On Tue, 2017-10-10 at 12:06 +0100, lejeczek wrote: > > On 26/09/17 13:15, Klaus Wenninger wrote: > > On 09/26/2017 02:06 PM, lejeczek wrote: > > > hi fellas > > > > > > can something like in the subject pacemaker do? And if yes then > > > how to > > > do it? > > > > You could bind ResourceA to nodeA and resource[xy] to node[xy] via > > location constraints. > > Afterwards you could make resource[xy] depend on ResourceA - > > without > > collocation. > > The actual commands (based on crmsh or pcs) to create these rules > > depend on thedistribution you are using. > > > > Regards, > > Klaus > > thanks, > I am probably hoping for too much(?) - without man made > constraints but sort of cluster logic would make these > decisions: wherever it decided to start/run resourceA then > rourceB would have to run on different(all or remaining > cluster's nodes). > > I'd only have to tell it something like: if you started > resourceA on nodeA then remaing(or maybe specific) resources > start on all but nodeA nodes.
Sure, that's simply a colocation constraint with a negative score. For details, see http://clusterlabs.org/doc/en-US/Pacemaker/1.1-pcs/htm l-single/Pacemaker_Explained/index.html#s-resource-colocation (and/or the help for whatever higher-level tools you're using) > > > > I'm looking into docs but before I'd gone through it all I hoped > > > an > > > expert could tell. > > > > > > many thanks, L. -- Ken Gaillot <kgail...@redhat.com> _______________________________________________ Users mailing list: Users@clusterlabs.org http://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