[ClusterLabs] corosync doesn't start any resource

Stefan Krueger Shadow_7 at gmx.net
Mon Jun 25 08:35:11 EDT 2018


Hello Andrei,

> Then you need to set symmetrical="false".
yep, now it seems to work now, thank you very much!

> I assume this would be "pcs constraint order set ...
> symmetrical=false".
yes almost:
pcs constraint order set nfs-server vm_storage ha-ip action=start setoptions symmetrical=false


Thank you very very much!

best regards
Stefan

> Gesendet: Samstag, 23. Juni 2018 um 22:13 Uhr
> Von: "Andrei Borzenkov" <arvidjaar at gmail.com>
> An: users at clusterlabs.org
> Betreff: Re: [ClusterLabs] corosync doesn't start any resource
>
> 22.06.2018 11:22, Stefan Krueger пишет:
> > Hello Andrei,
> > 
> > thanks for this hint, but I need this "special" order. In an other setup it works.
> > 
> 
> Then you need to set symmetrical="false". Otherwise pacemaker implicitly
> creates reverse order which leads to deadlock. I am not intimately
> familiar with pcs, I assume this would be "pcs constraint order set ...
> symmetrical=false".
> 
> > best regards
> > Stefan
> > 
> >> Gesendet: Freitag, 22. Juni 2018 um 06:57 Uhr
> >> Von: "Andrei Borzenkov" <arvidjaar at gmail.com>
> >> An: users at clusterlabs.org
> >> Betreff: Re: [ClusterLabs] corosync doesn't start any resource
> >>
> >> 21.06.2018 16:04, Stefan Krueger пишет:
> >>> Hi Ken,
> >>>
> >>>> Can you attach the pe-input file listed just above here?
> >>> done ;) 
> >>>
> >>> And thank you for your patience!
> >>>
> >>
> >> You delete all context which makes it hard to answer. This is not web
> >> forum where users can simply scroll up to see previous reply.
> >>
> >> Both your logs and pe-input show that nfs-server and vm-storage wait for
> >> each other.
> >>
> >> My best guess is that you have incorrect ordering for start and stop
> >> which causes loop in pacemaker decision. Your start order is "nfs-server
> >> vm-storage" and your stop order is "nfs-server vm-storage", while it
> >> should normally be symmetrical. Reversing order in one of sets makes it
> >> work as intended (verified).
> >>
> >> I would actually expect that asymmetrical configuration still should
> >> work, so I leave it to pacemaker developers to comment whether this is a
> >> bug or feature :)
> >>
> >> _______________________________________________
> >> 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
> >>
> > _______________________________________________
> > 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
> > 
> 
> _______________________________________________
> 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
>



More information about the Users mailing list