Red Hat Bugzilla – Bug 1243763
Pacemaker Systemd init doesn't wait the system to be online before starting, letting NFS resources to fail
Last modified: 2017-01-05 17:00:51 EST
Created attachment 1052613 [details] Patch with the changes needed to allow the cluster to deal correctly with NFS mount resources Description of problem: With pacemaker pacemaker-1.1.12 (22.el7_1.2) the pacemaker init script doesn't wait to have a proper network connection before coming up, and this drive the failure of NFS mount resources. This is particularly visible and disappointing when deploying OpenStack via osp-installer and choosing NFS. Version-Release number of selected component (if applicable): pacemaker-1.1.12-22.el7_1.2.x86_64 How reproducible: Always Steps to Reproduce: 1. add a NFS resource ocf::heartbeat:Filesystem 2. reboot one or all the nodes Actual results: When the cluster is up, the NFS resource will be disabled Expected results: The cluster wait to be online and then turn start all resources Additional info: Attached you can find a patch produced by Matteo Bernacchi that fix the pacemaker misbehaviour.
Created attachment 1052752 [details] systemd service.unit patch to allow the cluster to deal correctly with NFS mount resources
Comment on attachment 1052752 [details] systemd service.unit patch to allow the cluster to deal correctly with NFS mount resources Pacemaker doesn't require either of these items. The NFS unit or agent should be dealing with this.
See also: bug #1175005
I agree, it's a workaround for our specific problem. I've found this
Possibly something for the resource agent to take care of
(In reply to Andrea Perotti from comment #8) > I agree, it's a workaround for our specific problem. > > I've found this Can you confirm whether this is still an issue or not with the fixed rpcbind version in RHEL7.2 from: https://bugzilla.redhat.com/show_bug.cgi?id=1203820?