Performing an "ls", "cd", "find" on a one of our mounted directories causes the "ls", "cd", "find" process to hang (becomes defunct process).
LVM is used but all other LV's in the same VG are fine. Multipath reports all paths are up. A "pvscan" detects all disks correctly.
No errors evident in dmesg or /var/log/messages (default logging enabled).
Inode count is low (<10). This probelm is very intermittent.
Any ideas where to start trouble shooting. Hopefully I can perform a bit more details diags next time around.
This problem is intermittent and acorss different mount points/servers. A restart of the server is the only way I know how to resolve.
Call with Red Hat (including a sosreport) has come back with nothing.

You mentioned multipath, does that mean this is SAN of some kind ? ... maybe you have some latency somewhere in the storage system - FC switches, congestion etc ... is the storage shared ?

Yes, the storage is presented to the host via the default rhel51 MPIO driver via IBM SAN32B switches (Brocade) on an IBM DS8100 disk stroage system. All infrastrcuture is shared with many hosts.
My thinking though is that the problem exists higher up in the stack. The filesystem did not return until a restart of the host.
The filesystem was one of 3 logical volumes sharing a 2 disk volume group. The other 2 filesystems were unaffected. lvscan and pvscan picked up the physcial and logical volumes without issue. All idications are that the disk was accessible but not from a file system level.
We have sluggish read times from time to time on the disks stroage system throughout the day due to load.