocessing, for example, you may want to send a message to specific  
users informing them to stand by while a certain event occurs. For  
post-processing, you may want to disable login for a specific group of users if  
a particular network fails.  
5.1.3 Event Notification  
You can specify a command or user-defined script that provides notification  
(for example, mail) that an event is about to happen and that an event has  
just occurred, along with the success or failure of the event.  
This is done on the very same SMIT screen as in 5.1.2, “Pre- and Post-Event  
Processing” on page 122 in the NotifyCommand field.  
For example, a site may want to use a network_downnotification event to  
inform system administrators that traffic may have to be rerouted. Afterwards,  
you can use a network_upnotification event to tell system administrators that  
traffic can again be serviced through the restored network.  
Event notification in an HACMP cluster can also be done using pre- and  
post-event scripts, just by adding the script you want to execute for  
notification into the pre-and/or post-eventcommand script.  
5.1.4 Event Recovery and Retry  
You can specify a command that attempts to recover from an event command  
failure. If the retry count is greater than zero, and the recovery command  
succeeds, the event script command is rerun. You can also specify the  
number of times to attempt to execute the recovery command.  
122  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
For example, a file system cannot be unmounted, because of a process  
running on it. Then, you might want to kill that process first, before  
unmounting the file system, in order to get the event script done. Now, since  
the event script didn’t succeed in its first run, the Retry feature enables  
HACMP for AIX to retry it until it finally succeeds, or the retry count is  
reached.  
5.1.5 Notes on Customizing Event Processing  
You must declare a shell (for example #!/bin/sh) at the beginning of each  
script executed by the notify, recovery, and pre- or post-event processing  
commands.  
Notify, recovery, and pre- and post-event processing do not occur when the  
force option of the node_downevent is specified.  
Synchronizing the cluster configuration does not propagate the actual new or  
changed scripts; you must add these to each node manually. Also, it is  
allowed to have different contents in these scripts on different nodes in order  
to be able to act upon different environments. However, the name of these  
scripts, their location in the file system, and their permission bits have to be  
identical.  
5.1.6 Event Emulator  
To test the effect of running an event on your cluster, HACMP for AIX  
provides a utility to run an emulation of an event. This emulation lets you  
predict a cluster's reaction to an event as though the event actually occurred.  
The emulation runs on all active nodes in your cluster, and the output is  
stored in an output file. You can select the path and name of this output file  
using the EMU_OUTPUTenvironment variable, or, use the default  
/tmp/emuhacmp.out file on the node that invoked the Event Emulator.  
For more information on event emulation, see these chapters: “Administrative  
Facilities” in the HACMP for AIX, Version 4.3: Concepts and Facilities,  
SC23-4276, and “Monitoring an HACMP Cluster” in the HACMP for AIX,  
Version 4.3: Administration Guide, SC23-4279.  
5.2 Error Notification  
The AIX Error Notification facility detects errors matching predefined  
selection criteria and responds in a programmed way. The facility provides a  
wide range of criteria that you can use to define an error condition. These  
errors are called notification objects.  
Cluster Customization  
123  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Each time an error is logged in the system error log, the error notification  
daemon determines if the error log entry matches the selection criteria. If it  
does, an executable is run. This executable, called a notify method, can  
range from a simple command to a complex program. For example, the notify  
method might be a mail message to the system administrator or a command  
to shut down the cluster.  
Using the Error Notification facility adds an additional layer of high availability  
to the HACMP for AIX software. Although the combination of the HACMP for  
AIX software and the inherent high availability features built into the AIX  
operating system keeps single points of failure to a minimum, failures still  
exist that, although detected, are not handled in a useful way.  
Take the example of a cluster where an owner node and a takeover node  
share an SCSI disk. The owner node is using the disk. If the SCSI adapter on  
the owner node fails, an error may be logged, but neither the HACMP for AIX  
software nor the AIX Logical Volume Manager responds to the error. If the  
error has been defined to the Error Notification facility, however, an  
executable that shuts down the node with the failed adapter could be run,  
allowing the surviving node to take over the disk.  
5.3 Network Modules/Topology Services and Group Services  
The HACMP for AIX SMIT interface allows you to add, remove, or change an  
HACMP for AIX network module. You rarely need to add or remove any of  
those, however, you may want to change the failure detection rate of a  
network module.  
There are three values to choose from: Fast, Normal and Slow. The normal  
heartbeat rate is usually optimal. Speeding up or slowing down failure  
detection is an area where you can adjust cluster failover behavior.  
If you decide to change the failure detection rate of a network module, keep  
the following considerations in mind:  
• Failure detection is dependent on the fastest network linking two nodes.  
• Faster heartbeat rates may lead to false failure detections, particularly on  
busy networks. For example, bursts of high network traffic may delay  
heartbeats and this may result in nodes being falsely ejected from the  
cluster. Faster heartbeat rates also place a greater load on networks.  
• If your networks are very busy and you experience false failure detections,  
you can try changing the failure detection speed on the network modules  
to slow to avoid this problem.  
124  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
The failure rate of networks varies, depending on their characteristics. For  
example, for an Ethernet, the normal failure detection rate is two keepalives  
per second; fast is about four per second; slow is about one per second. For  
an HPS network, because no network traffic is allowed when a node joins the  
cluster, normal failure detection is 30 seconds; fast is 10 seconds; slow is 60  
seconds.  
The Change / Show Topology and Group Services Configuration screen  
includes the settings for the length of the Topology and Group services logs.  
The default settings are highly recommended. The screen also contains  
entries for heartbeat settings, but these are not operable (see HACMP/ES  
Installation and Administration Guide, SC23-4284, Chapter 18). The  
heartbeat rate is now set for each network module in the corresponding  
screen (see above).  
To learn more about Topology and Group Services, see Chapter 32 of the  
HACMP/ES Installation and Administration Guide, SC23-4284.  
5.4 NFS considerations  
For NFS to work correctly in an HACMP cluster environment, you have to  
take care of some special NFS characteristics.  
The HACMP scripts have only minimal NFS support. You may need to modify  
them to handle your particular configuration. The following sections contain  
some suggestions for handling a variety of issues.  
5.4.1 Creating Shared Volume Groups  
When creating shared volume groups, normally, you can leave the Major  
Number field blank and let the system provide a default for you. However,  
unless all nodes in your cluster are identically configured, you will have  
problems using NFS in an HACMP environment. The reason is that the  
system uses the major number as part of the file handle to uniquely identify a  
Network File System.  
In the event of node failure, NFS clients attached to an HACMP cluster  
operate exactly the way they do when a standard NFS server fails and  
reboots. If the major numbers are not the same, when another cluster node  
takes over the file system and re-exports it, the client application will not  
recover, since the file system exported by the node will appear to be different  
from the one exported by the failed node.  
Cluster Customization  
125  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
To prevent problems with NFS file systems in an HACMP cluster, make sure  
that each shared volume group has the same major number on all nodes. The  
lvlstmajorcommand lists the free major numbers on a node. Use this  
command on each node to find a major number that is free on all cluster  
nodes, then, record that number in the Major Number field on the Shared  
Volume Group/File System (Non-Concurrent Access) worksheet in Appendix  
A, Planning Worksheets, of the HACMP for AIX, Version 4.3: Planning Guide,  
SC23-4277 for a non-concurrent access configuration.  
Alternatively, if you use the Task Guide to create your shared volume groups,  
it will make sure that the major number is the same on all nodes that will  
share it.  
5.4.2 Exporting NFS File Systems  
The default scripts provided with HACMP do not use the /etc/exports file.  
Instead, the default scripts provided call a cl_export_fsutility that uses the  
exportfscommand with the -i flag and specifies the file system names stored  
in the HACMP ODM object class.  
Therefore export options specified in the /etc/exports file are ignored.  
However, export options may be specified by modifying the cl_export_fs  
utility. Alternately, the /etc/exports file can be used as is typical in an NFS  
environment by simply removing the -iflag from the exportfscommand in  
the cl_export_fs utility.  
5.4.3 NFS Mounting  
For HACMP and NFS to work together properly, you must be aware of the  
following mount issues:  
5.4.3.1 Creating NFS Mount Points on Clients  
A mount point is required in order to mount a file system with NFS. Mount  
points are required for NFS clients, not servers; however, you should be  
aware that a server can also be a client.  
5.4.4 Cascading Takeover with Cross Mounted NFS File Systems  
This section describes how to set up cascading resource groups with cross  
mounted NFS file systems.  
5.4.4.1 Server-to-Server NFS Cross Mounting  
HACMP allows you to configure a cluster so that servers can NFS-mount  
each other’s file systems. The following figure shows an example:  
126  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Figure 14. NFS Cross Mounts  
When Node A fails, Node B uses the cl_nfskillutility to close open files in  
Node A:/afs, unmounts it, mounts it locally, and re-exports it to waiting clients.  
After takeover, Node B has:  
/bfs locally mounted  
/bfs nfs-exported  
/afs locally mounted  
/afs nfs-exported  
Ensure that the shared volume groups have the same major number on the  
server nodes. This allows the clients to re-establish the NFS-mount  
transparently after the takeover.  
Caveats about Node Names and NFS  
In the configuration described above, the node name is used as the NFS  
hostname for the mount. This can fail if the node name is not a legitimate  
TCP/IP adapter label.  
To avoid this problem do one of the following:  
Cluster Customization  
127  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
• Ensure that node name and the service adapter label are the same on  
each node in the cluster  
or  
• Alias the node name to the service adapter label in the /etc/hosts file.  
5.4.5 Cross Mounted NFS File Systems and the Network Lock Manager  
If an NFS client application uses the Network Lock Manager, there are  
additional considerations to ensure a successful failover. Consider the  
following scenario: Node A has a file system mounted locally and exported for  
use by clients. Node B is an NFS client and mounts the exported file system  
for local use by an application that issues lock requests using the flock()  
system call. Node A fails. Node B then attempts to unmount the NFS  
mounted file system, mount it as a local file system, and export it for client  
use. However, the unmount fails because of outstanding lock requests  
against the file system.  
Adding the following lines to the cl_deactivate_nfs script will clear  
outstanding locks against the failed node and will allow the file system to be  
unmounted. However, it will result in the loss of all locks. Consider your  
configuration carefully. If you have non-cluster related NFS file systems  
where losing locks would be unacceptable, you may need to take appropriate  
steps before using this addition to the cl_deactivate_nfs script.  
Add the code below between the following two lines (three places):  
128  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
######## Add for NFS Lock Removal (start) ########  
######## Add for NFS Lock Removal (finish) ########  
###############################################################################  
#
# Name: cl_deactivate_nfs  
#
# Given a list of nfs-mounted filesystems, we try and unmount -f  
# any that are currently mounted.  
#
# Arguments: list of filesystems.  
#
###############################################################################  
PROGNAME="$0"  
MOUNTED="false"  
######## Add for NFS Lock Removal (start) ########  
STOPPED="false"  
######## Add for NFS Lock Removal (finish) ########  
SLEEP="2"  
if [ "$VERBOSE_LOGGING" = "high" ]  
then  
set -x  
fi  
set -u  
if [ $# -ne 0 ]  
then  
FILELIST=‘for i in $*; do /bin/echo $i; done | /bin/sort -r‘  
for fs in $FILELIST  
do  
# Is the filesystem mounted?  
# -s says only return status, -x says exact match  
# we use awk instead of cut because mount outputs  
# lots of leading blanks that confuse cut  
/etc/mount | awk ’{ print $2 }’ | fgrep -s -x "$fs"  
if [ $? -eq 0 ]  
then  
# At least one filesystem is mounted  
MOUNTED="true"  
# This filesystem is mounted  
######## Add for NFS Lock Removal (start) ########  
# Determine the host which is making the filesystem  
# available  
# This will clear any outstanding locks against the  
# failed node, not preserve their state, and is thus  
# considered a forceful move.  
host=‘/etc/mount|grep nfs|grep "$fs"|awk ’{ print $1 }’‘  
if [ -n "$host" ]  
then  
if [ "$STOPPED" = "false" ]  
then  
stopsrc -s rpc.lockd  
stopsrc -s rpc.statd  
STOPPED="true"  
Cluster Customization  
129  
Download from Www.Somanuals.com. All Manuals Search And Download.  
fi  
/bin/rm -f /etc/sm.bak/$host  
/bin/rm -f /etc/sm/$host  
/bin/rm -f /etc/state  
fi  
######## Add for NFS Lock Removal (finish) ########  
# Send a SIGKILL to all processes having open file  
# descriptors within this logical volume to allow  
# the unmount to succeed..  
cl_nfskill -k -u $fs  
fi  
done  
else  
cl_echo 27 ’$PROGNAME: Bad number of arguments ’ $PROGNAME  
exit 2  
fi  
# Make sure all processes have time to die  
# Only wait if at least one filesystem is mounted  
if [ "$MOUNTED" = "true" ]  
then  
sleep $SLEEP  
fi  
FILELIST=‘for i in $*; do /bin/echo $i; done | /bin/sort -r‘  
for fs in $FILELIST  
do  
# Is the filesystem mounted?  
# -s says only return status, -x says exact match  
# we use awk instead of cut because mount outputs  
# lots of leading blanks that confuse cut  
/etc/mount | awk ’{ print $2 }’ | fgrep -s -x "$fs"  
if [ $? -eq 0 ]  
then  
# At least one filesystem is mounted  
until /etc/umount -f $fs  
do  
sleep 2  
done  
fi  
done  
######## Add for NFS Lock Removal (start) ########  
if [ "$STOPPED" = "true" ]  
then  
startsrc -s rpc.statd  
startsrc -s rpc.lockd  
fi  
######## Add for NFS Lock Removal (finish) ########  
exit 0  
130  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Chapter 6. Cluster Testing  
Before you start to test the HACMP configuration, you need to guarantee that  
your cluster nodes are in a stable state. Check the state of the:  
• Devices  
• System parameters  
• Processes  
• Network adapters  
• LVM  
• Cluster  
• Other items such as SP Switch, printers, and SNA configuration  
6.1 Node Verification  
Here is a series of suggested actions to test the state of a node before  
including HACMP in the testing.  
6.1.1 Device State  
• Run diag -ain order to clean up the VPD.  
• Look in the errorlog for unusual errors by issuing the command errpt |  
moreor errpt -a | more.  
• Check that all devices are in the available state (lsdev -C | more).  
• Check that the SCSI addresses of adapters on shared buses are unique  
(lsattr -E -l ascsi0).  
• If you are using target mode SCSI networks, check the connection by  
issuing cat < /dev/tmscsi#.tmon the first node and cat /etc/hosts >  
/dev/tmscsi#.im(enter twice!) on the second node where #is the  
appropriate tmscsi device number. Repeat the test in the other direction.  
Note that cluster services must be stopped on both nodes to perform this  
test.  
To check a serial line between two nodes type stty < /dev/tty#on both  
nodes where #is the appropriate tty device number for the RS232  
heartbeat connection. Note that cluster services must be stopped on both  
nodes to perform this test.  
© Copyright IBM Corp. 1999  
131  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
6.1.2 System Parameters  
• Type dateon all nodes to check that all the nodes in the cluster are  
running with their clocks on the same time.  
• Ensure that the number of user licenses has been correctly set  
(lslicense).  
• Check high water mark and other system settings (smitty chgsys).  
• Type sysdumpdev -land sysdumpdev -eto ensure that the dump space is  
correctly set and that the primary dump device (lslv hd7) is large enough  
to accomodate a dump.  
• Check that applications to be controlled by HACMP are not started here,  
and that extraneous processes which might interfere with HACMP and/or  
dominate system resources are not started (more /etc/inittab).  
• Check list of cron jobs (crontab -l).  
6.1.3 Process State  
• Check the paging space usage by issuing lsps -a.  
• Look for all expected processes with ps -ef | more.  
• Check that the run queue is < 5 and that the CPU usage is at an  
acceptable level (vmstat 2 5).  
6.1.4 Network State  
• Type for example ifconfig lo0, ifconfig en0and ifconfig en1 to check the  
network adapter configuration, if you are using ethernet adapters. For  
other types of adapters, use the appropriate device name.  
To check the configuration of an SP Switch adapter, type:  
/usr/lpp/ssp/css/ifconfig css0.  
• Use netstat -ior netstat -into show the network configuration of the  
node.  
To check the alternate MAC address, issue netstat -v ent0 | more.  
• Look at mbufs sizing relative to requests for memory denied (netstat -m |  
more).  
• Type netstat -ror netstat -rAnto ensure that there are valid routes to the  
other cluster node interfaces and to clients.  
• Run no -a | moreand look at the setting of ipforwardingand  
ipsendredirects.  
132  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
• Check that all interfaces communicate (ping <ip-address> or ping -R  
<ip-address>).  
• List the arp table entries with arp -a.  
• Check the status of the TCP/IP daemons (lssrc -g tcpip).  
• Ensure that there are no bad entries in the /etc/hosts file, especially at the  
bottom of the file.  
• Verify that, if DNS is in use, the DNS servers are correctly defined (more  
/etc/resolv.conf).  
• Check the status of NIS by typing ps -ef | grep ypbindand lssrc -g yp.  
• The command exportfsshows non-HACMP controlled NFS exports.  
• Run snmpinfo -m dump -o /usr/sbin/cluster/hacmp.defs address show  
snmp information for Cluster network addresses (including the serial  
interfaces).  
6.1.5 LVM State  
• Ensure that the correct VG's are defined, that quorum and auto-varyon are  
correctly defined, and that the shared VG's are in the correct state (lsvg  
and lsvg -o).  
• Check that there are no stale partitions (lsvg -l).  
• Check that all appropriate file systems have been mounted and that none  
of the rootvg file systems are full (df -k).  
• Check that PVid's have been assigned where necessary and that there  
are no ghost disks(lspv).  
• Verify that all entries in the /etc/filesystems file are correct and that there  
are no erroneous entries (more /etc/filesystemsand lsfs).  
6.1.6 Cluster State  
• Check the status of the cluster daemons by issuing lssrc -g clusterand  
lssrc -g lock.  
• Run /usr/sbin/cluster/clstatto check the status of the cluster and the  
status of the network interfaces.  
• Check the cluster logfiles with tail -f /tmp/hacmp.out, more  
/usr/sbin/cluster/history/cluster.mmdd(mmdd = current date), tail -f  
/var/adm/cluster.logand more /tmp/cm.log.  
• Check that the nodename is correct (odmget HACMPcluster).  
Cluster Testing 133  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
• Verify the cluster configuration by running /usr/sbin/cluster/diag/clconfig  
-v ’-tr’.  
To show cluster configuration, run: /usr/sbin/cluster/utilities/cllscf.  
To show the clstrmgr version, type: snmpinfo -m dump -o  
/usr/sbin/cluster/hacmp.defs clstrmgr.  
6.2 Simulate Errors  
The following paragraphs will give you hints on how you can simulate  
different hardware and software errors in order to verify your HACMP  
configuration. As an example, we will use a cluster consisting of two nodes  
and a cascading resource group definition.  
The term “NodeF” is used for the node to be failed, and the term “NodeT” for  
the takeover node of NodeF's resource group.  
When executing the test plan, it is helpful to monitor cluster activities during  
failover with the following commands. Note that the /tmp/hacmp.out file is the  
most useful to monitor, especially if the Debug Level of the HACMP Run Time  
Parameters for the nodes has been set to “high”, and if the Application Server  
Scripts include the set -xflag and periodic echo commands.  
6.2.1 Adapter Failure  
The following sections cover adapter failure.  
6.2.1.1 Ethernet or Token Ring Interface Failure  
In case of an Ethernet or Token Ring interface failure, perform the following  
steps:  
• Check that all the nodes in the cluster are up and running.  
• Optional: Prune the error log on NodeF (errclear 0).  
• Monitor the cluster log files on NodeT.  
• Use ifconfig to shut off the appropriate service interface (but not the  
Administrative SP Ethernet) on NodeF (for example, ifconfig en0 down).  
This will cause the service IP address to failover to the standby adapter on  
NodeF.  
• Verify that the swap adapter has occurred (including MAC  
Addressfailover) and that HACMP has turned the original service interface  
back on as the standby interface.  
134  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
• Use ifconfig to swap the service address back to the original service  
interface back (ifconfig en1 down). This will cause the service IP address  
to failover back to the service adapter on NodeF.  
6.2.1.2 Ethernet or Token Ring Adapter or Cable Failure  
Perform the following steps in the event of an Ethernet or Token Ring adapter  
or cable failure:  
• Check, by way of the verification commands, that all the Nodes in the  
cluster are up and running.  
• Optional: Prune the error log on NodeF (errclear 0).  
• Monitor the cluster log files on NodeT.  
• Disconnect the network cable from the appropriate service interface (but  
not the Administrative SP Ethernet) on NodeF. This will cause the service  
IP and MAC addresses to failover to the standby adapter on NodeF.  
• Verify that the swap adapter has occurred.  
• Reconnect the network cable to the service interface. This will cause the  
original service interface to become the standby interface.  
• Initiate a swap adapter back to the original service interface by  
disconnecting the network cable from the new service interface (originally  
the standby interface). This will cause the service IP and MAC addresses  
to failover back to the service adapter on NodeF.  
• Verify that the swap adapter has occurred.  
• Reconnect the cable to the original standby interface.  
• Verify that the original standby interface is operating with the standby IP  
address.  
6.2.1.3 Switch Adapter Failure  
Perform the following steps in case of switch adapter failure:  
Note  
Do not disconnect live switch cables to simulate a switch failure!  
• Check, by way of the verification commands, that all the Nodes in the  
cluster are up and running.  
• Assign NodeF to be Eprimary.  
• Optional: Prune the error log on NodeF (errclear 0).  
• Monitor the cluster log files on NodeT.  
Cluster Testing 135  
Download from Www.Somanuals.com. All Manuals Search And Download.  
• Generate the switch error in the error log which is being monitored by  
HACMP Error Notification (for configuration see 2.6.2.1, “Single  
network_down event has been customized, bring down css0 (ifconfig  
css0 down) or fence out NodeF from the Control Workstation (Efence NodeF).  
• If the first failure simulation method is used, the switch failure will be  
detected in the error log (errpt -a | more) on NodeF and cause a node  
failover to NodeT. The other two methods will cause HACMP to detect a  
network_down event, with the same result. (Note that if there is another  
node in the cluster with a lower alphanumeric node name than NodeT,  
then that node will become Eprimary. HACMP does not take care of the  
Eprimary if a new SP-Switch is used).  
• Verify that failover has occurred (netstat -iand pingfor networks, lsvg -o  
and viof a test file for volume groups, ps -U <appuid>for application  
processes, and Eprimaryfor Eprimary).  
• Start HACMP on NodeF (smit clstart). NodeT will release NodeF's  
cascading Resource Groups and NodeF will take them back over, but  
NodeT (or a lower alphanumeric node) will remain Eprimary.  
• Verify that re-integration has occurred (netstat -iand pingfor networks,  
lsvg -oand viof a test file for volume groups, ps -U <appuid> for  
application processes, and Eprimaryfor Eprimary).  
6.2.1.4 Failure of a 7133 Adapter  
Perform the following steps in the event of a 7133 Adapter failure:  
• Check, by way of the verification commands, that all the Nodes in the  
cluster are up and running.  
• Optional: Prune the error log on NodeF (errclear 0).  
• Monitor cluster logfiles on NodeT if HACMP has been customized to  
monitor 7133 disk failures.  
• Pull all the cable from the SSA adapter.  
• The failure of the 7133 adapter should be detected in the error log (errpt  
-a | more) on NodeF or should be noted in the appropriate diagnostics  
tool, and the logical volume copies on the disks in drawer 1 will be marked  
stale (lsvg -l NodeFvg).  
• Verify that all sharedvg file systems and paging spaces are accessible (df  
-kand lsps -a).  
• Re-attach the cables.  
136  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
• Verify that all sharedvg file systems and paging spaces are accessible (df  
-kand lsps -a).  
6.2.2 Node Failure / Reintegration  
The following sections deal with issues of node failure and reintegration.  
6.2.2.1 AIX Crash  
Perform the following steps in the event of an AIX crash:  
• Check, by way of the verification commands, that all the Nodes in the  
cluster are up and running.  
• Optional: Prune the error log on NodeF (errclear 0).  
• If NodeF is an SMP, you may want to set the fast reboot switch (mpcfg -cf  
11 1).  
• Monitor cluster logfiles on NodeT.  
• Crash NodeF by entering cat /etc/hosts > /dev/kmem. (The LED on NodeF  
will display 888.)  
• The OS failure on NodeF will cause a node failover to NodeT.  
• Verify that failover has occurred (netstat -iand pingfor networks, lsvg -o  
and viof a test file for volume groups, and ps -U <appuid> for application  
processes).  
• Power cycle NodeF. If HACMP is not configured to start from /etc/inittab,  
(on restart) start HACMP on NodeF (smit clstart). NodeF will take back  
its cascading Resource Groups.  
• Verify that re-integration has occurred (netstat -iand pingfor networks,  
lsvg -oand viof a test file for volume groups, and ps -U <appuid> for  
application processes).  
6.2.2.2 CPU Failure  
Perform the following steps in the event of CPU failure:  
• Check, by way of the verification commands, that all the Nodes in the  
cluster are up and running.  
• Optional: Prune the error log on NodeF (errclear 0).  
• If NodeF is an SMP, you may want to set the fast reboot switch (mpcfg -cf  
11 1).  
• Monitor cluster logfiles on NodeT.  
• Power off NodeF. This will cause a node failover to NodeT.  
Cluster Testing 137  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
• Verify that failover has occurred (netstat -iand pingfor networks, lsvg -o  
and viof a test file for volume groups, and ps -U <appuid> for application  
processes).  
• Power cycle NodeF. If HACMP is not configured to start from /etc/inittab  
(on restart), start HACMP on NodeF (smit clstart). NodeF will take back  
its cascading Resource Groups.  
• Verify that re-integration has occurred (netstat -iand pingfor networks,  
lsvg -oand viof a test file for volume groups, and ps -U <appuid> for  
application processes).  
6.2.2.3 TCP/IP Subsystem Failure  
• Check, by way of the verification commands, that all the Nodes in the  
cluster are up and running.  
• Optional: Prune the error log on NodeF (errclear 0).  
• Monitor the cluster log files on NodeT.  
• On NodeF, stop the TCP/IP subsystem (sh /etc/tcp.clean) or crash the  
subsystem by increasing the size of the sb_max and thewall parameters to  
large values (no -o sb_max=10000; no -o thewall=10000) and ping NodeT.  
Note that you should record the values for sb_max and thewall prior to  
modifying them, and, as an extra check, you may want to add the original  
values to the end of /etc/rc.net.  
• The TCP/IP subsystem failure on NodeF will cause a network failure of all  
the TCP/IP networks on NodeF. Unless there has been some  
customization done to promote this type of failure to a node failure, only  
the network failure will occur. The presence of a non-TCP/IP network  
(RS232, target mode SCSI or target mode SSA) should prevent the cluster  
from triggering a node down in this situation.  
• Verify that the network_down event has been run by checking the  
/tmp/hacmp.out file on either node. By default, the network_down script  
does nothing, but it can be customized to do whatever is appropriate for  
that situation in your environment.  
• On NodeF, issue the command startsrc -g tcpip. This should restart the  
TCP/IP daemons, and should cause a network_up event to be triggered in  
the cluster for each of your TCP/IP networks.  
6.2.3 Network Failure  
• Check, by way of the verification commands, that all the Nodes in the  
cluster are up and running.  
• Optional: Prune the error log on NodeF (errclear 0).  
138  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
• Monitor the cluster log files on NodeT.  
• Disconnect the network cable from the appropriate service and all the  
standby interfaces at the same time (but not the Administrative SP  
Ethernet) on NodeF. This will cause HACMP to detect a network_down  
event.  
• HACMP triggers events dependent on your configuration of the  
network_down event. By default, no action is triggered by the  
network_down event.  
• Verify that the expected action has occurred.  
6.2.4 Disk Failure  
The following sections deal with issues of disk failure.  
6.2.4.1 Mirrored rootvg Disk (hdisk0) Failure  
Perform the following steps in case of mirrored rootvg disk (hdisk0) failure:  
• Check, by way of the verification commands, that all the Nodes in the  
cluster are up and running.  
• Verify that the bootlist contains hdisk0 and hdisk1, if for example, hdisk1 is  
the mirror of hdisk0 (bootlist -m normal -o).  
• Optional: Prune the error log on NodeF (errclear 0).  
• Monitor cluster logfiles on NodeT if HACMP has been customized to  
monitor SCSI disk failures.  
• Slide back cover/casing on NodeF to get access to hdisk0 (this may first  
require turning the key to service mode). Pull the power cable (several  
colored wires with a white plastic connector) from the rear of hdisk0 (the  
lower internal disk is hdisk0, and the upper internal disk is hdisk1 on most  
systems). If you have a hot-pluggable disk, just pull the disk out of the  
frame.  
• The failure of hdisk0 should be detected in the error log (errpt -a | more)  
on NodeF.  
• Verify that all rootvg file systems and paging spaces are accessible (df;  
lsps -a).  
• Shutdown (smit clstop; shutdown -F) and power off NodeF.  
• Turn key to normal mode, power on NodeF, and verify that the system  
boots correctly. Log in and verify that all the rootvg file systems have been  
mounted (df).  
• Shutdown (shutdown -F) and power off NodeF.  
Cluster Testing 139  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
• Reconnect hdisk0, close the casing, and turn the key to normal mode.  
• Power on NodeF then verify that the rootvg logical volumes are no longer  
stale (lsvg -l rootvg).  
6.2.4.2 7135 Disk Failure  
Perform the following steps in the event of a disk failure:  
• Check, by way of the verification commands, that all the Nodes in the  
cluster are up and running.  
• Optional: Prune the error log on NodeF (errclear 0).  
• Monitor cluster logfiles on NodeT if HACMP has been customized to  
monitor 7135 disk failures.  
• Mark a shared disk failed through smit (smit raidiant; RAIDiant Disk  
Array Manager -> Change/Show Drive Status -> select the appropriate  
hdisk -> select the appropriate physical disk -> F4 to select a Drive  
Status of 83 Fail Drive), or if the disk is hot pluggable, remove the disk.  
• The amber light on the front of the 7135 comes on, and can also be seen  
in SMIT (smit raidiant; RAIDiant Disk Array Manager -> List all SCSI  
RAID Arrays).  
• Verify that all sharedvg file systems and paging spaces are accessible (df  
and lsps -a).  
• If using RAID5 with Hot Spare, verify that reconstruction has completed to  
the Hot Spare, then un-mark or plug the failed disk back in. If using  
RAID1, sync the volume group (syncvg NodeFvg).  
• If using RAID5 without Hot Spare, mark the failed disk Optimal (smit  
raidiant; RAIDiant Disk Array Manager -> Change/Show Drive Status;  
select the appropriate hdisk -> select the appropriate physical disk ->  
F4 to select a Drive Status of 84 Replace and Reconstruct Drive).  
• Verify that the reconstruction has completed (smit raidiant;RAIDiant Disk  
Array Manager -> List all SCSI RAID Arrays).  
• Verify that all sharedvg file systems and paging spaces are accessible (df  
and lsps -a) and that the partitions are not stale (lsvg -l sharedvg). Also  
verify that the yellow light has turned off on the 7135.  
6.2.4.3 Mirrored 7133 Disk Failure  
• Check, by way of the verification commands, that all the Nodes in the  
cluster are up and running.  
• Optional: Prune the error log on NodeF (errclear 0).  
140  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
• Monitor cluster logfiles on NodeT if HACMP has been customized to  
monitor 7133 disk failures.  
• Since the 7133 disk is hot pluggable, remove a disk from drawer 1  
associated with NodeF's shared volume group.  
• The failure of the 7133 disk will be detected in the error log (errpt -a |  
more) on NodeF, and the logical volumes with copies on that disk will be  
marked stale (lsvg -l NodeFvg).  
• Verify that all NodeFvg file systems and paging spaces are accessible (df  
-k and lsps -a).  
• Plug the failed disk back in, then sync the volume group (syncvg NodeFvg).  
• Verify that all NodeFvg file systems and paging spaces are accessible (df  
-kand lsps -a) and that the partitions are not stale (lsvg -l NodeFvg).  
6.2.5 Application Failure  
By default, HACMP does not recognize application failures. With some  
additional configuration, it is possible to “teach” HACMP application failures  
and trigger events (for more information see 2.6.2.3, “Application Failure” on  
page 47).  
So, the way of testing application failures is strongly dependent on your  
configuration.  
Before you start to do the configuration and the testing application failure  
notification, analyse your application for possible failures. Then try to  
reproduce them.  
Cluster Testing 141  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
142  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Chapter 7. Cluster Troubleshooting  
Typically, a functioning HACMP cluster requires minimal intervention. If a  
problem occurs, however, diagnostic and recovery skills are essential. Thus,  
troubleshooting requires that you identify the problem quickly and apply your  
understanding of the HACMP for AIX software to restore the cluster to full  
operation.  
In general, troubleshooting an HACMP cluster involves:  
• Becoming aware that a problem exists  
• Determining the source of the problem  
• Correcting the problem  
Becoming aware of a problem is often through system messages on the  
console, end-users complaining about slow or unavailable services or  
through some sort of monitoring of your cluster. When an HACMP for AIX  
script or daemon generates a message, the message is written to the system  
console and to one or more cluster log files. Messages written to the system  
console may scroll off screen before you notice them. The following  
paragraphs provide an overview of the log files, which are to be consulted for  
cluster troubleshooting, as well as some information on specific cluster states  
you may find there.  
7.1 Cluster Log Files  
HACMP for AIX scripts, daemons, and utilities write messages to the  
following log files:  
Table 21. HACMP Log Files  
Log File Name  
Description  
/var/adm/cluster.log  
Contains time-stamped, formatted messages generated by  
HACMP for AIX scripts and daemons. In this log file, there  
is one line written for the start of each event, and one line  
written for the completion.  
/tmp/hacmp.out  
Contains time-stamped, formatted messages generated by  
the HACMP for AIX scripts. In verbose mode, this log file  
contains a line-by-line record of each command executed  
in the scripts, including the values of the arguments passed  
to the commands. By default, the HACMP for AIX software  
writes verbose information to this log file; however, you can  
change this default. Verbose mode is recommended.  
© Copyright IBM Corp. 1999  
143  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
Log File Name  
Description  
system error log  
Contains time-stamped, formatted messages from all AIX  
subsystems, including the HACMP for AIX scripts and  
daemons.  
/usr/sbin/cluster/  
history/cluster.mmdd  
Contains time-stamped, formatted messages generated by  
the HACMP for AIX scripts. The system creates a new  
cluster history log file every day that has a cluster event  
occurring. It identifies each day’s file by the filename  
extension, where mm indicates the month and dd indicates  
the day.  
/tmp/cm.log  
Contains time-stamped, formatted messages generated by  
HACMP for AIX clstrmgr activity. Information in this file is  
used by IBM Support personnel when the clstrmgr is in  
debug mode. Note that this file is overwritten every time  
cluster services are started; so, you should be careful to  
make a copy of it before restarting cluster services on a  
failed node.  
/tmp/cspoc.log  
Contains time-stamped, formatted messages generated by  
HACMP for AIX C-SPOC commands. Because the  
C-SPOC utility lets you start or stop the cluster from a  
single cluster node, the /tmp/cspoc.log is stored on the  
node that initiates a C-SPOC command.  
/tmp/dms_logs.out  
/tmp/emuhacmp.out  
Stores log messages every time HACMP for AIX triggers  
the deadman switch.  
Contains time-stamped, formatted messages generated by  
the HACMP for AIX Event Emulator. The messages are  
collected from output files on each node of the cluster, and  
cataloged together into the /tmp/emuhacmp.outlog file.In  
verbose mode (recommended), this log file contains a  
line-by-line record of every event emulated. Customized  
scripts within the event are displayed, but commands  
within those scripts are not executed.  
For a more detailed description of the cluster log files consult Chapter 2 of the  
HACMP for AIX, Version 4.3: Troubleshooting Guide, SC23-4280.  
7.2 config_too_long  
If the cluster manager recognizes a state change in the cluster, it acts upon it  
by executing an event script. However, some circumstances, like errors within  
the script or special conditions of the cluster, might cause the event script to  
144  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
hang. After a certain amount of time, by default 360 seconds, the cluster  
manager will issue a config_too_long message into the /tmp/hacmp.out file.  
The message issued looks like this:  
The cluster has been in reconfiguration too long;Something may be wrong.  
In most cases, this is because an event script has failed. You can find out  
more by analyzing the /tmp/hacmp.out file.The error messages in the  
/var/adm/cluster.log file may also be helpful. You can then fix the problem  
identified in the log file and execute the clruncmdcommand on the command  
line, or by using the SMIT Cluster Recovery Aids screen. The clruncmd  
command signals the Cluster Manager to resume cluster processing.  
Note, however, that sometimes scripts simply take too long, so the message  
showing up isn’t always an error, but sometimes a warning. If the message is  
issued, that doesn’t necessarily mean that the script failed or never finished.  
A script running for more than 360 seconds can still be working on something  
and eventually get the job done. Therefore, it is essential to look at the  
/tmp/hacmp.out file to find out what is actually happening.  
7.3 Deadman Switch  
The term “deadman switch” describes the AIX kernel extension that causes a  
system panic and dump under certain cluster conditions if it is not reset. The  
deadman switch halts a node when it enters a hung state that extends  
beyond a certain time limit. This enables another node in the cluster to  
acquire the hung node’s resources in an orderly fashion, avoiding possible  
contention problems.  
If this is happening, and it isn’t obvious why the cluster manager was kept  
from resetting this timer counter, for example because some application ran  
at a higher priority as the clstrmgrprocess, customizations related to  
performance problems should be performed in the following order:  
1. Tune the system using I/O pacing.  
2. Increase the syncdfrequency.  
3. If needed, increase the amount of memory available for the  
communications subsystem.  
4. Change the Failure Detection Rate.  
Each of these options is described in the following sections.  
Cluster Troubleshooting 145  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
7.3.1 Tuning the System Using I/O Pacing  
Use I/O pacing to tune the system so that system resources are distributed  
more equitably during large disk writes. Enabling I/O pacing is required for an  
HACMP cluster to behave correctly during large disk writes, and it is strongly  
recommended if you anticipate large blocks of disk writes on your HACMP  
cluster.  
You can enable I/O pacing using the smit chgsysfastpath to set high- and  
low-water marks. These marks are by default set to zero (disabling I/O  
pacing) when AIX is installed. While the most efficient high- and low-water  
marks vary from system to system, an initial high-water mark of 33 and a  
low-water mark of 24 provide a good starting point. These settings only  
slightly reduce write times, and consistently generate correct failover  
behavior from HACMP for AIX. If a process tries to write to a file at the  
high-water mark, it must wait until enough I/O operations have finished to  
make the low-water mark. See the AIX Performance Monitoring & Tuning  
Guide, SC23-2365 for more information on I/O pacing.  
7.3.2 Extending the syncd Frequency  
Edit the /sbin/rc.boot file to increase the syncdfrequency from its default  
value of 60 seconds to either 30, 20, or 10 seconds. Increasing the frequency  
forces more frequent I/O flushes and reduces the likelihood of triggering the  
deadman switch due to heavy I/O traffic.  
7.3.3 Increase Amount of Memory for Communications Subsystem  
If the output of netstat -m reports that requests for mbufs are being denied,  
or if errors indicating LOW_MBUFSare being logged to the AIX error report,  
increase the value associated with “thewall” network option. The default  
value is 25% of the real memory. This can be increased to as much as 50% of  
the real memory.  
To change this value, add a line similar to the following at the end of the  
/etc/rc.net file:  
no -o thewall=xxxxx  
where xxxxx is the value you want to be available for use by the  
communications subsystem. For example,  
no -o thewall=65536  
146  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
7.3.4 Changing the Failure Detection Rate  
Use the SMIT Change/Show a Cluster Network Modulescreen to change the  
failure detection rate for your network module only if enabling I/O pacing or  
extending the syncd frequency did not resolve deadman problems in your  
cluster. By changing the failure detection rate to “Slow”, you can extend the  
time required before the deadman switch is invoked on a hung node and  
before a takeover node detects a node failure and acquires a hung node’s  
resources. See the HACMP for AIX, Version 4.3: Administration Guide,  
SC23-4279 for more information and instructions on changing the Failure  
Detection Rate.  
Note  
I/O pacing must be enabled before completing these procedures; it  
regulates the number of I/O data transfers. Also, keep in mind that the Slow  
setting for the Failure Detection Rate is network specific.  
7.4 Node Isolation and Partitioned Clusters  
Node isolation occurs when all networks connecting nodes fail but the nodes  
remain up and running. One or more nodes can then be completely isolated  
from the others. A cluster in which this has happened is called a partitioned  
cluster. A partitioned cluster has two groups of nodes (one or more in each),  
neither of which cannot communicate with the other. Let’s consider a two  
node cluster where all networks have failed between the two nodes, but each  
node remains up and running.  
The problem with a partitioned cluster is that each node interprets the  
absence of keepalives from its partner to mean that the other node has failed,  
and then generates node failure events. Once this occurs, each node  
attempts to take over resources from a node that is still active and therefore  
still legitimately owns those resources. These attempted takeovers can cause  
unpredictable results in the cluster—for example, data corruption due to a  
disk being reset.  
To guard against a TCP/IP subsystem failure causing node isolation, the  
nodes should also be connected by a point-to-point serial network. This  
connection reduces the chance of node isolation by allowing the Cluster  
Managers to communicate even when all TCP/IP-based networks fail.  
It is important to understand that the serial network does not carry TCP/IP  
communication between nodes; it only allows nodes to exchange keepalives  
Cluster Troubleshooting 147  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
and control messages so that the Cluster Manager has accurate information  
about the status of its partner.  
When a cluster becomes partitioned, and the network problem is cleared after  
the point when takeover processing has begun so that keepalive packets  
start flowing between the partitioned nodes again, something must be done to  
restore order in the cluster. This order is restored by the DGSP Message.  
7.5 The DGSP Message  
A DGSP message (short for Diagnostic Group Shutdown Partition) is sent  
when a node loses communication with the cluster and then tries to  
re-establish communication.  
For example, if a cluster node becomes unable to communicate with other  
nodes, yet it continues to work through its process table, the other nodes  
conclude that the “missing” node has failed because they no longer are  
receiving keepalive messages from it. The remaining nodes then process the  
necessary events to acquire the disks, IP addresses, and other resources  
from the “missing” node. This attempt to take over resources results in the  
dual-attached disks receiving resets to release them from the “missing” node  
and the start of IP address takeover scripts.  
As the disks are being acquired by the takeover node (or after the disks have  
been acquired and applications are running), the “missing” node completes  
its process table (or clears an application problem) and attempts to resend  
keepalive messages and rejoin the cluster. Since the disks and IP addresses  
are in the process of being successfully taken over, it becomes possible to  
have a duplicate IP address on the network and the disks may start to  
experience extraneous traffic on the data bus.  
Because the reason for the “missing” node remains undetermined, you can  
assume that the problem may repeat itself later, causing additional down time  
of not only the node but also the cluster and its applications. Thus, to ensure  
the highest cluster availability, a DGSP message is sent to all nodes in one of  
the partitions. Any node receiving a DGSP message halts immediately, in  
order to not cause any damage on disks or confusion on the networks.  
In a partitioned cluster situation, the smaller partition (lesser number of  
nodes) is shut down, with each of its nodes getting a DGSP message. If the  
partitions are of equal size, the one with the node name beginning in the  
lowest name in the alphabet gets shut down. For example, in a cluster where  
one partition has NodeA and the other has NodeB, NodeB will be shut down.  
148  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
7.6 User ID Problems  
Within an HACMP cluster, you always have more than one node potentially  
offering the same service to a specific user or a specific user id.  
As the node providing the service can change, the system administrator has  
to ensure that the same user and group is known to all nodes potentially  
running an application. So, in case one node is failing, and the application is  
taken over by the standby node, a user can go on working since the takeover  
node knows that user under exactly the same user and group id.  
Since user access within an NFS mounted file system is granted based on  
user IDs, the same applies to NFS mounted file systems.  
For more information on managing user and group accounts within a cluster,  
“Managing User and Groups in a Cluster” of the HACMP for AIX, Version 4.3:  
Administration Guide, SC23-4279.  
7.7 Troubleshooting Strategy  
In order to quickly find a solution to a problem in the cluster, some sort of  
strategy is helpful for pinpointing the problem. The following guidelines  
should make the troubleshooting process more productive:  
• Save the log files associated with the problem before they become  
unavailable. Make sure you save the /tmp/hacmp.out and /tmp/cm.log files  
before you do anything else to try to figure out the cause of the problem.  
• Attempt to duplicate the problem. Do not rely too heavily on the user’s  
problem report. The user has only seen the problem from the application  
level. If necessary, obtain the user’s data files to recreate the problem.  
• Approach the problem methodically. Allow the information gathered from  
each test to guide your next test. Do not jump back and forth between  
tests based on hunches.  
• Keep an open mind. Do not assume too much about the source of the  
problem. Test each possibility and base your conclusions on the evidence  
of the tests.  
• Isolate the problem. When tracking down a problem within an HACMP  
cluster, isolate each component of the system that can fail and determine  
whether it is working. Work from top to bottom, following the progression  
described in the following section.  
Cluster Troubleshooting 149  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
• Go from the simple to the complex. Make the simple tests first. Do not try  
anything complex and complicated until you have ruled out the simple and  
obvious.  
• Do not make more than one change at a time. If you do, and one of the  
changes corrects the problem, you have no way of knowing which change  
actually fixed the problem. Make one change, test the change, and then, if  
necessary, make the next change.  
• Do not neglect the obvious. Small things can cause big problems. Check  
plugs, connectors, cables, and so on.  
• Keep a record of the tests you have completed. Record your tests and  
results, and keep an historical record of the problem, in case it reappears.  
150  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Chapter 8. Cluster Management and Administration  
This chapter covers all aspects of monitoring and managing an existing  
HACMP cluster. This includes a description of the different monitoring  
methods and tools available, how to start and stop the cluster, changing  
cluster or resource configurations, applying software fixes, user  
management, and other things.  
8.1 Monitoring the Cluster  
By design, HACMP for AIX compensates for various failures that occur within  
a cluster. For example, HACMP for AIX compensates for a network adapter  
failure by swapping in a standby adapter. As a result, it is possible that a  
component in the cluster could have failed and that you would be unaware of  
the fact. The danger here is that, while HACMP for AIX can survive one or  
possibly several failures, a failure that escapes your notice threatens a  
cluster’s ability to maintain a highly available environment.  
HACMP for AIX provides the following tools for monitoring an HACMP cluster:  
• The /usr/sbin/cluster/clstat utility, which reports the status of key  
cluster components—the cluster itself, the nodes in the cluster, and the  
network adapters connected to the nodes.  
• The HAViewutility, which monitors HACMP clusters through the NetView for  
AIX graphical network management interface. It lets users monitor  
multiple HACMP clusters and cluster components across a network from a  
single node.  
• The SMIT Show Cluster Servicesscreen, which shows the status of the  
HACMP for AIX daemons  
• The following log files: the /var/adm/cluster.log file, which tracks cluster  
events, the /tmp/hacmp.out file, which records the output generated by  
configuration scripts as they execute, the  
/usr/sbin/cluster/history/cluster.mmdd log file, which logs the daily cluster  
history, and the /tmp/cspoc.log file, which logs the status of C-SPOC  
commands executed on cluster nodes.  
When you monitor a cluster, use the clstatutility to examine the cluster and  
its components. Also, constantly monitor the /tmp/hacmp.out file. Use the  
SMIT Show Cluster Services screen to make sure that the necessary HACMP  
for AIX daemons are running on each node. Finally, if necessary, examine the  
other cluster log files to get a more in-depth view of the cluster status.  
© Copyright IBM Corp. 1999  
151  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
Consult the HACMP for AIX, Version 4.3: Troubleshooting Guide, SC23-4280,  
for help if you detect a problem with an HACMP cluster.  
8.1.1 The clstat Command  
HACMP for AIX provides the /usr/sbin/cluster/clstatcommand for  
monitoring a cluster and its components. The clstatutility is a clinfo client  
program that uses the Clinfo API to retrieve information about the cluster.  
Clinfo must be running on a node for this utility to work properly.  
The /usr/sbin/cluster/clstatutility runs on both ASCII and X Window  
Display clients in either single-cluster or multi-cluster mode. Multi-cluster  
mode requires that you use the -iflag when invoking the clstatutility. The  
client display automatically corresponds to the capability of the system. For  
example, if you run clstaton an X Window client, a graphical display for the  
utility appears. However, you can run an ASCII display on an X-capable  
machine by specifying the -a flag. In order to set up a connection to the  
cluster nodes, the /usr/sbin/cluster/etc/clhosts file must be configured on the  
client.  
The clstatutility reports whether the cluster is up, down, or unstable. It also  
reports whether a node is up, down, joining, leaving, or reconfiguring, and the  
number of nodes in the cluster. For each node, the utility displays the IP label  
and address of each network interface attached to the node, and whether that  
interface is up or down. See the clstatman page for additional information  
about this utility.  
8.1.2 Monitoring Clusters using HAView  
HAView is a cluster monitoring utility that allows you to monitor HACMP  
clusters using NetView for AIX. Using NetView, you can monitor clusters and  
cluster components across a network from a single management station.  
HAView creates and modifies NetView objects that represent clusters and  
cluster components. It also creates submaps that present information about  
the state of all nodes, networks, and network interfaces associated with a  
particular cluster. This cluster status and configuration information is  
accessible through NetView’s menu bar.  
HAView monitors cluster status using the Simple Network Management  
Protocol (SNMP). It combines periodic polling and event notification through  
traps to retrieve cluster topology and state changes from the HACMP  
management agent, that is, the Cluster SMUX peer daemon (clsmuxpd).  
152  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
More details on how to configure HAView and on how to monitor your cluster  
with HAView can be found in Chapter 3, “Monitoring an HACMP cluster” in  
HACMP for AIX, Version 4.3: Administration Guide, SC23-4279.  
8.1.3 Cluster Log Files  
HACMP for AIX writes the messages it generates to the system console and  
to several log files. Because each log file contains a different subset of the  
types of messages generated by HACMP for AIX, you can get different views  
of cluster status by viewing different log files. HACMP for AIX writes  
messages into the log files described below. See Chapter 2, “Examining  
Cluster Log Files”, in the HACMP for AIX, Version 4.3: Troubleshooting  
Guide, SC23-4280 for more information about these files.  
8.1.3.1 /var/adm/cluster.log  
The /var/adm/cluster.log file is the main HACMP for AIX log file. HACMP  
error messages and messages about HACMP for AIX-related events are  
appended to this log with the time and date when they occurred.  
8.1.3.2 /tmp/hacmp.out  
The /tmp/hacmp.out file records the output generated by the configuration  
and startup scripts as they execute. This information supplements and  
expands upon the information in the /var/adm/cluster.log file. To receive  
verbose output, the Debug Level run-time parameter should be set to high,  
which is the default.  
8.1.3.3 /usr/sbin/cluster/history/cluster.mmdd  
The /usr/sbin/cluster/history/cluster.mmdd file contains timestamped,  
formatted messages generated by HACMP for AIX scripts. The system  
creates a cluster history file whenever cluster events occur, identifying each  
file by the file name extension mmdd, where mm indicates the month and dd  
indicates the day.  
While it is more likely that you will use these files during troubleshooting, you  
should occasionally look at them to get a more detailed picture of the activity  
within a cluster.  
8.1.3.4 System Error Log  
The system error log file contains timestamped, formatted messages from all  
AIX subsystems, including HACMP for AIX scripts and daemons. Cluster  
events are logged as operator messages (error id: AA8AB241) in the system  
error log.  
Cluster Management and Administration 153  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
8.1.3.5 /tmp/cm.log  
Contains timestamped, formatted messages generated by HACMP for AIX  
clstrmgractivity. This file is typically used by IBM support personnel.  
8.1.3.6 /tmp/cspoc.log  
Contains timestamped, formatted messages generated by HACMP for AIX  
C-SPOC commands. The /tmp/cspoc.log file resides on the node that invokes  
the C-SPOC command.  
8.1.3.7 /tmp/emuhacmp.out  
The /tmp/emuhacmp.out file records the output generated by the event  
emulator scripts as they execute. The /tmp/emuhacmp.out file resides on the  
node from which the event emulator is invoked. You can use the environment  
variable EMUL_OUTPUT to specify another name and location for this file,  
but the format and information remains the same.  
With HACMP/ES, because of its RSCT technology, there are 3 more logfiles  
you may want to watch. These are:  
8.1.3.8 /var/ha/log/grpsvcs.<filename>  
Contains timestamped messages in ASCII format. These track the execution  
of internal activities of the grpsvcs daemon. IBM support personnel use this  
information for troubleshooting. The file gets trimmed regularly. Therefore,  
please save it promptly if there is a chance you may need it.  
8.1.3.9 /var/ha/log/topsvcs.<filename>  
Contains timestamped messages in ASCII format. These track the execution  
of internal activities of the topsvcs daemon. IBM support personnel use this  
information for troubleshooting. The file gets trimmed regularly. Therefore,  
please save it promptly if there is a chance you may need it.  
8.1.3.10 /var/ha/log/grpglsm  
The /var/ha/log/grpglsm file tracks the execution of internal activities of the  
grpglsm daemon. IBM support personnel use this information for  
troubleshooting. The file gets trimmed regularly. Therefore please save it  
promptly if there is a chance you may need it.  
8.2 Starting and Stopping HACMP on a Node or a Client  
This paragraph explains how to start and stop cluster services on cluster  
nodes and clients. It also describes how the Cluster-Single Point of Control  
154  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
(C-SPOC) utility can be used to start and stop cluster services on all nodes in  
cluster environments.  
Starting cluster services refers to the process of starting the HACMP for AIX  
daemons that enable the coordination required between nodes in a cluster.  
Starting cluster services on a node also triggers the execution of certain  
HACMP for AIX scripts that initiate the cluster. Stopping cluster services  
refers to stopping these same daemons on a node. This action may or may  
not cause the execution of additional HACMP for AIX scripts, depending on  
the type of shutdown you perform.  
8.2.1 HACMP Daemons  
The following lists the required and optional HACMP for AIX daemons.  
8.2.1.1 Cluster Manager daemon (clstrmgr)  
This daemon maintains the heartbeat protocol between the nodes in the  
cluster, monitors the status of the nodes and their interfaces, and invokes the  
appropriate scripts in response to node or network events. All cluster nodes  
must run the clstrmgrdaemon.  
8.2.1.2 Cluster SMUX Peer daemon (clsmuxpd)  
This daemon maintains status information about cluster objects. This daemon  
works in conjunction with the Simple Network Management Protocol (snmpd)  
daemon. All cluster nodes must run the clsmuxpddaemon.  
Note  
The clsmuxpddaemon cannot be started unless the snmpddaemon is  
running.  
8.2.1.3 Cluster Lock Manager daemon (cllockd)  
This daemon provides advisory locking services. The cllockddaemon may be  
required on cluster nodes if those nodes are part of a concurrent access  
configuration, but this is not necessarily so. Check with your application  
vendor to see if it is required.  
Note  
If the clsmuxpddaemon or the cllockddaemon cannot be started by the  
Cluster Manager (e. g. the ports are already in use), the Cluster Manager  
logs an error message and dies.  
Cluster Management and Administration 155  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
8.2.1.4 Cluster Information Program daemon (clinfo)  
This daemon provides status information about the cluster to cluster nodes  
and clients and invokes the /usr/sbin/cluster/etc/clinfo.rcscript in  
response to a cluster event. The clinfodaemon is optional on cluster nodes  
and clients. However, it is a prerequisite for running the clstat utility.  
With RSCT (RISC System Cluster Technology) on HACMP/ES Version 4.3,  
there are several more daemons.  
8.2.1.5 Cluster Topology Services daemon (topsvcsd)  
This daemon monitors the status of network adapters in the cluster. All  
HACMP/ES cluster nodes must run the topsvcsddaemon  
8.2.1.6 Cluster Event Management daemon (emsvcsd)  
This daemon matches information about the state of system resources with  
information about resource conditions of interest to client programs  
(applications, subsystems, and other programs).The emsvcsddaemon runs on  
each node of a domain.  
8.2.1.7 Cluster Group Services daemon (grpsvcsd)  
This daemon manages all of the distributed protocols required for cluster  
operation. All HACMP/ES cluster nodes must run the grpsvcsddaemon.  
8.2.1.8 Cluster Globalized Server Daemon daemon (grpglsmd)  
This daemon operates as a grpsvcsclient; its function is to make switch  
adapter membership global across all cluster nodes. All HACMP/ES cluster  
nodes must run the grpglsmddaemon.  
8.2.2 Starting Cluster Services on a Node  
You start cluster services on a node by executing the HACMP  
/usr/sbin/cluster/etc/rc.clusterscript. Use the Start Cluster Services SMIT  
screen to build and execute this command. The rc.cluster script initializes  
the environment required for HACMP by setting environment variables and  
then calls the /usr/sbin/cluster/utilities/clstart script to start the HACMP  
daemons. The clstart script is the HACMP script that starts all the cluster  
services. It does this by calling the SRC startsrccommand to start the  
specified subsystem or group.  
Using the C-SPOC utility, you can start cluster services on any node (or on all  
nodes) in a cluster by executing the C-SPOC  
/usr/sbin/cluster/utilities/cl_rc.clustercommand on a single cluster  
node. The C-SPOC cl_rc.clustercommand calls the rc.clustercommand to  
start cluster services on the nodes specified from the one node. The nodes  
156  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
           
are started in sequential order - not in parallel. The output of the command  
run on the remote node is returned to the originating node. Because the  
command is executed remotely, there can be a delay before the command  
output is returned.  
8.2.2.1 Automatically Restarting Cluster Services  
You can optionally have cluster services start whenever the system is  
rebooted. If you specify the -Rflag to the rc.cluster command, or specify  
restart or both in the Start Cluster Services SMIT screen, the rc.cluster  
script adds the following line to the /etc/inittab file.  
hacmp:2:wait:/usr/sbin/cluster/etc/rc.cluster -boot> /dev/console 2>&1  
# Bring up Cluster  
At system boot, this entry causes AIX to execute the  
/usr/sbin/cluster/etc/rc.clusterscript to start HACMP Cluster Services.  
Note  
Be aware that if the cluster services are set to restart automatically at boot  
time, you may face problems with node integration after a power failure and  
restoration, or you may want to test a node after doing maintenance work  
before having it rejoin the cluster.  
8.2.2.2 Starting Cluster Services with IP Address Takeover Enabled  
If IP address takeover is enabled, the /usr/sbin/cluster/etc/rc.clusterscript  
calls the /etc/rc.net script to configure and start the TCP/IP interfaces and to  
set the required network options.  
8.2.3 Stopping Cluster Services on a Node  
You stop cluster services on a node by executing the HACMP  
/usr/sbin/cluster/etc/clstopscript. Use the HACMP for AIX Stop Cluster  
Services SMIT screen to build and execute this command. The clstopscript  
stops an HACMP daemon or daemons. The clstopscript starts all the cluster  
services or individual cluster services by calling the SRC command stopsrc.  
Using the C-SPOC utility, you can stop cluster services on a single node or on  
all nodes in a cluster by executing the C-SPOC  
/usr/sbin/cluster/utilities/cl_clstopcommand on a single node. The  
C-SPOC cl_clstopcommand performs some cluster-wide verification and  
then calls the clstopcommand to stop cluster services on the specified  
nodes. The nodes are stopped in sequential order—not in parallel. The output  
of the command that is run on the remote node is returned to the originating  
Cluster Management and Administration 157  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
node. Because the command is executed remotely, there can be a delay  
before the command output is returned.  
8.2.3.1 When to Stop Cluster services  
You typically stop cluster services in the following situations:  
• Before making any hardware or software changes or other scheduled  
node shutdowns or reboots. Failing to do so may cause unintended cluster  
events to be triggered on other nodes.  
• Before certain reconfiguration activity. Some changes to the cluster  
information stored in the ODM require stopping and restarting the cluster  
services on all nodes for the changes to become active. For example, if  
you wish to change the name of the cluster, the name of a node, or the  
name of an adapter, you must stop and restart the cluster.  
8.2.3.2 Types of Cluster Stops  
When you stop cluster services, you must also decide how to handle the  
resources that were owned by the node you are removing from the cluster.  
You have the following options:  
Graceful  
In a graceful stop, the HACMP software shuts  
down its applications and releases its resources.  
The other nodes do not take over the resources of  
the stopped node.  
Graceful with Takeover In a graceful with takeover stop, the HACMP  
software shuts down its applications and releases  
its resources. The surviving nodes take over these  
resources. This is also called intentional failover.  
Forced  
In a forced stop, the HACMP daemons only are  
stopped, without releasing any resources. For  
example, the stopped node stays on its service  
address if IP Address Takeover has been enabled.  
It does not stop its applications, unmount its file  
systems or varyoff its shared volume groups. The  
other nodes do not take over the resources of the  
stopped node. Please note that the forced option  
is currently not supported at the Version 4.3 level  
in HACMP/ES, only in HACMP Classic.  
8.2.3.3 Abnormal Termination of a Cluster Daemon  
If the SRC detects that any HACMP daemon has exited abnormally (without  
being shut down using the clstopcommand), it executes the  
/usr/sbin/cluster/utilities/clexit.rc script to halt the system. This  
158  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
prevents unpredictable behavior from corrupting the data on the shared  
disks. See the clexit.rcman page for additional information.  
Important Note  
Never use the kill -9command on the clstrmgrdaemon. Using the kill  
command causes the clstrmgrdaemon to exit abnormally. This causes the  
SRC to run the /usr/sbin/cluster/utilities/clexit.rc script which halts the  
system immediately, causing the surviving nodes to initiate failover.  
8.2.4 Starting and Stopping Cluster Services on Clients  
Use the /usr/sbin/cluster/etc/rc.cluster script or the startsrccommand to  
start clinfoon a client, as shown below:  
/usr/sbin/cluster/etc/rc.cluster  
You can also use the standard AIX startsrccommand:  
startsrc -s clinfo  
Use the standard AIX stopsrccommand to stop clinfoon a client machine:  
stopsrc -s clinfo  
8.2.4.1 Maintaining Cluster Information Services on Clients  
In order for the clinfo daemon to get the information it needs, you must edit  
the /usr/sbin/cluster/etc/clhosts file. As installed, the clhosts file on an  
HACMP client node contains no hostnames or addresses. HACMP server  
addresses must be provided by the user at installation time. This file should  
contain all boot and service names or addresses of HACMP servers from any  
cluster accessible through logical connections with this client node. Upon  
startup, clinfouses these names or addresses to attempt communication  
with a clsmuxpdprocess executing on an HACMP server.  
An example list of hostnames/addresses in a clhosts file follows:  
n0_cl83 # n0 service  
n2_cl83 # n2 service  
n3_cl83 # n3 service  
For more detailed information on the clinfo command refer to Chapter 2,  
“Starting and Stopping Cluster Services”, HACMP for AIX, Version 4.3:  
Administration Guide, SC23-4279.  
Cluster Management and Administration 159  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
8.3 Replacing Failed Components  
From time to time, it will be necessary to perform hardware maintenance or  
upgrades on cluster components. Some replacements or upgrades can be  
performed while the cluster is operative, while others require planned  
downtime. Make sure you plan all the necessary actions carefully. This will  
spare you a lot of trouble.  
8.3.1 Nodes  
When maintaining or upgrading a node, cluster services must usually be  
stopped on the node. This means down time for the applications usually  
running on this node, at least during the takeover to other nodes.  
Consider the following points when replacing the whole or components of a  
node:  
• Make sure you have at least the same amount of RAM in the replacement  
system.  
• If your applications have been optimized for a particular processor or  
architecture, ensure that the new node is the same type of system.  
Uniprocessor applications may run slower on SMP systems.  
• Slot capacity of the new node must be the same or better.  
• Check the appropriate documentation for a proper adapter placement in  
your new node.  
• The license of your application may be dependent on the CPU ID. You  
may need to apply for a new license before trying to bring the new node  
into service.  
8.3.2 Adapters  
In order to replace or add an adapter, the node must be powered off. This  
means down time for the applications usually running on this node, at least  
during the takeover to other nodes.  
Consider the following points when replacing or adding adapters in a node:  
• Make sure that the adapter is your problem and not faulty cabling. Bad  
cables are much more common than defective adapters. Most network  
and SSA cables can be changed online. Do some testing, for example,  
exchange the cables or try to connect to another port in your hub to see if  
the hub is your problem.  
160  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
• The new adapter must be of the same type or a compatible type as the  
replaced adapter.  
• When replacing or adding an SCSI adapter, remove the resistors for  
shared buses. Furthermore, set the SCSI ID of the adapter to a value  
different than 7.  
8.3.3 Disks  
Disk failures are handled differently according to the capabilities of the disk  
type and the HACMP version you are running. Whether your data is still  
available after a disk crash, and whether you will need down time to  
exchange it, will depend on the following questions:  
• Is all the data on the failed disk mirrored to another disk, or is the failed  
disk part of a RAID array?  
• Will the volume group stay online (Quorum)?  
• Is the type of disk you are using hot-swappable?  
8.3.3.1 SSA/SCSI Disk Replacement (RAID)  
RAID arrays are typically designed for concurrent maintenance. No command  
line intervention should be necessary to replace a failed disk in a RAID array.  
Do the following steps in order to replace a disk that is a member of a RAID  
array:  
1. Remove the disk logically from the RAID array (for example with the  
appropriate SMIT menu). Removing a disk from a RAID array is known as  
reducing the RAID array. No more than one disk can be removed from an  
array at one time.  
2. Remove the failed disk and plug in the substitute disk.  
3. Add the replacement disk logically to the RAID array. All information from  
the original disk will be regenerated on the substitute disk. Once data  
regeneration has completed on the new disk, the array will return to its  
normal optimal mode of operation.  
8.3.3.2 Disk Replacement (Non-RAID) before HACMP version 4.3  
If LVM mirroring is used, some careful manual steps must be followed to  
replace a failed SCSI or SSA disk:  
1. Identify which disk has failed, using errpt, lspv, lsvg, diags.  
2. Remove all LV copies from the failed disk (rmlvcopy).  
3. Remove the disk from the VG (reducevg).  
Cluster Management and Administration 161  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
4. Logically remove the disk from the system (rmdev -l hdiskX -d; rmdev -l  
pdiskY -d if a SSA disk) on all nodes.  
5. Physically remove the failed disk and replace it with a new disk.  
6. Add the disk to the ODM (mkdevor cfgmgr)on all nodes.  
7. Add the disk to the shared volume group (extendvg).  
8. Increase the number of LV copies to span across the new disk (mklvcopy).  
9. Synchronize the volume group (syncvg)  
Note  
Steps 10 and 11 are only necessary in HACMP versions prior to 4.2. With  
HACMP 4.2 and later Lazy Update will export/import the volume group on  
the backup node in case of a takeover. However, it is necessary to update  
the PVID of the replaced disk on the backup nodes manually.  
10.Stop all the application(s) using the shared volume group, varyoff the  
shared volume group and export/import it on the backup node(s).  
Furthermore set the characteristics of the shared volume group  
(autovaryon and quorum) on the backup node(s), then vary it off again.  
11.Varyon the shared volume group on it’s “normal” node and start the  
application(s).  
8.3.3.3 Disk Replacement (Non-RAID) with HACMP version 4.3  
With the HACMP 4.3 enhancements to the C-SPOC LVM utilities, the disk  
replacement does not cause system down time, as long as the failed disk was  
part of a RAID array, or if all the LVs on it are mirrored to other disks, and the  
failed disk is hot-swappable.  
1. Identify which disk has failed using errpt, lspv, lsvg, diag.  
2. Remove all LV copies from the failed disk (smit cl_lvsc).  
3. Remove the disk from the VG (smit cl_vgsc).  
4. Logically remove the disk from the system (rmdev -l hdiskX -d, rmdev -l  
pdiskY -dif SSA disk)  
5. Physically remove the failed disk and replace it with a new disk.  
6. Add the new disk to the ODM (mkdevor cfgmgr).  
7. Add the new disk to the sharedvg (smit cl_vgsc).  
8. Increase the number of LV copies to span across the new disk (smit  
cl_lvsc).  
9. Sync the volume group (smit cl_syncvg).  
162  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
8.4 Changing Shared LVM Components  
Changes to VG constructs are probably the most frequent kind of changes to  
be performed in a cluster. As a system administrator of an HACMP for AIX  
cluster, you may be called upon to perform any of the following LVM-related  
tasks:  
• Creating a new shared volume group  
• Extending, reducing, changing, or removing an existing volume group  
• Importing, mirroring, unmirroring, or synchronizing mirrors of a volume  
group  
• Creating a new shared logical volume  
• Extending, reducing, changing, copying, or removing an existing logical  
volume (or a copy)  
• Creating a new shared file system  
• Extending, changing, or removing an existing file system  
The varyon of a shared volume group will only succeed if the information  
stored in the VGDA on the disks of the shared volume group and the  
information stored in the ODM are equal. After changes in the volume group  
(e. g. increasing the size of a file system), the information about the volume  
group in ODM and in the VGDA on the disks are still equal, but it will be  
different from the information in the ODM of a node that did not have the  
volume group varied on at the time of the change. In order to keep a takeover  
from failing, the volume group information must be synchronized. There are  
four distinct ways to keep all the volume group ODMs synchronized:  
• Manual Update  
• Lazy Update  
• C-SPOC  
TaskGuide  
Chapters 4 and 5 of the HACMP for AIX, Version 4.3: Administration Guide,  
SC23-4279, describe in detail how to change shared LVM components.  
8.4.1 Manual Update  
Sometimes, manual updates of shared LVM components are inevitable  
because you cannot do some of the tasks mentioned above with any of the  
tools. For example, neither with C-SPOC nor with TaskGuide or Lazy Update  
is it possible to remove a VG on all of the cluster nodes.  
Cluster Management and Administration 163  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
When changing shared LVM components manually, you will usually need to  
run through the following procedure:  
1. Stop HACMP on the node owning the shared volume group (sometimes a  
stop of the applications using the shared volume group may be sufficient).  
2. Make the necessary changes to the shared LVM components.  
3. Unmount all the file systems of the shared volume group.  
4. Varyoff the shared volume group.  
5. Export the old volume group definitions on the next node.  
6. Import the volume group from one of its disks on the next node. Make sure  
you use the same VG major number.  
7. Change the volume group to not auto-varyon at system boot time.  
8. Mount all the file systems of the shared volume group.  
9. Test the file systems.  
10.Unmount the file systems of the shared volume group.  
11.Varyoff the shared volume group.  
12.Repeat steps 6 through 11 for all the other nodes with an old ODM of the  
shared volume group.  
13.Start HACMP again on the node usually owning the shared volume group.  
8.4.2 Lazy Update  
For LVM components under the control of HACMP for AIX, you do not have to  
explicitly export and import to bring the other cluster nodes up-to-date.  
Instead, HACMP for AIX can perform the export and import when it activates  
the volume group during a failover. In a cluster, HACMP controls when  
volume groups are activated. HACMP for AIX implements a function, called  
Lazy Update, by keeping a copy of the timestamp from the volume group’s  
VGDA. AIX updates this timestamp whenever the LVM component is  
modified. When another cluster node attempts to vary on the volume group,  
HACMP for AIX compares its copy of the timestamp (kept in the  
/usr/sbin/cluster/etc/vg file) with the timestamp in the VGDA on the disk. If the  
values are different, the HACMP for AIX software exports and re-imports the  
volume group before activating it. If the timestamps are the same, HACMP for  
AIX activates the volume group without exporting and re-importing.  
The time needed for takeover expands by a few minutes if a Lazy Update  
occurs. A Lazy Update is always performed the first time a takeover occurs in  
order to create the timestamp file on the takeover node.  
164  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
Lazy Update has some limitations, which you need to consider when you rely  
on Lazy Update in general:  
• If the first disk in a sharedvg has been replaced, the importvgcommand  
will fail as Lazy Update expects to be able to match the hdisk number for  
the first disk to a valid PVID in the ODM.  
• Multi-LUN support on the SCSI RAID cabinets can be very confusing to  
Lazy Update as each LUN appears as a new hdisk known to only one  
node in the cluster (remember that Lazy Update works on LVM  
constructs).  
8.4.3 C-SPOC  
The Cluster Single Point of Control (C-SPOC) utility lets system  
administrators perform administrative tasks on all cluster nodes from any  
node in the cluster. These tasks are based on commonly performed AIX  
system administration commands that let you:  
• Maintain user and group accounts (see 8.8, “User Management” on page  
178).  
• Maintain shared Logical Volume Manager (LVM) components.  
• Control HACMP services on a cluster-wide basis (see Chapter 8.2,  
Without C-SPOC functionality, the system administrator must spend time  
executing administrative tasks individually on each cluster node. Using the  
C-SPOC utility, a command executed on one node is also executed on other  
cluster nodes. Thus C-SPOC minimizes administrative overhead and reduces  
the possibility of inconsistent node states. For example, to add a user, you  
usually must perform this task on each cluster node. Using C-SPOC,  
however, you issue a C-SPOC command once on a single node, and the user  
is added to all specified cluster nodes.  
C-SPOC also makes managing logical volume components and controlling  
cluster services more efficient. You can use the C-SPOC utility to start or stop  
cluster services on nodes from a single node.  
C-SPOC provides this functionality through its own set of cluster  
administration commands, accessible through SMIT menus and screens. To  
use C-SPOC, select the Cluster System Management option from the  
HACMP for AIX menu. See the HACMP for AIX, Version 4.3: Administration  
Guide, SC23-4279 for detailed information on using C-SPOC SMIT options.  
With C-SPOC, you can perform the following tasks:  
Cluster Management and Administration 165  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
• Shared volume groups  
• List all volume groups in the cluster.  
• Import a volume group (with HACMP 4.3 only).  
• Extend a volume group (with HACMP 4.3 only).  
• Reduce a volume group (with HACMP 4.3 only).  
• Mirror a volume group (with HACMP 4.3 only).  
• Unmirror a volume group (with HACMP 4.3 only).  
• Synchronize volume group mirrors (with HACMP 4.3 only).  
• Shared logical volumes  
• List all logical volumes by volume group.  
• Add a logical volume to a volume group (with HACMP 4.3 only).  
• Make a copy of a logical volume.  
• Remove a copy of a logical volume.  
• Show the characteristics of a logical volume.  
• Set the characteristics of a logical volume (name, size); this is only  
possible in non-concurrent mode and with HACMP 4.3.  
• Remove a logical volume.  
• Shared file systems (only applicable for non-concurrent VGs)  
• List all shared file systems.  
• Change/View the characteristics of a shared file system.  
• Remove a shared file system.  
C-SPOC has the following limitations:  
• C-SPOC does not offer the option for creating volume groups. Use the  
TaskGuide or standard AIX commands.  
• The new Volume Group must be imported manually to other nodes in the  
resource group (TaskGuide does it automatically).  
• The Volume Group must be defined in a resource group, and cluster  
resources must be synchronized, prior to using C-SPOC to manage it.  
• C-SPOC does not offer an option for creating file systems. Use standard  
AIX commands or SMIT menus to create file systems, and use C-SPOC to  
update the VG information on the other nodes.  
• C-SPOC cannot be used for concurrent shared LVM components prior to  
HACMP 4.3 for AIX.  
166  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
To use the SMIT shortcuts to C-SPOC, type smit cl_lvm or smit cl_conlvm for  
concurrent volume groups. Concurrent volume groups must be varied on in  
concurrent mode to perform tasks.  
8.4.4 TaskGuide  
The TaskGuide is a graphical interface that simplifies the task of creating a  
shared volume group within an HACMP cluster configuration. The TaskGuide  
presents a series of panels that guide the user through the steps of specifying  
initial and sharing nodes, disks, concurrent or non-concurrent access, volume  
group name and physical partition size, and cluster settings. The TaskGuide  
can reduce errors, as it does not allow a user to proceed with steps that  
conflict with the cluster’s configuration. Online help panels give additional  
information to aid in each step.  
8.4.4.1 TaskGuide Requirements  
TaskGuide is only available since HACMP for AIX version 4.3. Before you  
start the TaskGuide, make sure that:  
• You have a configured HACMP cluster in place  
• You are on a graphics capable terminal  
8.4.4.2 Starting the TaskGuide  
You can start the TaskGuide from the command line by typing:  
/usr/sbin/cluster/tguides/bin/cl_ccvg, or you can use the SMIT interface as  
follows:  
1. Type smit hacmp  
2. From the SMIT main menu, choose Cluster System Management ->  
Cluster Logical Volume Manager ->Taskguide for Creating a Shared  
Volume Group. After a pause, the TaskGuide Welcome panel appears.  
3. Proceed through the panels to create or share a volume group.  
8.5 Changing Cluster Resources  
In HACMP for AIX, you define each resource as part of a resource group.  
This allows you to combine related resources into a single logical entity for  
easier configuration and management. You then configure each resource  
group to have a particular kind of relationship with a set of nodes. Depending  
on this relationship, resources can be defined as one of three types:  
cascading, rotating, or concurrent access. You also assign a priority to each  
participating node in a cascading resource group chain.  
Cluster Management and Administration 167  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
To change the nodes associated with a given resource group, or to change  
the priorities assigned to the nodes in a resource group chain, you must  
redefine the resource group. You must also redefine the resource group if you  
add or change a resource assigned to the group. This section describes how  
to add, change, and delete a resource group.  
8.5.1 Add/Change/Remove Cluster Resources  
You can add, change and remove a resource group in an active cluster. You  
do not need to stop and then restart cluster services for the resource group to  
become part of the current cluster configuration.  
Use the following SMIT shortcuts:  
To add a resource group, use smit cm_add_grp.  
To remove a resource group, use smit cm_add_res.  
To change a resource group, use smit cm_add_res.  
Whenever you modify the configuration of cluster resources in the ODM  
on one node, you must synchronize the change across all cluster nodes.  
8.5.2 Synchronize Cluster Resources  
You perform a synchronization by choosing the Synchronize Cluster  
Resources option from the Cluster Resources SMIT screen.  
Note  
In HACMP for AIX, the event customization information stored in the ODM  
is synchronized across all cluster nodes when the cluster resources are  
synchronized. Thus, pre, post, notify, and recovery event script names  
must be the same on all nodes, although the actual processing done by  
these scripts can be different.  
The processing performed in synchronization varies depending on whether  
the Cluster Manager is active on the local node:  
Ιf the cluster manager is not active on the local node when you select this  
option, the ODM data in the DCD (Default Configuration Directory–for  
more information, see Chapter 3 in the HACMP for AIX, Version 4.3:  
Concepts and Facilities, SC23-4276) on the local node is copied to the  
ODMs stored in the DCDs on all cluster nodes.  
168  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
• If the Cluster Manager is active on the local node, synchronization triggers  
a cluster-wide, dynamic reconfiguration event. In dynamic reconfiguration,  
the configuration data stored in the DCD is updated on each cluster node,  
and, in addition, the new ODM data replaces the ODM data stored in the  
ACD (Active Configuration Directory) on each cluster node. The cluster  
daemons are refreshed and the new configuration becomes the active  
configuration. In the HACMP for AIX log file, reconfig_resource_release,  
reconfig_resource_acquire, and reconfig_resource_complete events mark  
the progress of the dynamic reconfiguration.  
• If the Cluster Manager is active on some cluster nodes but not on the local  
node, the synchronization is aborted.  
8.5.3 DARE Resource Migration Utility  
The HACMP for AIX software provides a Dynamic Reconfiguration (DARE)  
Resource Migration utility that allows for improved cluster management by  
allowing a system administrator to alter the placement of resource groups  
(along with their resources—IP addresses, applications, and disks) to specific  
cluster nodes using the cldarecommand. The command lets you move the  
ownership of a series of resource groups to a specific node in that resource  
group’s node list, as long as the requested arrangement is not incompatible  
with the current resource group configuration. It also lets you disable  
resource groups, preventing them from being acquired during a failover or  
reintegration.  
Dynamic resource group movement essentially lets a system administrator  
better use hardware resources within the cluster, forcing resource traffic onto  
one or more high-powered or better-connected nodes without having to shut  
down HACMP on the node from which the resource group is moved. Dynamic  
resource group movement also lets you perform selective maintenance  
without rebooting the cluster or disturbing operational nodes.  
Using the DARE Resource Migration utility does not affect other resource  
groups that might currently be owned by that node. The node that currently  
owns the resource group will release it as it would during a “graceful  
shutdown with takeover”, and the node to which the resource group is being  
moved will acquire the resource group as it would during a node failover.  
The following section covers the types and location keywords used in DARE  
resource migrations, and also how to use the cldare command and the -M  
flag to perform the migration.  
Cluster Management and Administration 169  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
8.5.3.1 Resource Migration Types  
Before performing a resource migration, decide if you will declare the  
migration stickyor non-sticky.  
Sticky Resource Migration  
A sticky migration permanently attaches a resource group to a specified  
node. The resource group attempts to remain on the specified node during a  
node failover or reintegration.  
Since stickiness is a behavioral property of a resource group, assigning a  
node as a sticky location makes the specified resource group a sticky  
resource. Older sticky locations are superseded only by new sticky migration  
requests for the same resource group, or they are removed entirely during  
non-sticky migration requests for the same resource group. If it is not  
possible to place a resource group on its sticky location (because that node is  
down), the normal resource policy is invoked, allowing the resource to  
migrate according to the takeover priority specified in the resource group’s  
node list.  
For both cascading and rotating resource groups, a normal resource policy  
means that other cluster nodes in the group’s node list are consulted at the  
time the sticky location fails to find the highest-priority node active. After  
finding the active node, cascading resource groups will continually migrate to  
the highest-priority node in the group’s node list (ultimately residing at the  
sticky location). Rotating resource groups stay put until the sticky location  
returns to the cluster.  
You can attach the optional keyword stickyto any migration you perform,  
regardless of the resource group configuration (rotating or cascading).  
However, with very few exceptions, you always use the sticky location for  
cascading configurations, and do not use it for rotating configurations.  
Non-Sticky Resource Migration  
Resource groups on nodes not designated sticky are by default transient,  
non-sticky resources. These resources are temporarily placed on the  
specified node with the highest priority in the node list until the next failover or  
reintegration occurs. Non-sticky resources are best suited for use with  
rotating resource group configurations because of this transient behavior.  
Because the normal behavior of cascading resources is to bound back to the  
highest available node in their node list, non-sticky migrations are usually not  
the best choice. The one instance in which a non-sticky migration of a  
cascading resource might make sense is if this resource has the  
170  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
INACTIVE_TAKEOVER flag set to false and has not yet started because its  
primary node is down.  
In general, however, only rotating resource groups should be migrated in a  
non-sticky manner. Such migrations are one-time events and occur similar to  
normal rotating resource group flavors. After migration, the resource group  
immediately resumes a normal rotating resource group failover policy, but  
from the new location.  
Note  
The cldarecommand attempts to perform all requested migrations  
simultaneously. If, for some reason, the command cannot simultaneously  
cause all specified resources to be released and cannot simultaneously  
reacquire them at the new locations, it fails, and no migrations occur.  
8.5.3.2 Locations  
You can specify the location for a resource group by entering a node name or  
a keyword.  
Node Name  
In most cases, you enter a node name in the location field to specify which  
node will contain sticky or non-sticky resource groups. Node names can be  
arbitrary and apply to both rotating and cascading resource group  
configurations.  
The DARE Resource Migration utility also provides the following special  
keywords you can use in the location field to determine the placement of  
migrated resource groups: defaultand stop. The defaultand stoplocations  
are special locations that determine resource group behavior and whether the  
resources can be reacquired.  
Default Location  
If you use the default keyword as the location specifier, the DARE Resource  
Migration utility removes all previous stickiness for the resource group and  
returns the resource group to its default failover behavior where node  
priorities apply (for either cascading or rotating resources). The use of a  
default destination for a cascading resource group returns it to its normal  
behavior (the resource group will migrate to the highest priority node currently  
up). Using a default destination for a rotating resource group releases the  
group from wherever it resides and lets the highest priority node with a boot  
address reacquire the resource.  
Cluster Management and Administration 171  
Download from Www.Somanuals.com. All Manuals Search And Download.  
If you do not include a location specifier in the location field, the DARE  
Resource Migration utility performs a default migration, again making the  
resources available for reacquisition.  
Note  
A default migration can be used to start a cascading resource group that  
has INACTIVE_TAKEOVER set to false and that has not yet started  
because its primary node is down.  
Stop Location  
The second special location keyword, stop, causes a resource group to be  
made inactive, preventing it from being reacquired, though it remains in the  
resource configuration. Its resources remain unavailable for reacquisition  
even after a failover or reintegration.  
8.5.3.3 Using the cldare Command to Migrate Resources  
The cldarecommand can be used to perform dynamic resource group  
migrations to other cluster nodes in conjunction with other cldare resource  
functionality. It lets you specify multiple resource groups and nodes on the  
command line, as long as the final resource group configuration is consistent.  
After some error checking, the resources are released and reacquired by the  
specified cluster nodes. Resource migration first releases all specified  
resources (wherever they reside in the cluster); then it reacquires these  
resources on the newly specified nodes.  
You can also use this command to swap resources on nodes in the resource  
group’s node list, but you cannot mix keywords—default, stop,and  
node—when using the cldarecommand.  
To migrate resource groups (and their resources) using the cldarecommand,  
enter the following command:  
cldare -M <resgroup name>:[location|[default|stop]][:sticky] ...  
where -Mspecifies migration, and where resource group names must be valid  
names of resource groups in the cluster. You can specify a node name (or  
special location) or the keyword stopor defaultafter the first colon. The node  
name must represent a cluster node that is up and in the resource group’s  
node list. You can specify a migration type after the second colon. Repeat this  
syntax on the command line for each resource group you want to migrate. Do  
not include spaces between arguments.  
172  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
Note that you cannot add nodes to the resource group list with the DARE  
Resource Migration utility. This task is performed through SMIT.  
Stopping Resource Groups  
If the location field of a migration contains the keyword stopinstead of an  
actual nodename, the DARE Resource Migration utility attempts to stop the  
resource group, which includes taking down any service label, unmounting  
file systems, and so on. You should typically supplement the keyword stop  
with the migration type stickyto indicate that the resource stays down, even  
if you reboot the cluster.  
As with sticky locations, sticky stop requests are superseded by new sticky  
migration requests for the same resource group, or they are removed by  
default,non-sticky migration requests for the same resource group. Thus, a  
stopped resource will be restarted at the time of the next migration request.  
Note  
Be careful when using a non-sticky stop request, since the resource group  
will likely be restarted at the next major cluster event. As a result, all  
non-sticky requests produce warning messages. A non-sticky stop could  
be used to halt a cascading resource group that has  
INACTIVE_TAKEOVER set to false during periods in which its primary  
node is down.  
8.5.3.4 Using the clfindres Command  
To help you locate resources placed on a specific node, the DARE Resource  
Migration utility includes a command, clfindres, that makes a best-guess  
estimate (within the domain of current HACMP configuration policies) of the  
state and location of specified resource groups. It also indicates whether a  
resource group has a sticky location, and it identifies that location.  
See Appendix A of the HACMP for AIX, Version 4.3: Administration Guide,  
SC23-4279, for the syntax and typical output of the clfindrescommand.  
8.5.3.5 Removing Sticky Markers When the Cluster is Down  
Sticky location markers are stored in the HACMPresource class in the  
HACMP ODM and are a persistent cluster attribute. While the cluster is up,  
you can only remove these locations by performing a subsequent non-sticky  
migration on the same resource group, using the defaultspecial location  
keyword or specifying no location.  
Cluster Management and Administration 173  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
Be aware that persistent sticky location markers are saved and restored in  
cluster snapshots. You can use the clfindrescommand to find out if sticky  
markers are present in a resource group.  
If you want to remove sticky location markers while the cluster is down, the  
defaultkeyword is not a valid method, since it implies activating the resource.  
Instead, when the cluster is down, you use a transient stoprequest, as in this  
example:  
cldare -v -M <resgroup name>:stop  
(The optional -v flag indicates that verification is skipped.)  
8.6 Applying Software Maintenance to an HACMP Cluster  
You can install software maintenance, called Program Temporary Fixes  
(PTFs), to your HACMP cluster while running HACMP for AIX cluster services  
on cluster nodes; however, you must stop cluster services on the node on  
which you are applying a PTF. As with everything else in a cluster, applying  
software fixes should be done in a controlled fashion.  
With the method described below, you might even be able to keep your  
mission-critical application up and running during the update process,  
provided that the takeover node is designed to carry its own load and the  
takeover load as well.  
The normal method of applying AIX fixes is to do the following:  
1. Use the smit clstopfastpath to stop cluster services on the node on which  
the PTF is to be applied. If you would like the resources provided by this  
node to remain available to users, stop cluster with takeover so that the  
takeover node will continue to provide these resources to users.  
2. Apply the software maintenance to this node using the procedure  
described in the documentation distributed with the PTF.  
3. Run the /usr/sbin/cluster/diag/clverifyutility to ensure that no errors  
exist after installing the PTF. Test the fix as thoroughly as possible.  
4. Reboot the node to reload any HACMP for AIX kernel extensions that may  
have changed as a result of the PTF being applied.  
If an update to the cluster.base.client.lib file set has been applied and you  
are using Cluster Lock Manager or Clinfo API functions, you may need to  
relink your applications.  
174  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
5. Restart the HACMP for AIX software on the node using the smit clstart  
fastpath and verify that the node successfully joined the cluster.  
6. Repeat Steps 1 through 5 on the remaining cluster nodes.  
Figure 15 below shows the procedure:  
Fallover of  
System A  
System A  
rejoins the  
cluster  
System A  
System B  
System A  
System B  
System A  
System B  
Apply PTFs  
to System A  
Test  
System A  
System A  
System B  
System A's  
Resource  
System B's  
Resource  
Figure 15. Applying a PTF to a Cluster Node  
Along with the normal rules for applying updates, the following general points  
should be observed for HACMP clusters:  
• Cluster nodes should be kept at the same AIX maintenance levels  
wherever possible. This will, of course, not be true while the update is  
being applied, but should be true at all other times.  
Cluster Management and Administration 175  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
• Cluster nodes should be running the same HACMP maintenance levels.  
There might be incompatibilities between various maintenance levels of  
HACMP, so you must ensure that consistent levels are maintained across  
all cluster nodes. The cluster must be taken down to update the  
maintenance levels.  
8.7 Backup Strategies  
HACMP software masks hardware failures in clustered RISC System/6000  
environments by quickly switching over to backup machines or other  
redundant components. However, installing HACMP is not a substitute for  
having a formal backup and recovery procedure.  
In general, a backup of user and system data is kept in case data is  
accidentally removed or in case of a disk failure. A formal backup process is  
really an insurance policy. You invest in the technology and time to back up  
systems so that, in the event of a problem, you can quickly rebuild the  
system.  
Since system and application backups are preferably done during periods of  
no usage (for instance, in the middle of the night), many installations  
implement an automated backup procedure using the AIX cron facility. While  
this is a very good procedure, the HACMP cluster environment presents  
some special challenges. The problem is, you never know which machine has  
your application data online, so you need to ensure that exactly the node that  
has a resource online will initiate the backup of data.  
It isn’t actually important which of the several backup commands you are  
using, what is important is the strategy. For the features and/or restrictions of  
backup commands like tar, cpio, dd or backup, refer to the AIX Commands  
Reference Version 4.3, SBOF-1877.  
8.7.1 Split-Mirror Backups  
No file system can be safely backed up while update activity is occurring. If  
you are going to have any assurance as to which updates are on the backup  
and which updates are not, you need to be able to demark exactly where the  
backup was made. Therefore, it may be difficult to do a good backup on  
systems that have applications or data that must be online continuously or  
offline for only a very short time. In some installations, the time required to do  
a full backup to an archival device, or even to another, might be longer than  
the availability requirements of the application will allow it to be offline. The  
mirroring capability of the AIX Logical Volume Manager (LVM) can be used to  
address this issue.  
176  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
8.7.1.1 How to do a split-mirror backup  
This same procedure can be used with just one mirrored copy of a logical  
volume. If you remove a mirrored copy of a logical volume (and file system),  
and then create a new logical volume (and file system) using the allocation  
map from that mirrored copy, your new logical volume and file system will  
contain the same data as was in the original logical volume.  
Now, you can mount this new file system (read-only is recommended), back it  
up, and you are really backing up a mirrored copy of the data in the original  
file system, as it was when we removed the mirror copy. Since this file  
system, created from the mirror copy, is mounted read-only, no inconsistency  
in the file system from the point at which you removed the mirror originally is  
created during the backup. After that, you can delete the new file system to  
release the physical partitions back to the free pool. Finally, you can add and  
synchronize a mirror back onto the original file system, and you are back to a  
mirrored mode of operation, with fully updated data.  
The splitlvcopycommand of AIX does much of the work required to  
implement this solution.  
We can summarize the steps to do a split-mirror backup of a file system as  
follows:  
1. Use the lsvg -l VGNAME command to take note of the logical volumet name  
that contains the file system you want to back up.  
2. Stop any application using the file system and unmount the file system.  
3. Use the splitlvcopycommand to break off one mirror of the logical  
volume, and create a new logical volume with its contents. For example, if  
the existing logical volume is named fslv, the command would be  
splitlvcopy -y newlv fslv.  
4. It is important to note that there is now one less mirror copy available to  
the user in fslv.  
5. Remount the file system and restart the application that was using it.  
6. You can see that the application and data are offline for only a very short  
time.  
7. Create a file system on your new logical volume and mount it read-only.  
This is to ensure that no update activity will occur in the new logical  
volume, and the consistency of the file system is guaranteed.  
8. Perform the backup on the new file system by any means desired, such as  
backup, tar, cpio,and pax.  
Cluster Management and Administration 177  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
9. After the backup is complete and verified, unmount and delete the new file  
system and the logical volume you used for it.  
10.Use the mklvcopycommand to add back the logical volume copy you  
previously split off to the fslv logical volume.  
11.Resynchronize the logical volume.  
Once the mirror copy has been recreated on the logical volume, the syncvg  
command will resynchronize all physical partitions in the new copy, including  
any updates that have occurred on the original copy during the backup  
process.  
It is always a good idea to check a backup for validity.  
8.7.2 Using Events to Schedule a Backup  
As described above, a crontabentry is often used for scheduling nightly  
backups during off-peak hours of the application. Now as you have several  
cluster nodes, each of them would need a crontabentry, in order to get its  
own data backed up. This crontabentry can determine whether only the  
“normal” data is backed up, i.e. the data this cluster node cares about during  
“normal” operations, or, in case of another’s node failure and a subsequent  
takeover of this node’s resources, backing up both of the cluster nodes’ data.  
Whenever one node takes over the reources of another node, the  
node_down_remote event has happened. You can use a post-event to the  
node_down_remote event to change the crontabentry from backing up only  
the local node’s data into backing up both nodes’ data.  
Furthermore, if the second node eventually comes up again and takes its  
resources back, you will see a node_up_remote event in your logs. Thus, you  
can configure a post-event to the node_up_remote event to change the  
crontab entry back to the “normal” setting.  
If you want to do a split-mirror backup, the crontab entry has to invoke a  
script, implementing the steps described above.  
A more detailed description of this procedure can be found in the redbook  
HACMP/6000 Customization Examples, SG24-4498, Chapter 6.  
8.8 User Management  
As 2.7, “User ID Planning” on page 48 described, on an HACMP cluster, the  
administrator has to take care of user and group IDs throughout the cluster. If  
178  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
they don’t match, the user won’t get anything done after a failover happened.  
So, the administrator has to keep definitions equal throughout the cluster.  
Fortunately, the C-SPOC utility, as of HACMP Version 4.3 and later, does this  
for you. When you create a cluster group or user using C-SPOC, it makes  
sure that it has the same group id or user id throughout the cluster.  
8.8.1 Listing Users On All Cluster Nodes  
To obtain information about all user accounts on cluster nodes (or about a  
particular user account), you can either use the AIX lsusercommand in rshto  
one cluster node after another, or use the C-SPOC cl_lsusercommand, or  
the C-SPOC SMIT List all the Users on the Cluster screen. The cl_lsuser  
command executes the AIX lsusercommand on each node. To obtain a  
listing of all user accounts in the cluster, you must specify the ALLargument.  
If you specify a user name that does not exist on one of the cluster nodes, the  
cl_lsusercommand outputs a warning message but continues execution of  
the command on other cluster nodes.  
Note  
If you have a Network Information Service (NIS) database installed on any  
cluster node, some user information may not appear when you use the  
cl_lsusercommand.  
8.8.2 Adding User Accounts on all Cluster Nodes  
Adding a user to the cluster involves three steps:  
1. Add an entry for the new user to the /etc/passwd file and other system  
security files.  
2. Create a home directory for the new user.  
3. Add the user to a group file.  
On AIX systems, you use the mkusercommand to perform these tasks. This  
command adds entries for the new user to various system security files,  
including /etc/passwd and /etc/security/passwd, adds the new user to a  
group, and creates a home directory for the new user. Every user account  
has a number of attributes associated with it. When you create a user, the  
mkusercommand fills in values for these attributes from the system default  
/usr/lib/security/mkuser.default file. You can override these default values by  
specifying an attribute and a value on the mkusercommand line.  
Cluster Management and Administration 179  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
To add a user on one or more nodes in a cluster, you can either use the AIX  
mkusercommand in a rshto one clusternode after the other, or use the  
C-SPOC cl_mkusercommand or the Add a User to the Cluster SMIT screen.  
The cl_mkusercommand calls the AIX mkusercommand to create the user  
account on each cluster node you specify. The cl_mkusercommand creates a  
home directory for the new account on each cluster node.  
8.8.3 Changing Attributes of Users in a Cluster  
On AIX systems, you can change any of the attributes associated with an  
existing user account by using the chusercommand. Using the chuser  
command, you specify the name of the user account you want to change and  
then specify the attributes with their new values. If you use the SMIT Change  
User Attributes screen, the complete list of user attributes is displayed and  
you can supply new values for any attributes. The chusercommand modifies  
the user information stored in the /etc/passwd file and the files in the  
/etc/security directory.  
To change the attributes of a user account on one or more cluster nodes, you  
can either use the AIX chusercommand in rshto one cluster node after the  
other, or use the C-SPOC cl_chusercommand or the C-SPOC Change User  
Attributes SMIT screen. The cl_chusercommand executes the AIX chuser  
command on each cluster node.  
Note  
Do not use the cl_chusercommand if you have an NIS (Network  
Information Service) database installed on any node in your cluster.  
Both cluster nodes must be active and a user with the specified name must  
exist on both the nodes for the change operation to proceed. Optionally, you  
can specify that the cl_chusercommand continue processing if the specified  
user name exists on any of the cluster nodes. See the cl_chusercommand  
man page for more information.  
8.8.4 Removing Users from a Cluster  
On AIX systems, you remove a user account by using the rmusercommand or  
the SMIT Remove a User From the System screen. Using the rmuser  
command you specify the name of the user account you want to remove and  
specify whether you want the user password and other authentication  
information removed from the /etc/security/passwd file.  
180  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
To remove a user account from one or more cluster nodes, you can either use  
the AIX rmusercommand on one cluster node after the other, or use the  
C-SPOC cl_rmusercommand or the C-SPOC Remove a User from the  
Cluster SMIT screen. The cl_rmusercommand executes the AIX rmuser  
command on all cluster nodes.  
Note  
The system removes the user account but does not remove the home  
directory or any files owned by the user. These files are only accessible to  
users with root authority or by the group in which the user was a member.  
8.8.5 Managing Group Accounts  
In order to manage a number of similar users as a single entity, AIX provides  
the administrator with the group concept. Members of one group share the  
same permissions, the same attributes and limits, and so on.  
Commands for managing group accounts are just like the user managing  
commands very much alike to the native AIX commands. The restrictions on  
NIS are just the same as for users, and therefore are not explained here in  
detail.  
For more detailed information, please refer to Chapter 12 of the HACMP for  
AIX, Version 4.3: Administration Guide, SC23-4279.  
8.8.6 C-SPOC Log  
Because these commands are running and executing while distributed  
amongst the cluster, it could happen that something doesn’t work exactly like  
it should. The C-SPOC utility, therefore, maintains a log on the initiating node.  
It can be found under /tmp/cspoc.log.  
Note that the initiating node doesn’t have to be the same in all cases, so the  
log file might be present on different cluster nodes, and doesn’t contain the  
same data.  
Cluster Management and Administration 181  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
182  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Chapter 9. Special RS/6000 SP Topics  
This chapter will introduce you to some special topics that only apply if you  
are running HACMP on the SP system.  
9.1 High Availability Control Workstation (HACWS)  
If you are thinking about what could happen to your SP whenever the Control  
Workstation might fail, you will probably think about installing HACWS for  
that. These paragraphs will not explain HACWS in full detail, but will  
concentrate on the most important issues for installation and configuration.  
For more details, refer to Chapter 3, “Installing and Configuring the High  
Availability Workstation”, in the IBM Parallel System Support Programs for  
AIX Installation and Migration Guide, GA22-7347, or to Chapter 4,”Planning  
for a High Availability Workstation”, in the IBM RS/6000 SP Planning Volume  
2, Control Workstation and Software Environment, GA22-7281.  
Some services of the control workstation (or cws for short) are vital, so a  
failure would impact your ability to manage an SP system. Also, the failure of  
the control workstation could cause the switch network to fail. HACWS covers  
the following cases with a fully functional environment:  
• Continues running your SP system after a cws failure  
• Shuts down the cws for deferred hardware and software maintenance  
without having a system outage  
• Maintains the SP system function and reliability when the cws fails  
• Fails over the cws to a backup  
9.1.1 Hardware Requirements  
To build a cluster consisting of two control workstations, you have to think  
about shared resources. The spdata file system holding the SDR data and  
other vital data has to be accessible from both control workstations, so, it has  
to be put onto a shared disk.  
The cws connects to the frames of an RS/6000 SP with RS232 lines as its  
supervisor network. If the RS/6000 SP consists of multiple frames, you will  
probably have an 8-port adapter installed in the Control Workstation in order  
to provide the needed number of ttys.  
To connect a backup cws to the frames, you need exactly the same tty port  
configuration as on the primary cws, that is, when frame 3 connects to tty3 on  
the primary cws, it has to connect to tty3 on the backup cws as well. Also, you  
© Copyright IBM Corp. 1999  
183  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
need to have the frame supervisors support dual tty lines in order to get both  
control workstations connected at the same time. Contact your IBM  
representative for the neccessary hardware (see Figure 16 on page 184).  
Both the tty network and the RS/6000 SP internal ethernet are extended to  
the backup cws. In contrast to standard HACMP, you don’t need to have a  
second ethernet adapter on the backup cws. In case you have only one, the  
HACWS software will work with ip aliasing addresses on one adapter.  
Primary CWS  
SP Ethernet  
e
l
ab  
C
ASD  
D
Ac  
e
v
t
i
v
Mirror  
e
Acti  
t
t
HACMP  
Serial  
Link  
y
C
a
HACMP  
Heartbeat  
b
l
Frame  
e
Supervisor  
Y-cable  
Inacti  
v
e
D
le  
Cab  
ASD  
tty  
e
tiv  
Cab  
le  
c
a
n
I
SP Ethernet  
Backup CWS  
Figure 16. A Simple HACWS Environment  
9.1.2 Software Requirements  
Both of the control workstations must have the same software installed, that  
is, they must be on the same AIX level, use the same PSSP software level  
and have to have HACMP on the same level as well. For example, if you want  
to use HACMP 4.3 for AIX, you have to use PSSP 3.1 and therefore AIX  
Version 4.3.2 on the primary cws and on the backup cws.  
9.1.3 Configuring the Backup CWS  
The primary cws is configured exactly as usual, as far as the AIX and PSSP  
software is concerned, as if there were no HACWS at all.  
184  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
The backup cws has to be installed with the same level of AIX and PSSP.  
Depending on the kerberos configuration of the primary cws, the backup cws  
has to be configured either as a secondary authentication server for the  
authentication realm of your RS/6000 SP when the primary cws is an  
authentication server itself, or as an authentication client when the primary  
cws is an authentication client of some other server. To do so will enable a  
correct kerberos environment on the backup cws; so, remote commands will  
succeed through kerberos authentication as on the primary cws.  
After the initial AIX and PSSP setup is done, the HACWS software has to be  
installed.  
9.1.4 Install High Availability Software  
On both control workstations, the HACMP software has to be installed now,  
according to the instructions in the HACMP for AIX, Version 4.3: Installation  
Guide, SC23-4278. Verification, as described in Chapter 10 of the HACMP for  
AIX, Version 4.3: Installation Guide, SC23-4278, should be performed. For  
HACWS control workstations, the ssp.hacws fileset has to be installed as  
well.  
9.1.5 HACWS Configuration  
Since the cws might have some daemons active that could interfere with the  
definition and configuration of networks, you have to stop them, in order to  
get the configuration done, with the command:  
/usr/sbin/hacws/spcw_apps -d  
This will stop the subsystems spmgr splogd hardmon sysctld supfilesrv  
sp_configdif they have been active with the corresponding SRC command.  
Now configure the serial network. You can either use target mode SCSI,  
target mode SSA or the raw RS-232 serial line, or any combination.  
Both machines, primary and backup, need to be configured to boot up on  
their boot address in order to not confuse a working cws at boot time of the  
backup cws.  
If not previously done, you have to migrate the /spdata file system to an  
external volume group, to make it accessible from both sides.  
After the /spdata file system is set up so that a varyonvgof its vg will work on  
either cws, you have to complete the Administration Tasks, like on an  
Special RS/6000 SP Topics 185  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
ordinary HACMP cluster, as it is described in Chapter 7 of the HACMP for  
AIX, Version 4.3: Installation Guide, SC23-4278.  
Now the cluster environment has to be configured. Define a cluster ID and  
name for your HACWS cluster and define the two nodes to HACMP.  
Adapters have to be added to your cluster definition as described before. You  
will have to add a boot adapter and a service adapter for both primary and  
backup cws. Now that the cluster topology is defined to HACMP, you can  
configure Network Modules as in the HACMP for AIX Installation Guide,  
SC23-4278, and synchronize the definitions across the two cluster nodes.  
With the ssp.hacws filesets comes a predefined start- and stop-script, which  
has to be defined to HACMP as part of the application server definition, which  
in turn has to be included in a resource group.  
Recommended settings for this resource group are:  
Resource Group Name  
Node Relationship  
[hacws_group1]  
[rotating]  
Participating Node Names [“nodename of primary cws” “nodename of  
backup cws”]  
Service IP label  
File System  
at least the hostname of the primary cws  
the name of the file system, most probably  
/spdata  
Volume Groups  
the name of the shared volume group containing  
/spdata  
Application Servers  
the name you gave the application server before  
9.1.6 Setup and Test HACWS  
Both the primary and backup cws have to be addressable by their hostname,  
in order to finish the configuration and check that everything is in order. So,  
check if the primary cws can address the backup cws by its hostname and  
vice versa. If not, use the ifconfigcommand to temporarily set the interface  
to the hostname on each cws. Do NOT use smit chinetfor this, since this  
would be a permanent change.  
Run the command:  
/usr/sbin/hacws/install_hacws -p primary_hostname -b backup_hostname -s  
on the primary cws to set up HACWS with the 2 node names.  
186  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
After that, identify the HACWS event scripts to HACMP by executing the  
/usr/sbin/hacws/spcw_addevents  
command, and verify the configuration with the  
/usr/sbin/hacws/hacws_verify  
command. You should also check the cabling from the backup cws with the  
/usr/sbin/hacws/spcw_verify_cabling  
command. Then reboot the primary and the backup cws, one after the other,  
and start cluster services on the primary cws with smit clstart. After cluster  
services is up and running, check that control workstation services, such as  
SDRGetObjects, are working as expected. If everything is fine, start up cluster  
services on the backup cws as well. Check for the completion of the cluster  
services startup with the following command:  
grep "SPCW_APPS COMPLETE" /tmp/hacmp.out  
Now you can cause a failover by stopping cluster services on the primary cws  
and see whether cws services are still available afterwards.  
9.2 Kerberos Security  
To understand security, we have to clarify some definitions first.  
Identification is the process by which an entity tells another who it is.  
Authentication is the process by which the other entity verifies this identity.  
Authorization is the process performed by an entity to check if an agent,  
whose identity has previously been authenticated, has or  
does not have the necessary privileges to carry out some  
action.  
Additionally, if information is transferred over an insecure network, as any  
TCP/IP network basically is, there is always a chance that someone is  
listening, so some sort of encryption is required.  
These issues are solved with kerberos.  
Special RS/6000 SP Topics 187  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
Kerberos  
Also spelled Cerberus - The watchdog of Hades, whose duty was to guard  
the entrance (against whom or what does not clearly appear); it is known to  
have had three heads.  
- Ambrose Bierce, The Enlarged Devil’s Dictionary  
The following is simply a shortened description on how kerberos works. For  
more details, the redbook Inside the RS/6000 SP, SG24-5145, covers the  
subject in much more detail.  
When dealing with authentication and Kerberos, three entities are involved:  
the client, who is requesting service from a server; the second entity, and the  
Key Distribution Center or Kerberos server, which is a machine that manages  
the database, where all the authentication data is kept and maintained.  
Kerberos is a third-party system used to authenticate users or services that  
are known to Kerberos as principals. The very first action to take regarding  
Kerberos and principals is to register the latter to the former. When this is  
done, Kerberos asks for a principal’s password, which is converted to a  
principal (user or service) 56-bit key using the DES (Data Encryption  
Standard) algorithm. This key is stored in the Kerberos server database.  
When a client needs the services of a server, the client must prove its identity  
to the server so that the server knows to whom it is talking.  
Tickets are the means the Kerberos server gives to clients to authenticate  
themselves to the service providers and get work done on their behalf on the  
services servers. Tickets have a finite life, known as the ticket life span.  
In Kerberos terms, to make a Kerberos authenticated service provider work  
on behalf of a client is a three-step process:  
• Get a ticket-granting ticket.  
• Get a service ticket.  
• Get the work done on the service provider.  
The main role of the ticket-granting ticket service is to avoid unnecessary  
password traffic over the network; so, the user should issue his password  
only once per session. What this ticket-granting ticket service does is to give  
the client systems a ticket that has a certain time span, whose purpose is to  
188  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
allow the clients to get service tickets to be used with other servers without  
the need to give them the password every time they request services.  
So, given a user has a ticket-granting ticket, if a user requests a kerberized  
service, he has to get a service ticket for it. In order to get one, the kerberized  
command sends an encrypted message, containing the requested service  
name, the machine’s name, and a time-stamp to the Kerberos server. The  
Kerberos server decrypts the message, checks whether everything is in  
order, and if so, sends back a service ticket encrypted with the service’s  
private key, so that only the requested service can decrypt it. The client sends  
his request along with the just received ticket to the service provider, who in  
turn decrypts and checks authorization, and then, if it is in order, provides the  
requested service to the client.  
9.2.1 Configuring Kerberos Security with HACMP Version 4.3  
With HACMP Version 4.3 there is a handy script to do the kerberos setup for  
you, called cl_setup_kerberos. It sets up all the IP labels defined to the  
HACMP cluster together with the needed kerberos principals, so that remote  
kerberized commands will work.  
On an SP the setup_authentcommand does the SP-related kerberos setup,  
which is based on the IP labels found in the SDR. Since the SDR does not  
allow multiple IP labels to be defined on the same interface, whereas HACMP  
needs to have multiple IP labels on one interface during IPAT, the kerberos  
setup for HACMP has to be redone, every time the setup_authentcommand is  
run explicitly or implicitly through the setup_servercommand.  
You can either do that manually, or use the cl_setup_kerberostool. To  
manually add the kerberos principals, use the kadmincommand. Necessary  
principals for kerberized operation in enhanced security mode are the  
(remote) rcmd principals and the godm principals. As always, a kerberos  
principal consists of a name, godm for example, an IP label, like  
hadave1_stby and a realm, so that the principal in its full length would look  
Now after adding all the needed principals to the kerberos database, you  
must also add them to the /etc/krb-srvtab file on the nodes. To do that, you  
will have to extract them from the database and copy them out to the nodes,  
replacing their kerberos file.  
Now you can extend root’s .klogin file and /etc/krb.realms file to reflect the  
new principals, and copy these files out to the node as well.  
Special RS/6000 SP Topics 189  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
After setting the cluster’s security settings to enhanced for all these nodes,  
you can verify that it is working as expected, for example, by running clverify,  
which goes out to the nodes and checks the consistency of files.  
9.3 VSDs - RVSDs  
VSDs (Virtual Shared Disks) and RVSDs (Recoverable Virtual Shared Disks)  
are SP-specific facilities that you are likely to use in an HACMP environment.  
9.3.1 Virtual Shared Disk (VSDs)  
Virtual Shared Disk (VSD) allows data in logical volumes on disks physically  
connected to one node to be transparently accessed by other nodes.  
Importantly, VSD supports only raw logical volumes, not file systems. The  
VSD facility is included in the ssp.csd.vsd fileset of PSSP.  
IBM developed VSD to enable Oracle’s parallel database on the SP. Oracle’s  
database architecture is strongly centralized. Any processing element, or  
node, must be able to “see” the entire database. In the case of the parallel  
implementation of Oracle, all nodes must have access to all disks of the  
database, regardless of where those disks are physically attached.  
Node X  
Node Y  
Application  
Application  
Cache  
LVM  
Cache  
LVM  
VSD  
VSD  
IP  
IP  
Disk DD  
Net DD  
Net DD  
Disk DD  
IP Network (SP Switch)  
lv_X  
lv_Y  
Figure 17. VSD Architecture  
190  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
With reference to Figure 17 above, imagine two nodes, Node X and Node Y,  
running the same application. The nodes are connected by the switch and  
have locally-attached disks. On Node X’s disk resides a volume group  
containing the raw logical volume lv_X. Similarly, Node Y has lv_Y. For the  
sake of illustration, let us suppose that lv_X and lv_Y together constitute an  
Oracle Parallel Server database to which the application on each node makes  
I/O requests.  
The application on Node X requests a piece of data in the database. After the  
node’s Virtual Memory Manager (VMM) determines that the data is not in  
memory, it talks not to the regular Logical Volume Manager (LVM), but rather  
to the VSD device driver. The VSD device driver is loaded as a kernel  
extension. Thus VSDs configured in the SP are known to the appropriate  
nodes at the kernel level.  
The VSD device driver can fetch the data from one of three places:  
1. From the VSD cache, if the data is still there from previous requests. VSD  
cache is shared by all VSDs configured on a node. Data is stored in 4KB  
blocks, a size optimized for Oracle Parallel Server. If your I/O patterns  
involve I/O operations larger than 4KB, we recommend disabling VSD  
cache, because its management becomes counterproductive.  
2. From lv_X, in which case the VSD device driver exploits Node X’s normal  
LVM and Disk Device Driver (Disk DD) pathway to fetch the data.  
3. From lv_Y, in which case the VSD device driver issues the request through  
the IP and Network Device Driver (Net DD) pathway to access Node Y. For  
performance, VSD uses its own stripped-down IP protocol. Once the  
request is passed up through Node Y’s Net DD and IP layers, Node Y’s  
VSD device driver accesses the data either from VSD cache or from lv_Y.  
The VSD server node uses the buddy buffer to temporarily store data for I/O  
operations originating at a client node, and to handle requests that are  
greater than the IP message size. In contrast to the data in the cache buffer,  
the data in a buddy buffer is purged immediately after the I/O operation  
completes. Buddy buffers are used only when a shortage in the switch buffer  
pool occurs, or, on certain networks with small IP message sizes (for  
example, Ethernet). The maximum and minimum size for the buddy buffer  
must be defined when the VSD is created. For best performance, you must  
ensure that your buddy buffer limits accommodate your I/O transaction sizes  
to minimize the packetizing workload of the VSD protocol. Buddy buffers are  
discussed in detail in IBM Parallel System Support Programs for AIX  
Managing Shared Disks, SA22-7279.  
Special RS/6000 SP Topics 191  
Download from Www.Somanuals.com. All Manuals Search And Download.  
The VSDs in this scenario are mapped to the raw logical volumes lv_X and  
lv_Y. Node X is a client of Node Y’s VSD, and vice versa. Node X is also a  
direct client of its own VSD (lv_X), and Node Y is a direct client of VSD lv_Y.  
VSD configuration is flexible. An interesting property of the architecture is  
that a node can be a client of any other node’s VSD(s), with no dependency  
on that client node owning a VSD itself. You could set up three nodes with  
powerful I/O capacity to be VSD servers, and ten application nodes, with no  
disk other than for AIX, PSSP, and the application executables, as clients of  
the VSDs on these server nodes.  
VSDs are defined in the SDR and managed by either SP SMIT panels or the  
VSD Perspective. VSDs can be in one of five states as shown in Figure 18 on  
Undefined  
define  
undefine  
VSD information is  
available in the SDR  
Defined  
cfgvsd  
cfgvsd  
Open/close and I/O  
requests fail  
Stopped  
preparevsd  
stopvsd  
I/O requests queued and  
open/close request serviced  
Suspended  
resumevsd  
suspendvsd  
Open/close and  
I/O requests serviced  
Active  
Available  
Figure 18. VSD State Transitions  
This figure shows the possible states of a VSD and the commands used to  
move between states. VSD configuration changes, or manual recovery of a  
failed VSD, require you to move the VSD between various states.  
The distributed data access aspect of VSD scales well. The SP Switch itself  
provides a very high-bandwidth, scalable interconnect between VSD clients  
and servers, while the VSD layers of code are efficient. The performance  
192  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
impact of servicing a local I/O request through VSD relative to the normal  
VMM/LVM pathway is very small. IBM supports any IP network for VSD, but  
we recommend the switch for performance.  
VSD provides distributed data access, but not a locking mechanism to  
preserve data integrity. A separate product such as Oracle Parallel Server  
must provide the global locking mechanism.  
9.3.2 Recoverable Virtual Shared Disk  
Recoverable Virtual Shared Disk (RVSD) adds availability to VSD. RVSD  
allows you to twin-tail disks, that is, physically connect the same group of  
disks to two or more nodes, and provide transparent failover of VSDs among  
the nodes. RVSD is a separately-priced IBM LPP.  
Fast IP Network  
VSD Server  
VSD Server  
VSD Client  
Node X  
Node Y  
Node Z  
rvsd_X  
rvsd_Y  
VSD  
VSD  
VSD  
rvsd_X  
(lv_X)  
rvsd_Y  
(lv_Y)  
Figure 19. RVSD Function  
With reference to Figure 19 above, Nodes X, Y, and Z form a group of nodes  
using VSD. RVSD is installed on Nodes X and Y to protect VSDs rvsd_X and  
rvsd_Y. Nodes X and Y physically connect to each other’s disk subsystems  
where the VSDs reside. Node X is the primary server for rvsd_X and the  
secondary server for rvsd_Y, and vice versa for Node Y. Should Node X fail,  
RVSD will automatically fail over rvsd_X to Node Y. Node Y will take  
ownership of the disks, varyon the volume group containing rvsd_X and make  
the VSD available. Node Y then serves both rvsd_X and rvsd_Y. Any I/O  
Special RS/6000 SP Topics 193  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
operation that was in progress, as well as new I/O operations against rvsd_X,  
are suspended until failover is complete. When Node X is repaired and  
rebooted, RVSD switches the rvsd_X back to its primary, Node X.  
The RVSD subsystems are shown in Figure 20 on page 194. The rvsd  
daemon controls recovery. It invokes the recovery scripts whenever there is a  
change in the group membership, which it is recognizing through the use of  
Group Services, which in turn relies on information from Topology Services.  
When a failure occurs, the rvsd daemon notifies all surviving providers in the  
RVSD node group, so they can begin recovery. Communication adapter  
failures are treated the same as node failures.  
The hc daemon is also called the Connection Manager. It supports the  
development of recoverable applications. The hc daemon maintains a  
membership list of the nodes that are currently running hc daemons and an  
incarnation number that is changed every time the membership list changes.  
The hc daemon shadows the rvsd daemon; recording the same changes in  
state and management of VSD that rvsd records. The difference is that hc  
only records these changes after rvsd processes them, to assure that RVSD  
recovery activities begin and complete before the recovery of hc client  
applications takes place. This serialization helps ensure data integrity.  
RSVD Daemons  
hc  
rvsd  
Group Services  
Topology Services  
Adapter  
Membership  
Node  
Membership  
Figure 20. RVSD Subsystem and HA Infrastructure  
IBM Certification Study Guide AIX HACMP  
194  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
9.4 SP Switch as an HACMP Network  
One of the fascinating things with an RS/6000 SP is the switch network. It has  
developed over time; so, currently there are two types of switches at  
customer sites. The “older” HPS or HiPS switch (High Performance Switch),  
also known as the TB2 switch, and the “newer” SP Switch, also known as the  
TB3 switch.  
The HPS switch is no longer supported with PSSP Version 3.1, and the same  
applies to HACMP/ES Version 4.3.  
The two different types of switches differ in their availability design from the  
hardware point of view significantly. For example, any fault service action on  
the HPS switch caused a total network disruption for a small fraction of time.  
For example, running an Estart to get new nodes up on the switch affected  
running nodes.  
The SP switch however, was designed to do all actions regarding the switch  
fabric on the link level, so only the selected node is affected. All the others  
continue working without even noticing that something has happened on the  
switch network.  
9.4.1 Switch Basics Within HACMP  
Although it has already been mentioned in other places, the following is a  
short summary of basics you have to remember when you configure a switch  
as a network to HACMP.  
• As the switch network is a point-to-point network, you must configure it to  
HACMP as a private network.  
• For IPAT to work on the switch network, you must enable ARP on the  
switch network. However, hardware address takeover is not supported  
with the switch.  
• If you configure IPAT, the service and boot addresses are ifconfig alias  
addresses on the css0adapter. Since there is currently no support for  
more than one switch adapter in a node, this is the way HACMP covers  
the normally-needed second adapter for redundancy.  
• The base address for the switch adapter, i.e. the switch address known to  
the SDR, should not be configured into an HACMP network. This would  
lead to confusion for the PSSP switch management software.  
• The netmask associated with the css0 base IP address is used as the  
netmask for all HACMP SP Switch network addresses.  
Special RS/6000 SP Topics 195  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
9.4.2 Eprimary Management  
The SP switch has an internal primary backup concept, where the primary  
node, known as the Eprimary, is backed up automatically by a backup node.  
So, in case any serious failure happens on the primary, it will resign from  
work, and the backup node will take over the switch network handling,  
keeping track of routes, working on events, and so on.  
HACMP/ES used to have an Eprimary management function with versions  
below 4.3; so, if you upgrade to Version 4.3 and also upgrade your switch to  
the SP switch, and you had configured Eprimary management previously  
within the HACMP definitions, you have to unmanage it.  
To check whether the Eprimary is set to be managed, issue the following  
command:  
odmget -q’name=EPRIMARY’ HACMPsp2  
If the switch is set to MANAGE, before changing to the new switch, run the  
script:  
/usr/es/sbin/cluster/events/utils/cl_HPS_Eprimary unmanage  
As the SP switch has its availability concept built-in, there is no need to do it  
outside the PSSP software, so, HACMP doesn’t have to take care of it any  
more.  
9.4.3 Switch Failures  
As mentioned before, a node in the SP is still restricted to have a maximum of  
one switch adapter installed. Therefore, even with the software being able to  
assign a new primary node within the SP and outside of HACMP, the switch  
adapter is still a single point of failure.  
If the switch adapter in a node resigns from work due to a software or  
hardware problem, the switch network is down for that node.  
If any application running on that node relies on the switch network, this  
means that the application has virtually died on that node. Therefore, it might  
be advisable to promote the switch network failure into a node failure, as  
Recovery” on page 46. HACMP would be able to recognize the network  
failure when you configure the switch network as an HACMP network, and  
thus would react with a network_down event, which in turn would shut down  
the node from HACMP, causing a takeover.  
196  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
In case this node was the Eprimary node on the switch network, and it is an  
SP switch, then the RS/6000 SP software would have chosen a new Eprimary  
independently from the HACMP software as well.  
Special RS/6000 SP Topics 197  
Download from Www.Somanuals.com. All Manuals Search And Download.  
198  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Chapter 10. HACMP Classic vs. HACMP/ES vs. HANFS  
So, why would you prefer to install one version of HACMP instead of  
another? This chapter summarizes the differences between them, to give you  
an idea in which situation one or the other best matches your needs. The  
certification test itself does not refer to these different HACMP flavors, but it is  
useful to know the differences anyway.  
The following paragraphs are based on the assumption that you are using  
Version 4.3. For an overview of previous Versions and their corresponding  
AIX levels, as well as the supported hardware to run on, see Table 3 on page  
10.1 HACMP for AIX Classic  
High Availability Cluster Multi-Processing for AIX (HACMP for AIX) Version  
4.3 comes in two flavors. One of them directly derives from previous versions,  
and therefore is called Classic, and the other, which utilizes another  
technology for heartbeating, is called HACMP Extended Scalability  
(HACMP/ES) see below for details.  
Basically, these two versions differ only in the way the cluster manager keeps  
track of the status of nodes, adapters and networks. In the Classic Version,  
this is done through the use of Network Interface Modules.  
Network Interface Modules (NIMs) monitor the nodes and network  
interfaces associated with a cluster. Each  
network module monitors one cluster network  
using one kind of communication protocol (for  
example, Ethernet or FDDI). Each network  
module is responsible for maintaining  
keepalive traffic with neighboring nodes as  
directed by the Cluster Controller, for  
providing a link to other nodes on the network  
it monitors, and for initiating adapter swaps on  
certain networks.  
10.2 HACMP for AIX / Enhanced Scalability  
HACMP/ES no longer used NIMs, but utilizes a technolgy that was originally  
developed on the RS/6000 SPs. Since PSSP Version 2.2. RS/6000 SP  
Systems come with the Phoenix technology for managing availability of the  
nodes. This technology was already designed as a basic instrument for  
© Copyright IBM Corp. 1999  
199  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
handling membership and event management by using heartbeats. On the  
SP, the original High Availability infrastructure was built on this technology,  
and HACMP/ES Version 4.3. is now another instance relying on it. As of AIX  
4.3.2 and PSSP 3.1, the High Availability infrastructure, which previously was  
tightly coupled to PSSP, was externalized into a package called RISC System  
Cluster Technology (RSCT). This package can be installed and run, not only  
on SP nodes, but also on regular RS/6000 systems. This allows HACMP/ES  
to also be available on non-SP RS/6000s as of Version 4.3.  
10.2.1 IBM RISC System Cluster Technology (RSCT)  
The High Availability services previously packaged with the IBM PSSP for  
AIX Availability Services, also known as the ssp.ha fileset, are now an  
integral part of the HACMP/ES software. The IBM RS/6000 Cluster  
Technology (RSCT) services provide greater scalability, notify distributed  
subsystems of software failure, and coordinate recovery and synchronization  
among all subsystems in the software stack.  
Packaging these services with HACMP/ES makes it possible to run this  
software on all RS/6000s, not just on SP nodes.  
RSCT Services include the following components:  
Event Manager  
A distributed subsystem providing a set of high  
availability services. It creates events by matching  
information about the state of system resources with  
information about resource conditions of interest to  
client programs. Client programs, in turn, can use event  
notifications to trigger recovery from system failures.  
Group Services  
A system-wide, fault-tolerant, and highly available  
facility for coordinating and monitoring changes to the  
state of an application running on a set of nodes. Group  
Services helps both in the design and implementation of  
fault-tolerant applications and in the consistent recovery  
of multiple applications. It accomplishes these two  
distinct tasks in an integrated framework.  
Topology Service A facility for generating heartbeats over multiple  
networks and for providing information about adapter  
membership, node membership, and routing. Adapter  
and node membership provide indications of adapter  
and node failures respectively. Reliable Messaging uses  
the routing information to route messages between  
nodes around adapter failures.  
200  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
         
See Part 4 of HACMP for AIX, Version 4.3: Enhanced Scalability Installation  
and Administration Guide, SC23-4284, for more information on these  
services.  
10.2.2 Enhanced Cluster Security  
With HACMP Version 4.3 comes an option to switch security Mode between  
Standard and Enhanced.  
Standard  
Synchronization is done through the /.rhostsremote command  
facilities. To avoid the compromised security that the presence  
of this file presents, the administrator is strongly encouraged to  
remove these files after the synchronization/verification is done.  
Enhanced Kerberos authentication is used for remote commands. That  
means the kerberos daemons can decide whether a remote host  
is who they claim to be. This is done by granting access on the  
basis of tickets, which are provided only to those hosts having  
the correct identification.  
10.3 High Availability for Network File System for AIX  
The HANFS for AIX software provides a reliable NFS server capability by  
allowing a backup processor to recover current NFS activity should the  
primary NFS server fail.  
The HANFS for AIX software supports only two nodes in a cluster.  
HANFS for AIX is based on High Availability Cluster Multi-Processing for AIX,  
Version 4.3 (HACMP for AIX Classic) product architecture, which ensures  
that critical resources, configured as part of a cluster, are highly available for  
processing. The HANFS for AIX software extends HACMP for AIX by taking  
advantage of AIX extensions to the standard NFS functionality that enable it  
to handle duplicate requests correctly and restore lock state during NFS  
server failover and reintegration.  
Note  
A cluster cannot be mixed, that is, have some nodes running the HANFS  
for AIX software and other nodes running the HACMP for AIX software. A  
single cluster must either have all nodes running the HANFS for AIX  
software or all nodes running the HACMP for AIX software. Distinct HANFS  
and HACMP clusters, however, are allowed on the same physical network.  
HACMP Classic vs. HACMP/ES vs. HANFS 201  
Download from Www.Somanuals.com. All Manuals Search And Download.  
       
10.4 Similarities and Differences  
All three products have the basic structure in common. They all use the same  
concepts and structures. So, a cluster or a network, in the HACMP context, is  
the same, no matter what product is being used. There is always a Cluster  
Manager controlling the node, keeping track of the cluster’s status, and  
triggering events. The differences are in the technologies being used  
underneath, or in some special cases, the features available.  
The technique of keeping track of the status of a cluster by sending and  
receiving heartbeat messages is the major difference between HACMP  
Classic and HACMP/ES Version 4.3. HACMP Classic uses the network  
modules (NIMs) for this purpose. These communicate their results straight  
through to the HACMP Cluster Manager. HACMP/ES, uses the facilities of  
RSCT, namely Topology Services, Group Services, and Event Management,  
for its heartbeating. Since Version 4.3, the restriction to run HACMP/ES on  
RS/6000 SP systems only has been withdrawn. However, if you run it on an  
RS/6000 SP, you need to have PSSP Version 3.1 installed. As the HPS  
Switch is no longer supported with PSSP Version 3.1, you need to upgrade to  
the SP Switch, in case you haven’t already, or you will have a switchless  
system.  
You can still run HACMP Classic on RS/6000 SP Nodes, just as on  
standalone RISC System/6000s. It has no references into the PSSP code  
whatsoever.  
HANFS for AIX Version 4.3 is basically, a modified HACMP Classic,  
enhanced with the capability of the takeover node to recover current NFS  
activity, should the primary NFS server fail. By means of AIX extensions to  
standard NFS functionality, HANFS for AIX is enabled to handle duplicate  
requests correctly or restore the lock state in case of an NFS server failover  
or reintegration. Remember though, that HANFS is somewhat restricted, in  
that it only supports two-node clusters and cascading resource groups.  
10.5 Decision Criteria  
Your decision of what type of high availability software you are going to use  
can be based on various criteria. Existing hardware is one of them.  
If you still use the “old” HPS Switch and don’t want to lose its functionality,  
you are bound to the use of PSSP 2.4 or lower. Therefore HACMP Classic or  
HACMP/ES up to Version 4.2.2 only is the choice for you.  
202  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
For switchless RS/6000 SP systems or SPs with the newer SP Switch, the  
decision will be based on a more functional level.  
Event Management is much more flexible in HACMP/ES, since you can  
define custom events. These events can act on anything that haemdcan  
detect, which is virtually anything measurable on an AIX system. How to  
customize events is explained in great detail in the redbook HACMP  
Enhanced Scalability, SG24-2081.  
If you have an NFS server that you need to make highly available, especially  
if it is heavily used, and NFS file locking is a major issue, you will need to run  
HANFS for AIX.  
HACMP Classic vs. HACMP/ES vs. HANFS 203  
Download from Www.Somanuals.com. All Manuals Search And Download.  
204  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Appendix A. Special Notices  
This publication is intended to help System Administrators, System Engineers  
and other System Professionals to pass the IBM HACMP Certification Exam.  
The information in this publication is not intended as the specification for any  
of the following programming interfaces: HACMP, HACMP/ES, HANFS or  
HACWS. See the PUBLICATIONS section of the IBM Programming  
Announcement for those products for more information about what  
publications are considered to be product documentation.  
References in this publication to IBM products, programs or services do not  
imply that IBM intends to make these available in all countries in which IBM  
operates. Any reference to an IBM product, program, or service is not  
intended to state or imply that only IBM’s product, program, or service may be  
used. Any functionally equivalent program that does not infringe any of IBM’s  
intellectual property rights may be used instead of the IBM product, program  
or service.  
Information in this book was developed in conjunction with use of the  
equipment specified, and is limited in application to those specific hardware  
and software products and levels.  
IBM may have patents or pending patent applications covering subject matter  
in this document. The furnishing of this document does not give you any  
license to these patents. You can send license inquiries, in writing, to the IBM  
Director of Licensing, IBM Corporation, 500 Columbus Avenue, Thornwood,  
NY 10594 USA.  
Licensees of this program who wish to have information about it for the  
purpose of enabling: (i) the exchange of information between independently  
created programs and other programs (including this one) and (ii) the mutual  
use of the information which has been exchanged, should contact IBM  
Corporation, Dept. 600A, Mail Drop 1329, Somers, NY 10589 USA.  
Such information may be available, subject to appropriate terms and  
conditions, including in some cases, payment of a fee.  
The information contained in this document has not been submitted to any  
formal IBM test and is distributed AS IS. The information about non-IBM  
(“vendor”) products in this manual has been supplied by the vendor and IBM  
assumes no responsibility for its accuracy or completeness. The use of this  
information or the implementation of any of these techniques is a customer  
responsibility and depends on the customer's ability to evaluate and integrate  
them into the customer's operational environment. While each item may have  
© Copyright IBM Corp. 1999  
205  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
been reviewed by IBM for accuracy in a specific situation, there is no  
guarantee that the same or similar results will be obtained elsewhere.  
Customers attempting to adapt these techniques to their own environments  
do so at their own risk.  
Any pointers in this publication to external Web sites are provided for  
convenience only and do not in any manner serve as an endorsement of  
these Web sites.  
Any performance data contained in this document was determined in a  
controlled environment, and therefore, the results that may be obtained in  
other operating environments may vary significantly. Users of this document  
should verify the applicable data for their specific environment.  
The following document contains examples of data and reports used in daily  
business operations. To illustrate them as completely as possible, the  
examples contain the names of individuals, companies, brands, and  
products. All of these names are fictitious and any similarity to the names and  
addresses used by an actual business enterprise is entirely coincidental.  
Reference to PTF numbers that have not been released through the normal  
distribution process does not imply general availability. The purpose of  
including these reference numbers is to alert IBM customers to specific  
information relative to the implementation of the PTF when it becomes  
available to each customer according to the normal IBM PTF distribution  
process.  
The following terms are trademarks of the International Business Machines  
Corporation in the United States and/or other countries:  
AIX  
Application System/400  
AS/400  
AT  
BookManager  
HACMP/6000  
IBM  
NetView  
POWERserver  
RS/6000  
SP1  
CT  
Home Director  
Micro Channel  
POWERparallel  
RISC System/6000  
SP  
System/390  
XT  
Ultrastar  
400  
The following terms are trademarks of other companies:  
C-bus is a trademark of Corollary, Inc.  
206  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Java and HotJava are trademarks of Sun Microsystems, Incorporated.  
Microsoft, Windows, Windows NT, and the Windows 95 logo are trademarks  
or registered trademarks of Microsoft Corporation.  
PC Direct is a trademark of Ziff Communications Company and is used  
by IBM Corporation under license.  
Pentium, MMX, ProShare, LANDesk, and ActionMedia are trademarks or  
registered trademarks of Intel Corporation in the U.S. and other  
countries.  
Network File System and NFS are trademarks of SUN Microsystems, Inc.  
SUN Microsystems is a trademark of SUN Microsystems, Inc.  
UNIX is a registered trademark in the United States and other  
countries licensed exclusively through X/Open Company Limited.  
Other company, product, and service names may be trademarks or  
service marks of others.  
Special Notices 207  
Download from Www.Somanuals.com. All Manuals Search And Download.  
208  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Appendix B. Related Publications  
The publications listed in this section are considered particularly suitable for a  
more detailed discussion of the topics covered in this redbook.  
B.1 International Technical Support Organization Publications  
For information on ordering these ITSO publications see “How to Get ITSO  
• A Practical Guide to Serial Storage Architecture for AIX, SG24-4599  
• HACMP Enhanced Scalability, SG24-2081  
• HACMP Enhanced Scalability Handbook, SG24-5328  
HACMP Enhanced Scalability: User-Defined Events, SG24-5327  
HACMP/6000 Customization Examples, SG24-4498  
High Availability on the RISC System/6000 Family, SG24-4551  
Inside the RS/6000 SP, SG24-5145  
Monitoring and Managing IBM SSA Disk Subsystems, SG24-5251  
AIX Version 4.3 Migration Guide, SG24-5116  
B.2 Redbooks on CD-ROMs  
Redbooks are also available on CD-ROMs. Order a subscription and  
receive updates 2-4 times a year at significant savings.  
CD-ROM Title  
Subscription  
Number  
Collection Kit  
Number  
System/390 Redbooks Collection  
SBOF-7201  
SBOF-7370  
SBOF-7240  
SBOF-6899  
SBOF-6898  
SBOF-7270  
SBOF-7230  
SBOF-7205  
SBOF-8700  
SBOF-7290  
SK2T-2177  
SK2T-6022  
SK2T-8038  
SK2T-8039  
SK2T-8044  
SK2T-2849  
SK2T-8040  
SK2T-8041  
SK2T-8043  
SK2T-8037  
Networking and Systems Management Redbooks Collection  
Transaction Processing and Data Management Redbook  
Lotus Redbooks Collection  
Tivoli Redbooks Collection  
AS/400 Redbooks Collection  
RS/6000 Redbooks Collection (HTML, BkMgr)  
RS/6000 Redbooks Collection (PostScript)  
RS/6000 Redbooks Collection (PDF Format)  
Application Development Redbooks Collection  
© Copyright IBM Corp. 1999  
209  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
B.3 Other Publications  
These publications are also relevant as additional sources of information:  
IBM RS/6000 SP: Planning, Volume 2, Control Workstation and Software  
Environment, GA22-7281  
IBM PSSP for AIX: Installation and Migration Guide, GA22-7347  
IBM PSSP for AIX: Managing Shared Disks, SA22-7279  
Adapters, Devices, and Cable Information for Multiple Bus Systems,  
SA38-0516  
Adapters, Devices, and Cable Information for Micro Channel Bus  
Systems, SA38-0533  
PCI Adapter Placement Reference, SA38-0538  
AIX Commands Reference, SBOF-1877  
• AIX Performance Monitoring and Tuning Guide, SC23-2365  
AIX HACMP for AIX,Version 4.3: Concepts and Facilities, SC23-4276  
AIX HACMP for AIX,Version 4.3: Planning Guide, SC23-4277  
AIX HACMP for AIX,Version 4.3: Installation Guide, SC23-4278  
AIX HACMP for AIX,Version 4.3: Administration Guide, SC23-4279  
AIX HACMP for AIX,Version 4.3: Troubleshooting Guide, SC23-4280  
AIX HACMP for AIX,Version 4.3: Programming Locking Applications,  
SC23-4281  
AIX HACMP for AIX,Version 4.3: Programming Client Applications,  
SC23-4282  
AIX HACMP for AIX,Version 4.3: HANFS for AIX Installation and  
Administration Guide, SC23-4283  
AIX HACMP for AIX,Version 4.3: Enhanced Scalability Installation and  
Administration Guide, SC23-4284  
AIX HACMP for AIX,Version 4.3: Master Index and Glossary, SC23-4285  
210  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
How to Get ITSO Redbooks  
This section explains how both customers and IBM employees can find out about ITSO redbooks,  
CD-ROMs, workshops, and residencies. A form for ordering books and CD-ROMs is also provided.  
This information was current at the time of publication, but is continually subject to change. The latest  
information may be found at http://www.redbooks.ibm.com/.  
How IBM Employees Can Get ITSO Redbooks  
Employees may request ITSO deliverables (redbooks, BookManager BOOKs, and CD-ROMs) and  
information about redbooks, workshops, and residencies in the following ways:  
Redbooks Web Site on the World Wide Web  
http://w3.itso.ibm.com/  
PUBORDER – to order hardcopies in the United States  
Tools Disks  
To get LIST3820s of redbooks, type one of the following commands:  
TOOLCAT REDPRINT  
TOOLS SENDTO EHONE4 TOOLS2 REDPRINT GET SG24xxxx PACKAGE  
TOOLS SENDTO CANVM2 TOOLS REDPRINT GET SG24xxxx PACKAGE (Canadian users only)  
To get BookManager BOOKs of redbooks, type the following command:  
TOOLCAT REDBOOKS  
To get lists of redbooks, type the following command:  
TOOLS SENDTO USDIST MKTTOOLS MKTTOOLS GET ITSOCAT TXT  
To register for information on workshops, residencies, and redbooks, type the following command:  
TOOLS SENDTO WTSCPOK TOOLS ZDISK GET ITSOREGI 1998  
REDBOOKS Category on INEWS  
Online – send orders to: USIB6FPL at IBMMAIL or DKIBMBSH at IBMMAIL  
Redpieces  
For information so current it is still in the process of being written, look at "Redpieces" on the  
Redbooks Web Site (http://www.redbooks.ibm.com/redpieces.html). Redpieces are redbooks in  
progress; not all redbooks become redpieces, and sometimes just a few chapters will be published  
this way. The intent is to get the information out much more quickly than the formal publishing process  
allows.  
© Copyright IBM Corp. 1999  
211  
Download from Www.Somanuals.com. All Manuals Search And Download.  
     
How Customers Can Get ITSO Redbooks  
Customers may request ITSO deliverables (redbooks, BookManager BOOKs, and CD-ROMs) and  
information about redbooks, workshops, and residencies in the following ways:  
Online Orders – send orders to:  
IBMMAIL  
Internet  
In United States  
In Canada  
Outside North America  
usib6fpl at ibmmail  
caibmbkz at ibmmail  
dkibmbsh at ibmmail  
Telephone Orders  
United States (toll free)  
Canada (toll free)  
1-800-879-2755  
1-800-IBM-4YOU  
Outside North America  
(long distance charges apply)  
(+45) 4810-1020 - German  
(+45) 4810-1620 - Italian  
(+45) 4810-1270 - Norwegian  
(+45) 4810-1120 - Spanish  
(+45) 4810-1170 - Swedish  
(+45) 4810-1320 - Danish  
(+45) 4810-1420 - Dutch  
(+45) 4810-1540 - English  
(+45) 4810-1670 - Finnish  
(+45) 4810-1220 - French  
Mail Orders – send orders to:  
IBM Publications  
Publications Customer Support 144-4th Avenue, S.W.  
P.O. Box 29570  
Raleigh, NC 27626-0570  
USA  
IBM Publications  
IBM Direct Services  
Sortemosevej 21  
DK-3450 Allerød  
Denmark  
Calgary, Alberta T2P 3N5  
Canada  
Fax – send orders to:  
United States (toll free)  
Canada  
Outside North America  
1-800-445-9269  
1-800-267-4455  
(+45) 48 14 2207 (long distance charge)  
1-800-IBM-4FAX (United States) or (+1) 408 256 5422 (Outside USA) – ask for:  
Index # 4421 Abstracts of new redbooks  
Index # 4422 IBM redbooks  
Index # 4420 Redbooks for last six months  
On the World Wide Web  
Redbooks Web Site  
http://www.redbooks.ibm.com  
IBM Direct Publications Catalog  
http://www.elink.ibmlink.ibm.com/pbl/pbl  
Redpieces  
For information so current it is still in the process of being written, look at "Redpieces" on the  
Redbooks Web Site (http://www.redbooks.ibm.com/redpieces.html). Redpieces are redbooks in  
progress; not all redbooks become redpieces, and sometimes just a few chapters will be published  
this way. The iinntteennttiissttooggeettththeeininfoforrmmaatitoionnoouuttmmuucchhmqourieckqeuricthkalynththaenftohremfaolrmpuabl lpisuhbilnisghpinrgocpersoscess  
allows.  
212  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
IBM Redbook Order Form  
Please send me the following:  
Title  
Order Number  
Quantity  
First name  
Company  
Address  
City  
Last name  
Postal code  
Country  
Telephone number  
Telefax number  
VAT number  
Invoice to customer number  
Credit card number  
Credit card expiration date  
Card issued to  
Signature  
We accept American Express, Diners, Eurocard, Master Card, and Visa. Payment by credit card not  
available in all countries. Signature mandatory for credit card payment.  
213  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
214  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
List of Abbreviations  
GODM  
GUI  
Global Object Data  
Manager  
AIX  
Advanced Interactive  
Executive  
Graphical User  
Interface  
APA  
All Points Addressable  
APAR  
Authorized Program  
Analysis Report  
HACMP  
HANFS  
HCON  
IBM  
High Availability Cluster  
Multi-Processing  
The description of a  
problem to be fixed by  
IBM defect support.  
This fix is delivered in a  
PTF (see below).  
High Availability  
Network File System  
Host Connection  
Program  
ARP  
Address Resolution  
Protocol  
International Business  
Machines Corporation  
ASCII  
American Standard  
Code for Information  
Interchange  
I/O  
IP  
Input/Output  
Interface Protocol  
IPL  
Initial Program Load  
(System Boot)  
AS/400  
CDF  
Application System/400  
Cumulative Distribution  
Function  
ITSO  
International Technical  
Support Organization  
CD-ROM  
Compact Disk - Read  
Only Memory  
JFS  
KA  
KB  
Kb  
Journaled File System  
Keepalive Packet  
kilobyte  
CLM  
Cluster Lock Manager  
CLVM  
Concurrent Logical  
Volume Manager  
kilobit  
CPU  
CRM  
Central Processing Unit  
LAN  
LU  
Local Area Network  
Concurrent Resource  
Manager  
Logical Unit (SNA  
definition)  
DE  
Differential Ended  
Data Link Control  
Deadman Switch  
LUN  
LVM  
Logical Unit (RAID  
definition)  
DLC  
DMS  
DNS  
DSMIT  
Logical Volume  
Manager  
Domain Name Service  
MAC  
MB  
Medium Access Control  
megabyte  
Distributed System  
Management Interface  
Tool  
MIB  
Management  
Information Base  
FDDI  
Fiber Distributed Data  
Interface  
MTBF  
Mean Time Between  
Failure  
F/W  
GB  
Fast and Wide (SCSI)  
Gigabyte  
© Copyright IBM Corp. 1999  
215  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
NETBIOS  
Network Basic  
SPOF  
Single Point of Failure  
Input/Output System  
SPX/IPX  
Sequenced Package  
Exchange/Internetwork  
Packet Exchange  
NFS  
NIM  
Network File System  
Network Interface  
Module (This is the  
definition of NIM in the  
HACMP context. NIM in  
the AIX 4.1 context  
stands for Network  
Installation Manager).  
SRC  
System Resource  
Controller  
SSA  
Serial Storage  
Architecture  
TCP  
Transmission Control  
Protocol  
NIS  
Network Information  
Service  
TCP/IP  
Transmission Control  
Protocol/Interface  
Protocol  
NVRAM  
Non-Volatile Random  
Access Memory  
UDP  
User Datagram  
Protocol  
ODM  
POST  
PTF  
Object Data Manager  
Power On Self Test  
UPS  
Uninterruptible Power  
Supply  
Program Temporary Fix  
A fix to a problem  
described in an APAR  
(see above).  
VGDA  
VGSA  
WAN  
Volume Group  
Descriptor Area  
Volume Group Status  
Area  
RAID  
Redundant Array of  
Independent (or  
Inexpensive) Disks  
Wide Area Network  
RISC  
SCSI  
SLIP  
Reduced Instruction  
Set Computer  
Small Computer  
Systems Interface  
Serial Line Interface  
Protocol  
SMIT  
SMP  
System Management  
Interface Tool  
Symmetric  
Multi-Processor  
SMUX  
SNA  
SNMP (see below)  
Multiplexor  
Systems Network  
Architecture  
SNMP  
SOCC  
Simple Network  
Management Protocol  
Serial Optical Channel  
Converter  
216  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Index  
using 179  
cldare command 172  
clfindres 173  
clinfo 156  
cllockd 155  
clsmuxpd 155  
clstat 152  
Symbols  
/.rhosts file  
editing 59  
/etc/hosts file  
and adapter label 38  
/sbin/rc.boot file 146  
/usr/sbin/cluster/godm daemon 59  
clstrmgr 155  
Cluster ID 101  
Cluster Manager 30  
cluster nodes  
A
abbreviations 215  
Abnormal Termination 158  
acronyms 215  
synchronizing 111  
Cluster Planning  
cluster services  
starting  
Adapter Failure 134  
Adapter Function 38  
Adapter Hardware Address 104  
Adapter Identifier 104  
adapter label 38  
on clients 159  
stopping  
on clients 159  
Cluster Snapshot 113  
Cluster Status Events 121  
Cluster Topology 100  
cluster topology  
defining cluster ID 101  
Clverify 111  
Concurrent Access 82  
concurrent access mode  
quorum 90  
Concurrent Disk Access Configuration 34  
Concurrent Resource Group 30  
config_too_long 121, 144  
Configuring Target Mode SCSI 64  
Configuring Target Mode SSA 65  
CPU Failure 137  
adding  
cluster definition 101  
user accounts 179  
advantages of SSA 25  
AIX Parameter Settings 56  
Application Failure 141  
application failure 47  
application server 41  
Application Servers 110  
ARP cache 40  
ATM 13  
B
Backup Strategies 176  
boot adapter 39  
CPU Options  
cron 58  
cross mount 41  
cross mounting  
NFS filesystems 126  
C-SPOC 165  
C
Cabling Considerations 60  
capacity requirements  
Cascading Resource Group 29  
cascading resource groups  
NFS crossmounting issues 126  
changing  
user accounts 180  
cl_lsuser command  
D
daemons  
godm 59  
DARE 169  
Deadman Switch 145  
defining  
using 179  
cl_mkuser command  
hardware addresses 40  
© Copyright IBM Corp. 1999  
217  
Download from Www.Somanuals.com. All Manuals Search And Download.  
 
DGSP message 148  
Disk Capacities 19  
Disk Failure 139  
heartbeats 11  
home directories 49  
Hot Standby Configuration 30  
hot standby configuration 30  
dual-network 36  
Dynamic Reconfiguration 169  
I
I/O Pacing 56, 146  
Importing 86  
E
editing  
/.rhosts file 59  
emsvcsd 156  
Enhanced Cluster Security 201  
Eprimary 196  
Inactive Takeover 29  
Installing HACMP 93  
IP Address Takeover 14  
Error Notification 45, 123  
Ethernet 13  
Event Customization 44, 117  
Event Emulator 123  
Event Manager 200  
Event Notification 44  
J
journaled file system log (jfslog) 85  
K
Kerberos 187  
L
F
Lazy Update 164  
Licensing Methods 43  
Log Files 143, 153  
Failed Components 160  
Failure Detection Rate 147  
fast recovery 110  
FCS 13  
FDDI 13  
M
fiber channel 13  
forced stop 158  
Forcing a Varyon 90  
major number 125  
Manual Update 163  
Monitoring the Cluster 151  
mounting  
NFS 126  
Mutual Takeover Configuration 32  
mutual takeover configuration 32  
G
graceful stop 158  
graceful with takeover stop 158  
Group Services 200  
grpglsmd 156  
N
grpsvcsd 156  
name serving  
cron considerations 58  
network adapter 38  
Network Adapter Events 121  
network adapters  
H
HACMP Daemons 155  
HACMP/ES  
adapter label 38  
starting on clients 159  
stopping on clients 159  
HACWS 183  
HANFS for AIX 201  
Hardware Address Swapping 12  
hardware address swapping 40  
planning 40  
defined 38  
Network Attribute 37  
Network Events 120  
Network Failure 138  
Network Modules 105  
Network Name 37  
Network Option Settings 57  
HAView 151  
218  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Network Topology 35  
networks  
point-to-point 36  
NFS  
S
SCSI  
target mode 38  
SCSI Disks 26  
mounting filesystems 126  
takeover issues 126  
NFS cross mount 41  
NFS Exports 41  
service adapter 38  
service ticket 188  
Shared LVM Component Configuration 81  
Shared LVs and Filesystems 84  
Shared VGs 82  
NFS mount 41  
NIM 199  
NIS 58  
single point of failure  
SLIP 13  
Node Events 117  
Node Failure / Reintegration 137  
Node isolation 147  
node relationships 108  
non-concurrent access  
quorum 90  
SOCC 13  
software maintenance 174  
SP Switch 14, 195  
splitlvcopy command 177  
Split-Mirror Backups 176  
SSA  
Non-Sticky Resource Migration 170  
advantages 25  
SSA Adapters 20  
SSA disk subsystem  
configuring adapter router 67  
standby adapter 9, 39  
starting  
cluster services on clients 159  
Sticky Resource Migration 170  
stopping  
cluster service on clients 159  
Switch Failures 196  
Symmetric Multi-Processor (SMP)  
syncd 146  
P
partitioned cluster 147  
password 49  
point-to-point connection 36  
principal 188  
private network 38  
PTFs) 174  
public network 37  
R
RAID on SSA Disks 72  
RAID on the 7133 Disk Subsystem 24  
RAID vs. Non-RAID 22  
Recoverable Virtual Shared Disk 193  
resource groups 108  
Resource Planning 28  
RISC System Cluster Technology 200  
Rootvg Mirroring 51  
synchronizing  
resource configuration 111  
Synchronizing Cluster Resources 111  
Synchronizing the Cluster Definition 106  
T
Takeover and Reintegration 112  
target mode SCSI 38  
Target-mode SCSI 14, 16  
Target-mode SSA 14, 16  
Task Guide 90  
Rotating Resource Group 29  
Rotating Standby Configuration 31  
rotating standby configuration 31  
RS232 15  
RSCT 200  
Rules  
TaskGuide 167  
TCP/IP Network Types 11  
TCP/IP Subsystem Failure 138  
Third-Party Takeover Configuration 33  
third-party takeover configuration 33  
Ticket 188  
IBM 7190 21  
Rules for SSA Loops 20  
Run-Time Parameters 110  
RVSD 193  
ticket-granting ticket 188  
219  
Download from Www.Somanuals.com. All Manuals Search And Download.  
Token-Ring 13  
Topology Service 200  
topsvcsd 156  
U
Upgrading 96  
user accounts  
adding 179  
changing 180  
creating 179  
removing 180  
User and Group IDs 48  
V
VGDA 88  
VGSA 88  
Virtual Shared Disk (VSDs) 190  
X
xhacmpm 101  
220  
IBM Certification Study Guide AIX HACMP  
Download from Www.Somanuals.com. All Manuals Search And Download.  
ITSO Redbook Evaluation  
IBM Certification Study Guide AIX HACMP  
SG24-5131-00  
Your feedback is very important to help us maintain the quality of ITSO redbooks. Please complete  
this questionnaire and return it using one of the following methods:  
• Use the online evaluation form found at http://www.redbooks.ibm.com  
• Fax this form to: USA International Access Code + 1 914 432 8264  
• Send your comments in an Internet note to [email protected]  
Which of the following best describes you?  
_ Customer _ Business Partner  
_ None of the above  
_ Solution Developer  
_ IBM employee  
Please rate your overall satisfaction with this book using the scale:  
(1 = very good, 2 = good, 3 = average, 4 = poor, 5 = very poor)  
Overall Satisfaction  
__________  
Please answer the following questions:  
Was this redbook published in time for your needs?  
If no, please explain:  
Yes___ No___  
What other redbooks would you like to see published?  
Comments/Suggestions:  
(THANK YOU FOR YOUR FEEDBACK!)  
© Copyright IBM Corp. 1999  
221  
Download from Www.Somanuals.com. All Manuals Search And Download.  
   
IBM Certification Study Guide AIX HACMP  
SG24-5131-00  
Download from Www.Somanuals.com. All Manuals Search And Download.  

Grizzly Lathe H8103 User Manual
Harbor Freight Tools Automobile Parts 66632 User Manual
Harbor Freight Tools Drill 47158 User Manual
Hearth and Home Technologies Indoor Fireplace ST HVB CE User Manual
Heatcraft Refrigeration Products Humidifier 25005601 User Manual
Honeywell Video Gaming Accessories W7750A User Manual
Hotpoint Freezer HSZ3022NF User Manual
HP Hewlett Packard Air Compressor 716 User Manual
HP Hewlett Packard Network Router StorageWorks 1 8 Tape Autoloader User Manual
Humminbird Fish Finder 531374 1_A User Manual