IBM SG24-5131-00 User Manual

© Copyright IBM Corp. 1999

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.

The following sections describe important details of cluster node setup.

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

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