[Pacemaker] Dose Pacemaker-1.1 support suicide STONITH plugin?

Dejan Muhamedagic dejanmm at fastmail.fm
Thu Apr 25 14:11:51 UTC 2013


On Thu, Apr 25, 2013 at 10:48:38PM +0900, Takatoshi MATSUO wrote:
> Hi Dejan
> 
> 2013/4/25 Dejan Muhamedagic <dejanmm at fastmail.fm>:
> > Hi,
> >
> > On Thu, Apr 25, 2013 at 04:16:09PM +0900, Takatoshi MATSUO wrote:
> >> Hi
> >>
> >> I am trying to use suicide STONITH plugin with Pacemaker-1.1.10-rc1(138556c),
> >> but it doesn't work well.
> >> Does Pacemaker-1.1 support suicide plugin ?
> >
> > The suicide plugin may be useful only in single node clusters.
> 
> Yes, I know.
> 
> > Because it's not possible to get information about the outcome
> > of the fencing operation. Why would you want to use it?
> 
> I think it's useful for two nodes clusters which have no STONITH device too.
> It can fence when failure of stop occurs.

OK. Though I'd advise to use some real fencing device.

> It seems that Pacemaker-1.0.12 + suicide plugin work well, but

Yes, IIRC, there was an exception for the suicide plugin to run
on the node it can fence (normally, it's not allowed).

> Pacemaker-1.1 doesn't.
> Does Pacemaker-1.1 have no compatibility?

I wouldn't know, sorry.

Thanks,

Dejan

> Thanks,
> Takatoshi MATSUO
> 
> > Thanks,
> >
> > Dejan
> >
> >> I use cluster-glue-1.0.11.
> >>
> >>
> >> My configuration
> >> ---
> >> property \
> >>     no-quorum-policy="ignore" \
> >>     stonith-enabled="true" \
> >>     startup-fencing="false"
> >>
> >> rsc_defaults \
> >>     resource-stickiness="INFINITY" \
> >>     migration-threshold="1"
> >>
> >> primitive Stonith1 stonith:suicide
> >> primitive Stonith2 stonith:suicide
> >>
> >> fencing_topology \
> >>     16-sl6: Stonith1 \
> >>     17-sl6: Stonith2
> >>
> >> primitive dummy ocf:heartbeat:Dummy \
> >>     op start   timeout="10s" interval="0s"  on-fail="restart" \
> >>     op monitor timeout="10s" interval="10s" on-fail="restart" \
> >>     op stop    timeout="10s" interval="0s"  on-fail="fence"
> >>
> >> location rsc_location-Stonith1 Stonith1 \
> >>     rule -INFINITY: #uname eq 16-sl6
> >> location rsc_location-Stonith2 Stonith2 \
> >>     rule -INFINITY: #uname eq 17-sl6
> >> ---
> >>
> >> Logs of DC
> >> ---
> >> 16-sl6 crmd[3143]:   notice: te_fence_node: Executing reboot fencing
> >> operation (9) on 16-sl6 (timeout=60000)
> >> 16-sl6 stonith-ng[3139]:   notice: handle_request: Client
> >> crmd.3143.0f97946a wants to fence (reboot) '16-sl6' with device
> >> '(any)'
> >> 16-sl6 stonith-ng[3139]:   notice: handle_request: Forwarding complex
> >> self fencing request to peer 17-sl6
> >> 16-sl6 stonith-ng[3139]:   notice: remote_op_done: Operation reboot of
> >> 16-sl6 by 17-sl6 for crmd.3143 at 17-sl6.ef50f979: No route to host
> >> ----
> >>
> >>
> >> Regards,
> >> Takatoshi MATSUO
> >>
> >> _______________________________________________
> >> Pacemaker mailing list: Pacemaker at oss.clusterlabs.org
> >> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
> >>
> >> Project Home: http://www.clusterlabs.org
> >> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> >> Bugs: http://bugs.clusterlabs.org
> >
> > _______________________________________________
> > Pacemaker mailing list: Pacemaker at oss.clusterlabs.org
> > http://oss.clusterlabs.org/mailman/listinfo/pacemaker
> >
> > Project Home: http://www.clusterlabs.org
> > Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> > Bugs: http://bugs.clusterlabs.org
> 
> _______________________________________________
> Pacemaker mailing list: Pacemaker at oss.clusterlabs.org
> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
> 
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs: http://bugs.clusterlabs.org




More information about the Pacemaker mailing list