[Pacemaker] Colocation constraint to External Managed Resource

Robert H. pacemaker at elconas.de
Wed Oct 16 11:32:52 UTC 2013


> Right.
> Failure timeout is only evaluated on the next pengine run,
> so if nothing else happens, it takes up to recheck-interval ...

Good to know.

>> had to set failure-timeout to the same value (failure-timeout="5m").
>
> You should set it smaller than the recheck interval, I think.
> Or you may need to wait *two* recheck intervals
> for the failure timeout to take effect.
>
> Did you try with failure-timeout < recheck-interval?
> Problems?

I did test with failure-timeout < recheck-interval, then for some time 
the ressource is shown as "failed", while it is already online again.


Regards,
Robert

-- 
--
Robert




More information about the Pacemaker mailing list