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

6.1Node Verification

Here is a series of suggested actions to test the state of a node before including HACMP in the testing.

6.1.1Device State

Run diag -ain order to clean up the VPD.

Look in the errorlog for unusual errors by issuing the command errpt more or 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#.tm on 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

Page 149
Image 149
IBM SG24-5131-00 manual Cluster Testing, Node Verification, Device State