[ClusterLabs] Antwort: Re: clear pending fence operation

[philipp.achmueller at arz.at]

Andrei Borzenkov <arvidjaar at gmail.com> schrieb am 07.07.2015 10:03:26:

> Von: Andrei Borzenkov <arvidjaar at gmail.com>
> An: Cluster Labs - All topics related to open-source clustering 
> welcomed <users at clusterlabs.org>
> Datum: 07.07.2015 10:04
> Betreff: Re: [ClusterLabs] clear pending fence operation
> 
> On Tue, Jul 7, 2015 at 10:41 AM,  <philipp.achmueller at arz.at> wrote:
> > hi,
> >
> > is there any way to clear/remove pending stonith operation on cluster 
node?
> >
> > after some internal testing i got following status:
> >
> > Jul  4 12:18:02 XXX crmd[1673]:   notice: te_fence_node: Executing 
reboot
> > fencing operation (179) on XXX (timeout=60000)
> > Jul  4 12:18:02 XXX stonith-ng[1668]:   notice: handle_request: Client
> > crmd.1673.1867d504 wants to fence (reboot) 'XXX' with device '(any)'
> > Jul  4 12:18:02 XXX stonith-ng[1668]:   notice: 
initiate_remote_stonith_op:
> > Initiating remote operation reboot for XXX:
> > 3453b93d-a13a-4513-b05b-b79ad85ff992 (0)
> > Jul  4 12:18:03 XXX stonith-ng[1668]:    error: remote_op_done: 
Operation
> > reboot of XXX by <no-one> for crmd.1673 at XXX.3453b93d: Generic 
Pacemaker
> > error
> > Jul  4 12:18:03 XXX crmd[1673]:   notice: tengine_stonith_callback: 
Stonith
> > operation 2/179:23875:0:134436dd-4df8-44a2-bf4a-ec6276883edd: Generic
> > Pacemaker error (-201)
> > Jul  4 12:18:03 XXX crmd[1673]:   notice: tengine_stonith_callback: 
Stonith
> > operation 2 forXXX failed (Generic Pacemaker error): aborting 
transition.
> > Jul  4 12:18:03 XXX crmd[1673]:   notice: abort_transition_graph: 
Transition
> > aborted: Stonith failed (source=tengine_stonith_callback:697, 0)
> > Jul  4 12:18:03 XXX crmd[1673]:   notice: tengine_stonith_notify: Peer 
XXX
> > was not terminated (reboot) by <anyone> for XXX: Generic Pacemaker 
error
> > (ref=3453b93d-a13a-4513-b05b-b79ad85ff992) by client crmd.1673
> >
> > so, node XXX is still online, i want to get cluster back to stable
> >
> 
> If you are using sufficiently recent pacemaker, you can use
> "stonith_admin --confirm"; be sure to actually stop all resources on
> victim node in this case.

ok, i use v1.1.12. 
so in my case there were 7 VirtualDomain RGs online, do i have to migrate 
them to there scheduled destination location, or is this not relevant?

warning: stage6: Scheduling Node XXX for STONITH
notice: LogActions: Move    vma#011(Started XXX -> WWW)
notice: LogActions: Move    vmb#011(Started XXX -> WWW)
notice: LogActions: Move    vmc#011(Started XXX -> YYY)
notice: LogActions: Move    vmd#011(Started XXX -> YYY)
notice: LogActions: Move    vme#011(Started XXX -> ZZZ)
notice: LogActions: Move    vmf#011(Started XXX -> ZZZ)
notice: LogActions: Move    fmg#011(Started XXX -> YYY)

thank you!


> 
> On older pacemaker using crmsh "crm node clearstate" does the same.
> 
> _______________________________________________
> Users mailing list: Users at clusterlabs.org
> http://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

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.clusterlabs.org/pipermail/users/attachments/20150707/6b5cc4eb/attachment-0003.html>