When I need to restart a service that is part of a cluster I
typically run:

  1. crm_attribute -t crm_config -n is_managed_default -v false
  2. restart the service
  3. crm_attribute -t crm_config -n is_managed_default -v true

This works well except the fail counter for the resource is still
incremented while is_managed_default=false. This often causes the
service to migrate to a different node when management is enabled.

Is there a way to avoid the fail counter being incremented?

Or is there a more preferred method to disable the cluster, while
doing maintenance?

thanks, Jesse

_______________________________________________
Pacemaker mailing list
[email protected]
http://list.clusterlabs.org/mailman/listinfo/pacemaker

Reply via email to