Node that is rejoining the cluster is immediately fenced after a "not installed" error from an exportfs resource in a RHEL 6 Update 5 High Availability cluster with pacemaker

Solution Unverified - Updated -

Issue

  • A node that has rebooted is starting pacemaker and logs errors referencing "not installed" in relation to an exportfs resource, following a message that the configured directory "does not exist or is not a directory. Following this the node gets fenced by the other node
  Dec 28 18:12:26 [7123] node2       crmd:     info: do_lrm_rsc_op:     Performing key=22:3:0:376790c0-18ba-402f-a87c-5cd437a72ff7 op=myEXPORT_stop_0
  Dec 28 18:12:26 [7120] node2       lrmd:     info: log_execute:   executing - rsc:myEXPORT action:stop call_id:43
exportfs(EXPORT-DEPLOY)[8268]:  2014/12/28_18:12:26 ERROR: /nfs/export does not exist or is not a directory
  Dec 28 18:12:27 [7120] node2       lrmd:     info: log_finished:  finished - rsc:myEXPORT action:stop call_id:43 pid:8268 exit-code:5 exec-time:1001ms queue-time:0ms
Dec 28 17:12:28 [6970] node1    pengine:    error: native_create_actions:        Resource myEXPORT (ocf::exportfs) is active on 2 nodes attempting recovery
[...]
Dec 28 17:12:28 [6970] node1    pengine:  warning: native_create_actions:        See http://clusterlabs.org/wiki/FAQ#Resource_is_Too_Active for more information.
[...]
Dec 28 17:12:30 [6970] node1    pengine:   notice: unpack_rsc_op:        Preventing myEXPORT from re-starting on node2: operation stop failed 'not installed' (rc=5)
Dec 28 17:12:30 [6970] node1    pengine:  warning: unpack_rsc_op:        Processing failed op stop for myEXPORT on node2: not installed (5)
Dec 28 17:12:30 [6970] node1    pengine:  warning: pe_fence_node:        Node node2 will be fenced because of resource failure(s)
Dec 28 17:12:30 [6970] node1    pengine:  warning: stage6:       Scheduling Node node2 for STONITH
  • When starting pacemaker after booting up, a node's logs show that an exportfs resource is "Too Active" and that the directory is already exported before the cluster has started it

  • System crashes just after full boot when starting pacemaker with an exportfs resource

Environment

  • Red Hat Enterprise Linux (RHEL) 6 Update 5 with the High Availability Add On
  • pacemaker
  • resource-agents releases prior to 3.9.5-12.el6
  • One or more exportfs resources configured in the cluster
    • exportfs resource's directory attribute corresponds to a Filesystem resource also managed by the cluster

Subscriber exclusive content

A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.

Current Customers and Partners

Log in for full access

Log In

New to Red Hat?

Learn more about Red Hat subscriptions

Using a Red Hat product through a public cloud?

How to access this content