0

Dear ServerFault users i've got a question here.

My setup is following: Two node cluster, one service group which consists of 4 tiers of resources. Group is activated and probed on both nodes and basically it failovers normally.
But i need one specific thing which is as follows: when any resource of upper level ( parent ) fails 3 times, failover happens and all the service group is being switched to second node.
For that, i've set some parameters and marked all those resources as 'Critical'.
But at the same time i need the child resource for those above, which they depend on, to be bringing down whole group and thus causing package switch on first fault. on way is to set RestartLimit to 1 and mark it as critical. But these properties are common for all resources on 'Application' or 'Service Group' level, so no possibility to 'personalize' resources behavior.

Thanks in advance!

hmallett
  • 2,425
  • 14
  • 26
Nikolay Uloga
  • 3
  • 1

1 Answers1

0

When any critical resource in a service group faults the service group should fail over. Normally you'd have all resources in a service group marked as critical.

Am I right in thinking that when an agent probes a resource and discovers it offline when it should be online, you wish to attempt to restart the resource twice before marking it as faulted? If so, then yes you can modify the resource's properties to do that.

If you wish to modify individual resources' properties, then on newer versions of VCS (5.0 and later I think) you can simply modify the properties. On previous versions you'd need to copy the resource type to a new resource, and modify the copy.

hmallett
  • 2,425
  • 14
  • 26