--- Andrew Beekhof <[EMAIL PROTECTED]> schrieb:
> On Mon, May 19, 2008 at 11:52 AM, Jozef Slezacek
> <[EMAIL PROTECTED]> wrote:
> > Hi,
> >
> > I set the default_resource_stickiness to INFINITY
> so resources don't fail-back randomly.
> > One test scenario (on node ldirector2) in which I
> damage the resource configuration (ldirectord) so it
> can't start up, the node get's STONITHed as expected
> and the resources fail over.
> >
> > However, when it comes up again and heartbeat is
> started, the monitor operation tries to start on
> ldirector2
>
> no, it doesn't
>
> monitors do what the word monitor implies - they
> check the current
> status of the resource.
I was under the impression that monitor checks status
only on the active node where the resource is running.
> yours is failing at startup but using the general
> "something is wrong"
> error code instead of the "invalid configuration"
> error code.
Why is the resource starting up (and failing) in the
first place if stickiness is set to INFINITY ?
> in such cases the only safe thing to do is assume
> the resource is in
> some way active and recover it.
>
> >, fails again and get's STONITHed. After the
> stonith, the resource does not fail over anymore.
> >
> > How can I prevent the monitor to try to come up
> on a node after reboot, even though it's not
> supposed to run any resources ? (stickiness
> INFINITY)
> >
> > -- snip --
> > May 19 10:51:11 ldirector1 pengine: [3471]: ERROR:
> unpack_rsc_op: Remapping ldirector_monitor_0 (rc=1)
> on ldi
> > rector2 to an ERROR
> > May 19 10:51:11 ldirector1 pengine: [3471]: ERROR:
> native_add_running: Resource
> ocf::ldirectord:ldirector app
> > ears to be active on 2 nodes.
> > -- snip --
> >
> > The hb_report attached is between 10:50:52 and
> 10:52:03 where I started up the ldirector2 node
> until the point where it gets stonithed.
> >
> > Regards
> > Jo
> >
__________________________________________________________
Gesendet von Yahoo! Mail.
Dem pfiffigeren Posteingang.
http://de.overview.mail.yahoo.com
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
