Dominik Klein wrote: > So resource "cups" runs on node "apollo" by default.
Yes apollo is the production (sort of) node and artemis where I test patches, etc hence prefer running everything on 1 node and have the other as a failover. > If you reboot apollo and apollo comes back into cluster, you have > +INFINITY for cups on apollo and on the node cups currently runs on. So > then it's heartbeats decision where to run the resource and it looks > like it "thought" apollo would be the better location. Yes. According to my understanding it meant to be an INFINITY<>INFINITY situation where I expected no action to be taken. AFAIK only -INFINITY wins over INFINITY. Anyway I am not planning to file a bug (unless you guys convince me) yet, it needs further testing I suppose. It didn't happen in the past when I had score=100 only for VMs which was more than enough to start the resource on the preferred node initially. I consider moving back to a smaller score at this stage. Thanks for all, Ivan _______________________________________________ Linux-HA mailing list Linux-HA@lists.linux-ha.org http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems