[Linux-HA] crm_mon confused?

[Christian Rishøj]

On 10/31/07, Ian Turner <vectro at vectro.org> wrote:

> That is there so that even if crm_failcount=100 for a node, the
> resource_failure_stickiness will not bring the total score below zero. I want
> services to keep running if possible; if there is a better way, I'd love to
> hear it.

Also interested in this.

Additionally, it would be nice if I could configure Heartbeat to be
more eager in trying to recover failed resources. Suppose my database
times out when stopping. Now it's banned, with a failed stop action.
I'd like Heartbeat to check the resource again, and if it restarts
properly, then to clear the failed status. Any hints on accomplishing
this would be much appreciated.

Thanks

Christian