We've had an incident where due to an error in one of our volumes the pool was deactivated. The server is XEN Guest NW6.5 SP7 on a SLES 10 SP2 X86-64 Host, storage is on a SAN which has reported no errors, SYS is on the same SAN but a different LUN. The LUN that has the pools and volumes in question is 1TB in size which the pool takes 100%.
The only lines in the server error log from the time of the incident are as follows

Code:
18-09-2008   1:00:29 pm:    COMN-3.26-1196
   Severity = 4  Locus = 3  Class = 0
   NSS-3.00-5002: Pool LONDON/MAIN is being deactivated.
   An error (20012(nameTree.c[45])) at block 2(file block -2)(ZID 6) has compromised volume integrity.

18-09-2008   1:00:29 pm:    COMN-3.26-1195
   Severity = 4  Locus = 3  Class = 0
   NSS-3.00-5000: Volume LONDON/STAFF is being deactivated.
   An error (20012(nameTree.c[45])) at block 2(file block -2)(ZID 6) has compromised volume integrity.
After been unable to reactivate the pool we rebooted the server and the pool and volumes all activated fine and lasted about 2 minutes before the same issue occurred with a slightly different set of log lines

Code:
18-09-2008   1:09:12 pm:    COMN-3.26-1196
   Severity = 4  Locus = 3  Class = 0
   NSS-3.00-5002: Pool LONDON/MAIN is being deactivated.
   An error (20012(beastTree.c[510])) at block 237911777(file block -237911777)(ZID 1) has compromised volume integrity.

18-09-2008   1:09:12 pm:    COMN-3.26-1195
   Severity = 4  Locus = 3  Class = 0
   NSS-3.00-5000: Volume LONDON/STAFF is being deactivated.
   An error (20012(beastTree.c[510])) at block 237911777(file block -237911777)(ZID 1) has compromised volume integrity.
We attempted a NSS /PoolRebuild which failed. We created a new SAN LUN and created a new set of Pool and Volumes and restored from backup.The original LUN with the corrupt poll and volumes was moved to a different server to try the rebuild on. This rebuild succeeded and we were able to copy newer files from the now repaired Volumes to the newly created Volumes. At this stage we can find no cause for the issue occurring in the first place and are interested if anyone has some ideas on how we can locate the cause.