Service set with recovery="restart-disable" or "disable" will still relocate to another node if it fails during start or restart in a RHEL 6 cluster with rgmanager
Issue
- With a cluster service with
recovery="restart-disable"
, if a status check on a resource in that service fails, we'll attempt to restart, as expected. But if a start operation as part of that restart fails, then we'll still relocate to another node rather than disabling. This seems incorrect. -
The RHEL HA Guide states:
The restart-disable policy indicates that the system should attempt to restart the service in place if it fails, but if restarting the service fails the service will be disabled instead of being moved to another host in the cluster.
However the behavior when a restart fails doesn't seem to match this.
- If I use
recovery="disable"
, if there's a failure when starting the service, it still relocates to another node. Is there any way to not have it do this?
Environment
- Red Hat Enterprise Linux (RHEL) 6 with the High Availability Add On
rgmanager
- One or more services in
/etc/cluster/cluster.conf
withrecovery="restart"
orrecovery="restart-disable"
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.