Dell Unity: Expired-Snap Deletion Is Slow
Summary: Impact- Under certain conditions, snaps appear to be in a Destroying state perpetually. This may even lead to unexpected instabilities under certain conditions.
Symptoms
Under certain condition, an internal snap deletion operation called snap-truncate (of expired user snaps) may take a long time and would seem to be ever going.
This issue can be examine in Unisphere per file system in the Snap page, or in CLI:
svc_storagecheck --list | grep -c Destroying 1917 Impact Statement: In case of expected or any unexpected SP reboots while a truncate operation is in progress, a problem can occur leading to Unisphere and Replication instabilities that would require Support L2 Recoveries. This may even lead to FileSystem offline condition and Replication Restart from Scratch (unrecoverable session).
FileSystem extension task may also be stopping responding as a result of this condition.
Cause
Under certain conditions, an expired-snaps deletion task takes longer time to process due to resources limitations.
Replication RPO may have direct cause to the number of refreshes per FileSytem, and as a result # of expired snaps pending truncation.
Resolution
In case the total number of snaps is not decreasing gradually and consistently, contact Dell Support and request further assistance on this knowledge base article.
Consider increasing the RPO value if replication is set up (reducing the number of updates per hour), and allow some time (~24 hours) to examine if the number of snaps pending deletion is decreased.
Immediate Workarounds:
Change All Active Snaps Expiration Time Frame to: Never Expire
This would ensure reduction of the pending deletion overall number.
Additional possible actions: Reduction of number of snaps per file system
Additional Information
Unisphere Error code 0x7d13001