[Pacemaker] monitor interval too short

[Dejan Muhamedagic]

Hi,

On Sat, Sep 29, 2012 at 02:10:26PM +0000, James Harper wrote:
> I have a xen resource with op's configured like:
> 
>         op stop interval="0" timeout="300s" \
>         op migrate_from interval="0" timeout="300s" \
>         op migrate_to interval="0" timeout="300s" \
>         op monitor interval="10s" timeout="90s" \
> 
> but it seems like every time I make any change to the crm configuration, I get a heap of timeouts, even though I can't see anything anywhere that is introducing delays into the system (all the xm commands I test while this is happening return more-or-less immediately, and certainly within a second or so)

"Any change" is what? A change to this particular resource? Which
operations timeout?

> So I put some logging into the Xen resource agent, and while it is mostly ticking along being called every 10 seconds like it should, suddenly it starts getting called every second. Is there an explanation for why it does this? Calling xm for every vm every 10 seconds is fine, but calling it every 1 second for every vm is going to introduce delays. It's never going to get to 90 seconds, but maybe the timeout is getting reduced by the same amount?

The timeout is never reduced. The next monitor operation is
scheduled once the current one has finished. I've never heard of
that the monitor interval gets reduced by itself. Are you really
sure that that happens? If so, please open a bugzilla and attach
a hb_report.

Thanks,

Dejan

> Pacemaker is Debian packaged 1.1.7-1.
> 
> Thanks
> 
> James
> 
> 
> _______________________________________________
> 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