[ClusterLabs] Antw: Re: Antw: clearing failed actions

[Attila Megyeri]

Thanks.

We have several clusters working fine since several years, without STONITH, so we did not really bother to implement it.

As for the failed actions - I cannot recall since when these are not cleared, but they aren't.

When I check the pengine log on the DC, at every recheck interval I see lines like:

...    pengine:  warning: unpack_rsc_op:    Processing failed op start for jboss_admin1 on ctadmin1: unknown error (1)

or 
...    pengine:  warning: unpack_rsc_op:    Processing failed op monitor for jboss_abssrv2 on ctabs2: unknown error (1)


And these are the failed actions  visible in the crm_mon -f as well:


Failed actions:
    jboss_admin1_start_0 (node=ctadmin1, call=120, rc=1, status=Timed Out, last-rc-change=Thu Jun  1 14:17:31 2017
, queued=40001ms, exec=0ms
): unknown error
    jboss_abssrv2_monitor_10000 (node=ctabs2, call=106, rc=1, status=complete, last-rc-change=Thu Jun  1 14:13:36 2017
, queued=0ms, exec=0ms
): unknown error


If I do a resource cleanup, the errors are gone.

At the same time I see no actions on the mentioned nodes - this log is from the DC...

On the mentioned notes a regular monitoring operation is performed, and reults in 0 - no error.

What am I missing here?



> -----Original Message-----
> From: Ulrich Windl [mailto:Ulrich.Windl at rz.uni-regensburg.de]
> Sent: Thursday, June 1, 2017 8:34 AM
> To: users at clusterlabs.org
> Subject: [ClusterLabs] Antw: Re: Antw: clearing failed actions
> 
> >>> Digimer <lists at alteeve.ca> schrieb am 01.06.2017 um 00:03 in Nachricht
> <50aad2be-185b-0348-6a93-987034c9cb3a at alteeve.ca>:
> [...]
> > I don't know, but according to Ken's last email, what you're seeing is
> > expected. I replied because of the miss understanding of the rolls
> > quorum and fencing play. Running a cluster without fencing is dangerous.
> 
> I'd recommend this: Enable a working STONITH. Then if you see your cluster
> never uses STONITH, and everything works fine, and you feel you don't
> needit, then you can get rid of it.
> But don't try the other way 'round: Omit STONITH, expecting the cluster
> would work flawlessly, then (not) add STONITH.
> 
> Regards,
> Ulrich
> 
> 
> 
> 
> _______________________________________________
> Users mailing list: Users at clusterlabs.org
> http://lists.clusterlabs.org/mailman/listinfo/users
> 
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs: http://bugs.clusterlabs.org