As part of my normal housekeeping, I keep a close eye on my DDB’s. I’ve learned enough lessons over the years to look closely at pending delete records, run regular DDB verifications, and purge orphan files.
Recently we have had two separate issues where physical pruning of data was not occurring on the physical Storage.
The first was a corruption of Storage Permissions on our Isilon that prevent physical pruning.
The second was related to corrupt records in the DDB that prevented physical pruning.
In both cases the DDB looked healthy. None of the regular indicators such as a high “Number of Pending Prunable Records” showed a problem. The graphs showed significant drops in Primary and Secondary Blocks.
Both these problems had been there for a long period of time and were only detected because we were looking for an anomaly. In the second case we were fortunate that we were carry out bulk deletes and some activity occurred that would not have happened during normal operation that gave us a significant clue.
My question is, what other regular health checks do I need to run to ensure that my DDB and physical pruning is healthy. The regular checks failed to detect these problems
