>>> Andreas Kurz <[email protected]> schrieb am 09.11.2011 um 15:48 in >>> Nachricht <[email protected]>: > On 11/09/2011 01:22 PM, Ulrich Windl wrote: > > Hi! > > > > I have a question : If I have a resource group with pacemaker > (pacemaker-1.1.5-5.9.11.1 on SLES11 SP1) that has several resources, and I > set one resource to "unmanaged", the group should not be affected, right? > > I also have a colocation like > > colocation col_ip2 inf: prm_ip2 cln_foo:Slave > > > > When I set "prm_ip2" to unmanaged the clone (ms) resource cln_foo did a > promote (slave to master on node where prm_ip2 is, and a demote master to > slave on another node) action. > > The idea of the colocation was to have the IP address on the node where the > slave instance of cln_foo is running. I hope the setup was fine. > > > > Misconfiguration or software bug? > > Try using the whole group and not only the ip in the colocation > constraint ... if the cluster sees no need to bind a resource to a > specific node it feels free to relocate resources if scores ... or in > this case promotion scores are the same. > > Though I am not sure if this behavior is intended for unmanaged > resources in the case you described ... might be a feature ;-) > > Generally I suggest to use maintenance mode so you can be sure cluster > does really nothing with any resource while you are doing your admin tasks.
Good point: Can I activate "maintenance mode" with the crm shell? AFAIK, "Maintenance mode" as done by the gui just puts all resources in to "unmanaged" mode. Did I miss something? Reagards, Ulrich _______________________________________________ Linux-HA mailing list [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
