NSS allows you to temporarily deactivate individual storage pools to fix volume problems instead of bringing down the server. However, when you deactivate a storage pool, users do not have access to any of the volumes in that pool. All of the volumes on the pool are part of the verify or rebuild process.
The purpose of the Pool Verify and Pool Rebuild utilities is to make sure you have a valid metadata structure for a pool. Use the utilities only when you have problems with the pool’s metadata structure.
Verifying a pool does not fix any problems. It is a read-only assessment of the pool’s metadata structure to identify the types of errors, the severity of errors, and in what volumes the errors occur.
Rebuilding a pool restores the consistency of the pool’s metadata structure. Rebuilding a pool does not restore lost data and does not repair the data integrity of corrupted data.
Rebuilding a pool does not fix problems for the following:
Journaling errors
Hardware and media errors
File system trustee assignments, trustee rights, and inherited rights filters
File system attributes for files and directories
Opportunistic locking
Content of files
Volume errors are typically transactions left unfinished during a system crash of some kind. Most volume errors are fixed automatically during volume mount when NSS resolves the journaled errors. If the pool can be mounted, mount its volume to allow the NSS journaling feature to repair any transactional errors that occurred during a system failure.
Afterwards, there if there are still problems, use diagnostic tools to rule out hardware problems as the cause.
If non-hardware errors persist, and if you have a viable backup to restore the pool to the last known good state, restore the backup to recover the pool and restore the data. It is probably not necessary to verify or rebuild the pool.
If non-hardware errors persist, and if you do not have a viable backup, use the following Pool Verify utilities to identify any errors in the pool’s metadata:
Review the verification log to determine the type and severity of problems with the pool’s metadata.
If all of the following conditions exist, then you should rebuild the pool to restore its metadata integrity.
Errors were not corrected by mounting the volume, or you could not mount the volume.
Errors were not caused by media or hardware problems, or they persisted after correcting any media or hardware issues.
You have no viable backup of the pool’s volumes to restore the pool to the last known good state.
The Pool Verify process reports errors in the physical integrity of any of the volumes’ metadata that would definitely cause data corruption if no action is taken.
More data will be lost from continued data corruption than will be lost from rebuilding the pools now.
WARNING:You should rebuild a pool only as a last resort to restore the consistency of the pool’s metadata. The rebuild repairs the metadata; it does not recover lost data or repair the integrity of the data itself. Data loss occurs during a rebuild if the utility must prune leaves in the data structure to restore metadata consistency.
If the Pool Verify process did not report errors, but you cannot create files or directories, you should run rebuild with the ReZID option. For information, see Section 17.3, ReZIDing Volumes in an NSS Pool.
If you are not sure whether you can tolerate a system rebuild, take a pool snapshot and run the rebuild against the pool snapshot instead. Then if the rebuild is acceptable, you can replace the pool with the rebuilt snapshot. For information about pool snapshots, see Section 18.0, Managing NSS Pool Snapshots.
If necessary, rebuild the pool’s metadata by using the following utilities: