On Wed, Dec 16, 2009 at 11:21 AM, Andrew Beekhof <and...@beekhof.net> wrote: > On Tue, Dec 15, 2009 at 5:37 PM, Rasto Levrinc <rasto.levr...@linbit.com> > wrote: >> Hi Beekhof, >> >> I've found again a Pacemaker bug or I am doing again something wrong. :) >> >> So I have a group with two dummy resources that are both unmanaged but >> running. The group has colocation constraint with score -INFINITY with the >> third dummy resource. The group resources are running on one node and the >> third dummy resource on the other. >> >> So far so good. >> >> Now I set the first resource in the group to be managed by pacemaker and >> all managed resources decide to move to the opposite nodes. This I would >> not expect. Resource stickiness is set. If I reverse the colocation >> constraint (exchange "rsc" and "with-rsc") it works ok. > > is-managed doesn't cause the move, it just allows pacemaker to perform it. > the real issue is the constraint. > > Seems like you might have found a bug in the way the scores are merged > for groups. >
Fixed: http://hg.clusterlabs.org/pacemaker/stable-1.0/rev/bff0ba217110 _______________________________________________ Pacemaker mailing list Pacemaker@oss.clusterlabs.org http://oss.clusterlabs.org/mailman/listinfo/pacemaker