[ClusterLabs] shutdown and restart of complete cluster due to power outage with UPS

[Klaus Wenninger]

On 01/24/2019 02:15 PM, Lentes, Bernd wrote:
>
> ----- On Jan 22, 2019, at 6:35 PM, Andrei Borzenkov arvidjaar at gmail.com wrote:
>
>
>> This is another problem - if cluster requires stonith, it won't statr
>> resources with another node UNCLEAN and fencing attempt apparently failed.
> Let's assume a running two-node cluster. Now node1 needs to be fenced because of some failure.
> node2 fences node 1, but, although fencing was successful, for whatever reason, the cluster
> does not recognize the fencing and node1 stays UNCLEAN.
> Does that mean that node2 can't take the resources from node 1 and start them ?

Yes, that is the aim of fencing - to assure that a node where we can't
really
say if it is running certain resources or not is brought to a defined
state (off).
And of course it has to be verified that this fencing action has succeeded.
But usually you won't run into big trouble if a fencing action actually goes
OK as you described but somehow misses to convey that information to
the node that triggered the fencing.
Fencing is usually gonna be repeated till it succeeds.

Klaus
>
> Bernd
>  
>
> Helmholtz Zentrum Muenchen
> Deutsches Forschungszentrum fuer Gesundheit und Umwelt (GmbH)
> Ingolstaedter Landstr. 1
> 85764 Neuherberg
> www.helmholtz-muenchen.de
> Aufsichtsratsvorsitzende: MinDirig'in Petra Steiner-Hoffmann
> Stellv.Aufsichtsratsvorsitzender: MinDirig. Dr. Manfred Wolter
> Geschaeftsfuehrer: Prof. Dr. med. Dr. h.c. Matthias Tschoep, Heinrich Bassler
> Registergericht: Amtsgericht Muenchen HRB 6466
> USt-IdNr: DE 129521671
>
> _______________________________________________
> Users mailing list: Users at clusterlabs.org
> https://lists.clusterlabs.org/mailman/listinfo/users
>
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs: http://bugs.clusterlabs.org