[Linux-cluster] Bind mounts with HA ext3

[Johnson, Eric]

I've got an RHEL 5.4 NFSv4 failover cluster that uses cluster controlled
ext3 file systems on CLVM. In order to bring all the mount points into
the NFS root, I have to use a bind mount.

Eg. Cluster mounts the file system at /fs1 then I have a script resource
that creates the bind mount to /nfsroot/fs1

However, this seems to confuse the resource manager into thinking the
file system is mounted in a location it is not configured to:

Sep 28 14:33:12 node1 clurgmgrd: [4150]: <warning> Device
/dev/nfsvg/fs1lv is mounted on /nfsroot/fs1 instead of /fs1
Sep 28 14:33:12 node1 clurgmgrd: [4150]: <warning> Device
/dev/nfsvg/fs2lv is mounted on /nfsroot/fs2 instead of /fs2
Sep 28 14:33:12 node1 clurgmgrd: [4150]: <warning> Device
/dev/nfsvg/fs3lv is mounted on /nfsroot/fs3 instead of /fs3

I can't set the real mount point into the /nfsroot directory due to an
application that expects them somewhere else on the same cluster.

This warning doesn't *seem* to impact actual operation of the cluster,
so I've just suppressed the messages through an rsyslog rule, but is
there a way to inform the cluster to ignore the bind mount? I'm using
RHEL 5.4 32-bit with all the latest updates as of today, including:

kernel-2.6.18-164.el5
cman-2.0.115-1.el5_4.2
rgmanager-2.0.52-1.el5
openais-0.80.6-8.el5

Eric