I'm running OES SP2 (Linux) fully patched with NCS. The boxes use QLogic
HBAs (2460LE) attached to EMC Symmetrix with EMC PowerPath and Naviagent.
The volumes are traditional volumes and are configured with CSM segment

I seem to be experiencing a problem where occasionally one node in the
cluster will load the dreaded evmsd_worker thread and lock container
operations. I can't migrate resources at this point. I can restart that
box and I appear to have a rolling evmsd_worker issue that slowly spreads
across the cluster. Eventually restarting all nodes will remove the
locking evmsd_worker threads.

What are some potential causes of this condition? Is there something I
can do to prevent this from occurring?

I also seem to get lots of evms engine messages about dirty metadata on
the csm segments. Nothing in docs on that either other then evms
attempting to correct it.