Sun Cluster fail-over Operation

[thomasr]

Scenario:
Cluster of 4 SUN 4500s , Informatica running on server1, server 1 dies and server2 detects the failure via no heartbeat response.
Server2 reaction:
Server2 now takes over the shared disk space. Do I need to tailor (write scripts) to restart the Informatica process on server2 or can I simply have Informatica running hot standby, check a common logfile within the shared disk space to determine where to pickup the processing? [sig][/sig]

 

[garyj4]

To write your own custom failover scripts:

You only need to write the start, stop, and abort scripts. But you must, I repeat, must have the second server start it up. It cannot be running and just take it over.

Take a look at the scripts in /opt/SUNWcluster/ha/nfs

These are the startup/stop scripts for nfs.

Good luck.

Gary Jarrell.

 

[jcode]

Just remember that sun does not support home grown data service scripts - not that you have to tell them when you call in - but just a point for future reference -jc