man crm_failcount

fail count will not be reset automatically, has to be done manually. As
well, crm_verify and (if any "failed start" or "failed op" appear)
crm_resource -C

It is not possible for the cluster to "forget" about failure counts, as
otherwise a failing resource could be forever bouncing around the
cluster and causing transitions (which depending on configuration could
be affecting other resources).

Not too sure about the details myself... but resetting the failure
counts is necessary to allow the resource to run again, and clearing the
resources seems to restore sanity in my experience.

Can somebody explain what the effect of "failed operations" or "failed
to start" messages in crm_verify actually is please?

In any case, your biggest problem is that the resource fails in the
first place. That's what needs fixing!

Yan


Steinhauer Juergen wrote:
> Hi all!
> 
> After several hours of reading the documentation and trying different
> setups I'm asking for help....
> 
> A two node cluster,
> default-resource-failure-stickiness = -100,
> default-resource-stickiness = 100,
> one resource group with two resources,
> both nodes have a preference of 300.
> 
> Node A is active, after 3 failures of the same resource node B becomes
> active.
> Why? Because of (300 -300) + (100+100) /100 =2 (according to the FAQ)?
> Then node B fails 4 times and node A gets active again. Why 4 times?
> Now, node A fails another 3 times. Failure counter is now 6 and the
> service stops on all nodes.
> 
> The other way round:
> Node B is active, after 2 failures node A gets active. Then 5 failures
> on A, B takes over again. Now B fails 3 times (counter is now 5) and the
> service stops.
> 
> Can anybody explain me the details?
> 
> Thanks in advance.
> 
> Regards.
> 
> 
> 
> 
> 
> _______________________________________________
> Linux-HA mailing list
> Linux-HA@lists.linux-ha.org
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
> See also: http://linux-ha.org/ReportingProblems
_______________________________________________
Linux-HA mailing list
Linux-HA@lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Reply via email to