Bug 709398

Description of problem:
During the shutdown of a service group, a cluster node removes the virtual IP address from its ethernet interface, removes the export, and then tries to unmount the logical volume.  When it tries to unmount the logical volume, rgmanager debug reports that the mount point is "still in use by 1 other service(s)".  The service group is disabled, but the mount points remain on the node. If I manually unmount the logical volumes at this point, it succeeds.  It should also be noted that the force unmount option is specified as on for each of the mount points. This should kill the processes using the mount point and allow it to succeed, but does not.

Version-Release number of selected component (if applicable):
3.0.12-11

How reproducible:
Always

Steps to Reproduce:
1. Create a cluster configuration with the service groups set up like in the attached cluster.conf file
2. Enable a service group on a node
3. Disable the service group
  
Actual results:
When disabling the service group, it fails to unmount the mount points associated with the service group. After the service group is disabled, if I manually run "umount /home/smashmnt04", it succeeds. If the service group is started on another node, this results in the same mount point being mounted on more than one node at a time.
During the service group shutdown, rgmanager reports:
<debug>  Not unmounting clusterfs:lvol4 - still in use by 1 other service(s)
Not unmounting clusterfs:lvol4 - still in use by 1 other service(s)

Expected results:
In RHEL6.0, this exact configuration would result in the mount points being unmounted when the service "service-192.168.20.100-1" was disabled.

Additional info:
The exact same configuration on RHEL5 (U5, U6), and RHEL6.0 has been tested and does not experience this problem.

Additionally, I tried configuring a service group as follows, but this resulted in exactly the same problem:

<service domain="kswc-achilles1" exclusive="1" name="service-192.168.20.100-1" recovery="relocate">
<clusterfs ref="lvol0">
<nfsexport ref="ex-sm00">
<nfsclient ref="@local-sm00">
<nfsclient ref="@client-sm00">
<ip ref="192.168.20.100"/>
</nfsclient>
</nfsclient>
</nfsexport>
</clusterfs>
<clusterfs ref="lvol4">
<nfsexport ref="ex-sm04">
<nfsclient ref="@local-sm04">
<nfsclient ref="@client-sm04">
<ip ref="192.168.20.100"/>
</nfsclient>
</nfsclient>
</nfsexport>
</clusterfs>
</service>