01-27-2012 09:02 AM
Is there anyway to configure a service group to fail over when another service group faults? I took a look at the VCS user guide and the group dependencies they have listed doesn't match my requirement.
Solved! Go to Solution.
01-27-2012 11:40 AM
Yes, you can use triggers.
Suppose you want SG_B to fail over when SG_A faults
You can use:
resfault: This will be triggered when resource fails so you could configure this to failover SG_B when a resource faults in SG_A
preonline: This will be triggered when a SG onlines, so you could configure this for SG_A so that when it onlines with reason "FAULT" (this is passed as an argument to preonline script) then the script fails over SG_B.
To configure trigger scripts, copy them from /opt/VRTSvcs/bin/sample_triggers to /opt/VRTSvcs/bin/triggers and modify scripts accordingly. Most triggers are automatically called if they are in the triggers directory, but the preonline trigger is an exception and this also requires you set PreOnline service group attribute to 1 for SGs that you want preonline to run for.
The sample triggers will explain what arguments are passed to the script and the VCS admin guide will explain when triggers are called.
SG dependencies are usually a better solution than triggersif SG dependencies could be used so you should check that "offline" dependency cannot be used.
Mike
01-27-2012 11:40 AM
Yes, you can use triggers.
Suppose you want SG_B to fail over when SG_A faults
You can use:
resfault: This will be triggered when resource fails so you could configure this to failover SG_B when a resource faults in SG_A
preonline: This will be triggered when a SG onlines, so you could configure this for SG_A so that when it onlines with reason "FAULT" (this is passed as an argument to preonline script) then the script fails over SG_B.
To configure trigger scripts, copy them from /opt/VRTSvcs/bin/sample_triggers to /opt/VRTSvcs/bin/triggers and modify scripts accordingly. Most triggers are automatically called if they are in the triggers directory, but the preonline trigger is an exception and this also requires you set PreOnline service group attribute to 1 for SGs that you want preonline to run for.
The sample triggers will explain what arguments are passed to the script and the VCS admin guide will explain when triggers are called.
SG dependencies are usually a better solution than triggersif SG dependencies could be used so you should check that "offline" dependency cannot be used.
Mike
01-28-2012 08:22 AM
I am wondering on why you want to failover SG B when there is a fault on SG A.
Can you please elaborate more on your setup and what exactly you want ?
A example of hastatus -sum will be useful.
02-13-2012 05:51 AM
Mike - Thanks!
Anoop - Its a rather complicated solution that involves another product