NFS cluster service with nfsexport and nfsclient resources fails to stop in RHEL 6

Solution Verified - Updated -

Issue

  • Stopping a cluster service with <fs> or <clusterfs>, <nfsexport>, and <nfsclient> resource fails when clients are mounting using NFS version 4
  • nfs over gfs2 service group does not stop cleanly, failing to unmount clusterfs resource, and ending up in a failed state, despite force_unmount and self_fence being enabled
  • node is not fencing itself with self_fence set on a clusterfs resource when it fails to unmount the file system
  • nfs service goes into failed state instead of failing over:
Nov 27 10:06:42 rgmanager Stopping service service:test
Nov 27 10:06:43 rgmanager [ip] Removing IPv4 address 192.168.143.199/24 from eth1
Nov 27 10:06:54 rgmanager [nfsclient] Removing export: *:/mnt/lv1
Nov 27 10:06:55 rgmanager [fs] unmounting /mnt/lv1
Nov 27 10:06:56 rgmanager [fs] umount failed: 1
Nov 27 10:06:56 rgmanager [fs] Sending SIGTERM to processes on /mnt/lv1
Nov 27 10:07:02 rgmanager [fs] unmounting /mnt/lv1
Nov 27 10:07:02 rgmanager [fs] umount failed: 1
Nov 27 10:07:02 rgmanager [fs] Sending SIGKILL to processes on /mnt/lv1
Nov 27 10:07:02 rgmanager stop on fs "clust-lv1" returned 1 (generic error)

Environment

  • Red Hat Enterprise Linux (RHEL) 6 with the High Availability Add On
  • rgmanager
  • A cluster service containing one or more clusterfs or fs resources with force_unmount set, and nfsexport and nfsclient resource(s) in /etc/cluster/cluster.conf
  • One or more clients mounting the export using nfs version 4 (vers=4 mount option, which is the default in RHEL 6)

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