cancel
Showing results for 
Search instead for 
Did you mean: 

Restart clustered application without triggereing failover?

shaggy62
Level 3
Hello,
I have a number of apps setup in simple failover cluster configurations across multiple clusters. Most of these are java application servers. Normally if an app needs to be restarted I can simply switch the app from one node to another (and back if need be) and that is fine. However there are times, particularly in non-production clusters, where application developers need to restart the application service and aren't interfacing with the cluster to do so. I have done this by disabling the service group on the secondary node(s) and then performing the restart, but I'd like to be able to restart the applications and not trigger a failover operation at all. So the question is how do I do that? Is there a way I can tell the cluster I'm performing an opertion like that?
1 ACCEPTED SOLUTION

Accepted Solutions

AHerr
Level 5
Employee Accredited Certified
Instead of disabling the service group you could just freeze the group.  This has the effect of preventing a failover if a resource comes offline outside of the cluster.
When you are finished with your external operations, then you should run the probe command to ensure that the resource is back online and then unfreeze the service group.
The one thing to note in this scenario.  If the physical machine was to fail while the SG is frozen, that group will not automatically be brought online on a failover node.
Because it is frozen it will not take any action on the group (online/offline/restart of resources).

View solution in original post

3 REPLIES 3

AHerr
Level 5
Employee Accredited Certified
Instead of disabling the service group you could just freeze the group.  This has the effect of preventing a failover if a resource comes offline outside of the cluster.
When you are finished with your external operations, then you should run the probe command to ensure that the resource is back online and then unfreeze the service group.
The one thing to note in this scenario.  If the physical machine was to fail while the SG is frozen, that group will not automatically be brought online on a failover node.
Because it is frozen it will not take any action on the group (online/offline/restart of resources).

shaggy62
Level 3
That sounds like exactly what I need. I'll write a wrapper script that freezes the SG, restarts the application, probes the SG and then unfreezes the SG. Thanks for the explanation.


Marianne
Level 6
Partner    VIP    Accredited Certified

Why not  simply train the admin guys to use the cluster GUI or HA commands to offline/online resources?

Another option would be to disable automatic failover or configure automatic restart when a resource goes offline.
The best is to read the entire chapter in the VCS users guide called Controlling VCS behavior and carefully consider all possible consequences. Find the VCS Admin/User Guide for your O/S and VCS version here:
https://vos.symantec.com/documents