[ClusterLabs] Resources stopped due to unmanage

[Pavel Levshin]

Hello.


I've just expiriensed a fault in my pacemaker-based cluster. Seriously, 
I'm completely disoriented after this. Hopefully someone can give me a 
hint...


Two-node cluster runs few VirtualDomains along with their common 
infrastructure (libvirtd, NFS and so on). It is Pacemaker 1.1.16 
currently. Resources have ordering and colocation constraints, which 
must ensure proper start and stop order. Unfortunately, these 
constraints have unwanted side effects. In particular, due to mandatory 
ordering constraints, the cluster tends to restart libvirtd when I need 
to stop a VM. This was on the list prevoiusly: 
https://lists.clusterlabs.org/pipermail/users/2016-October/004288.html


And today I've tried to perform maintenance task on one of VMs. I've typed:

pcs resource unmanage vm3

and all other VMs have been suddenly stopped. Seriously?!!!

Logs show that the cluster performed internal custom_action "vm3_stop 
(unmanaged)" and "vm3_start (unmanaged)", and then this triggered 
libvirtd_stop, which leads to every VM stopping due to colocation.

The question is: is there a sane way to run VMs under pacemaker's 
control? If yes, is it described somewhere?


--
Pavel