On 2011-10-16 18:39, kevins7189 wrote: > > I cannot seem to get my 2 cluster setup to work they way I want it do, or > believe I think it should. > Issues I have so far are: > First, whenever one of the two cluster nodes goes off-line (forcing them to > do so during testing, but shutdown or stopping openais), then I restart > openais on the cluster that I rebooted or stopped openais on, the resources > restart on the same active node. This cannot happen. I was under the > impression the global variable "resource stickiness" was supposed to be keep > this from happening, I've tried both 0 and INFINITY. I must not be > understanding something correctly. But, its important to me to not have the > resources restart at all if they are already happily running.
0 is equivalent to not setting a stickiness at all, INFINITY means you're practically never allowing them to migrate. Both are usually a bad choice. You would normally use a positive numerical value that is _not_ INFINITY. What example prompted you to use this? > Second, if I stop openais on the Active node, the resources do not migrate. > I can see this using LCMC even before I try it. When I stop openais, I can > see in LCMC that the resources will not migrate, but if I try it anyway, > there is an option in the Resource Group to "Remove Migration Constraint". > What is the migration constraint it is removing, and how can I remove it > permanently to make resources migrate like they should? I repeat myself -- we really need an FAQ list here --, but giving your cluster configuration only is inconclusive. We also need to know your cluster status to help you figure out what's wrong. Create a CIB dump (which contains both) with "cibadmin -Q", upload that to pastebin or whatever your favorite service is, and share the URL here. > Using Sles11 SP1 HA. I'm sure SUSE support would be happy to help, no? Why not file a ticket with them? Cheers, Florian -- Need help with High Availability? http://www.hastexo.com/now _______________________________________________ Linux-HA mailing list [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
