[ClusterLabs] Antw: Re: Q: Stickyness across resource restarts
Ulrich.Windl at rz.uni-regensburg.de
Tue Aug 21 05:38:54 UTC 2018
>>> Ken Gaillot <kgaillot at redhat.com> schrieb am 20.08.2018 um 17:11 in
<1534777874.6465.7.camel at redhat.com>:
> On Thu, 2018‑08‑16 at 12:50 +0200, Ulrich Windl wrote:
>> I have a question, specifically whether this behavior changed not too
>> long ago:
>> AFAIR, stickyness only affected running resources, i.e. during
>> restart a new (=lightly loaded) node was selected for the resource.
>> However when I restarted a resource today (SLES11 SP4), the resource
>> was started on a busy node, even one node had no resources at all
>> after being restarted earlier.
>> The docs are not very precise on that.
> It appears stickiness still applies if the stop and start are scheduled
> in the same transition (e.g. failure recovery). That makes sense
> because the actions are scheduled based on the state of the resource at
> the start of the transition (active). There haven't been any
> intentional changes in stickiness behavior in many years. I haven't
> tested the behavior with older versions to see if it's different
Thanks for that. Interestingly I restarted another resource that is configured
just like the other one, and that resource was placed to the empty node.
So if it works sometimes, and sometimes not, I suspect it's a bug ;-)
Actually I have an open support call for this, but still without results.
> Ken Gaillot <kgaillot at redhat.com>
> Users mailing list: Users at clusterlabs.org
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs: http://bugs.clusterlabs.org
More information about the Users