[Pacemaker] Both nodes restart with external/ssh STONITH agent

[Dejan Muhamedagic]

Hi,

On Wed, Jan 05, 2011 at 09:47:36AM +0100, Simon Jansen wrote:
> Hi,
> 
> I configured the external/ssh STONITH agent on my testing cluster that is
> based on virtual machines. This agent is just for testing purposes and will
> be replaced by an hardwarebased one when the cluster goes into live
> operation.
> So to test the functionality I disconnected the cluster interconnect on one
> node. After that both nodes are restarted by the stonith process. This ends
> up in an endless loop of system restarts on both nodes till the cluster
> interconnect comes up again.
> 
> I configured the agent as follows:
> primitive resStonith stonith:external/ssh \
>         params hostlist="node1 node2"
> 
> I tried to clone the resource but without success. The system acts in the
> same way.
> 
> Is this the normal behaviour of the ssh agent or did I misconfigured the
> agent?

Yes, it's probably normal. external/ssh uses at(1) to queue the
reboot command (which actually includes a sleep beforehand), so
there is a short window in which both nodes may run at(1)
successfully.

Thanks,

Dejan

> -- 
> 
> 
> Regards,
> Simon Jansen
> 
> 
> ---------------------------
> Simon Jansen
> 64291 Darmstadt

> _______________________________________________
> 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