When cluster nodes are started simultaneously rgmanager fails to start service with lvm resource in RHEL 6
Issue
- While testing for recovery after a power outage, both nodes are brought up at the same time. Rgmanager attempts to start both services on both nodes, and they fail.
- LVM resources fail to start if another node is starting rgmanager at the same time:
Jul 05 13:23:46 rgmanager Failed to activate logical volume, myvg/mylv
Jul 05 13:23:46 rgmanager Attempting cleanup of vgsan01/mylv
Jul 05 13:23:47 rgmanager Failed second attempt to activate vgsan01/nfs00
Jul 05 13:23:47 rgmanager start on lvm "nfs-lvm" returned 1 (generic error)
Jul 5 13:23:46 node2 lvm[10266]: Error locking on node node1.example.com: Volume is busy on another node
- I'm using HA-LVM with
clvmd
. I manually activated clustered volumes on one or more nodes in the cluster, and then those resources failed to start as part of their service on another node.
Environment
- Red Hat Enterprise Linux (RHEL) 5 or 6 with the Resilient Storage Add On
- Using HA-LVM with the
clvmd
variantlocking_type = 3
in/etc/lvm/lvm.conf
- A service in
/etc/cluster/cluster.conf
containing anlvm
resource
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.