[Linux-HA] default_resource_stickiness = INFINITY and resource monitoring

[Andrew Beekhof]

On 11/7/06, Paulo F. Andrade <pfca at mega.ist.utl.pt> wrote:
> Well it doesn't seem to be the case.
> I'm using Heartbeat v2.0.7, i have tested with
> default_resource_stickiness = 100, and the resource failed-over at
> count = 7. Meaning I had to kill my resource 7 times for it to fail-
> over.
>
> Then I reset the fail_count to 0, and changed the resource_stickiness
> to INFINITY in the GUI.
>
> I'm currently with fail_count = 9 and it still hasn't failed-over. It
> always restarts in the current node.
>
> What am I missing? When is a resource considered unhealthy?
>
> Paulo F. Andrade 52439 at IST
> mailto: pfca at mega.ist.utl.pt
>
>
> On 2006/11/03, at 23:57, Lars Marowsky-Bree wrote:
>
> > On 2006-11-03T15:30:57, "Paulo F. Andrade" <pfca at mega.ist.utl.pt>
> > wrote:
> >
> >> Will resource monitoring have any effect when
> >> default_resource_stickiness is set to INFINITY?
> >
> > Yes.
> >
> >> Will the resource failover if, for example the monitor operation on a
> >> daemon fails?
> >
> > Yes.
> >
> > stickiness only applies if the resource is healthy.

No it doesn't.  Remember how you asked it to always apply because
moving it may also involve moving other (healthy) resources?

> >> The way I figure it, is that it's pretty useless since the score will
> >> be reduced (by whats set in default_resource_failure_stickiness) but
> >> it won't have any effect (at least for some _long_ time), since the
> >> initial score is INIFINITY.
> >>
> >> Is this correct?
> >
> > No.
> >
> > _______________________________________________
> > Linux-HA mailing list
> > Linux-HA at lists.linux-ha.org
> > http://lists.linux-ha.org/mailman/listinfo/linux-ha
> > See also: http://linux-ha.org/ReportingProblems
>
> _______________________________________________
> Linux-HA mailing list
> Linux-HA at lists.linux-ha.org
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
> See also: http://linux-ha.org/ReportingProblems
>