[Pacemaker] stonith question

Dejan Muhamedagic dejanmm at fastmail.fm
Wed May 25 05:21:33 EDT 2011


Hi,

On Tue, May 24, 2011 at 01:40:49PM -0500, Phil Armstrong wrote:
> Hi,
> 
> I have a 2 node cluster with stonith disabled:
> 
>        <nvpair id="cib-bootstrap-options-stonith-enabled"
> name="stonith-enabled" value="false"/>
> 
> This is on an x86_64 liux system. I was having trouble getting a
> system dump. I would drop into kdb, then enter kdump. The dump
> process would appear to initiate, then it would hang, and then the
> machine would reboot itself as if it had been stonith'd. This was
> when I had both nodes active in the HA cluster.
> 
> I stopped openais on both nodes and I was able to get all the darn
> system dumps I wanted. But before that (when openais was on) I could
> find no evidence in the logs of the node that remained up that it
> had decided to shoot the node doing the kdump, and of course once I
> dropped into kdb there was no evidence of anything in that node's
> log.
> 
> My question is is there ever any situation with stonith disabled
> that a node would shoot either itself or another node in the
> cluster.

I don't think so.

You may also want to take a look at external/kdumpcheck. That
should allow you to use fencing and get the kdump.

Thanks,

Dejan

> I can get an hb_report if someone thinks that would be
> helpful, but for now I'm just posing the above question.
> 
> Thanks.
> 
> Phil
> 
> 
> -- 
> 	Phil Armstrong       pma at sgi.com
> 	Phone: 651-683-5561  VNET 233-5561
> 
> 
> _______________________________________________
> 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://developerbugs.linux-foundation.org/enter_bug.cgi?product=Pacemaker




More information about the Pacemaker mailing list