APAR status
Closed as program error.
Error description
The Spectrum Protect Plus maintenance job ends with the status "Partial" as seen in the following job log example : SUMMARY,,<timestamp>,,CTGGA2398,Starting job for policy Maintenance. id -> <JobID>. IBM Spectrum Protect Plus version 10.1.5-2181. ... WARN,,<timestamp>,2,CTGGA1770,Deleting storage snapshot for policy <policy_name> Failed to destroy storage Snapshot message :"Failed to delete snapshot: Command failed: cannot create bookmark 'vpool1/fs<xx>#snap<yyyy>': bookmark exists" type :"SnapshotDeleteError" ... DETAIL,,<timestamp>,,CTGGA0814,Task 2 : Hypervisor Copy completed with status PARTIAL zfs bookmarks are used by Spectrum Protect Plus when the snapshot to delete on the source vSnap host is in a replication relationship and it is the last present. This permits to send only the incremental changes of a subsequent snapshot on a next replication instead of sending a full sized base and snapshot data if no previous snapshot or bookmark exists. In the vSnap log we can see that prior to delete the snapshot, a zfs bookmark creation is attempted and fails because there is already one existing. This aborts the snapshot deletion : [<timestamp>] INFO pid-<12345> vsnap.api API request started: DELETE /snapshot/<yyyy>? | Body: None [<timestamp>] INFO pid-<12345> vsnap.common.util Authorized user: vsnap_user [<timestamp>] INFO pid-<12345> vsnap.zfs Deleting snapshot id <yyyy> [<timestamp>] WARNING pid-<12345> vsnap.zfs Creating bookmark for snapshot id <yyyy> because relationship <replication_ID> may need it for future syncs [<timestamp>] INFO pid-<12345> vsnap.linux.system Executing command: zfs bookmark vpool1/fs105@snap<yyyy> vpool1/fs<xx>#snap<yyyy> [<timestamp>] ERROR pid-<12345> vsnap.linux.system Return code 1: zfs bookmark vpool1/fs<xx>@snap<yyyy> vpool1/fs<xx>#snap<yyyy> [<timestamp>] ERROR pid-<12345> vsnap.linux.system Ouput: ["cannot create bookmark 'vpool1/fs<xx>#snap<yyyy>': bookmark exists"] An existing bookmark would for example be possible if a previous snapshot deletion process started, created the bookmark but then abnormally stopped for any reason before removing the snapshot. This situation would be seen with the following command on the replication source vSnap host : zfs list -t filesystem,snapshot,bookmark | grep fs<xx> which would list for example : vpool1/fs<xx> 43K 122G 27K /vsnap/vpool1/fs<xx> vpool1/fs<xx>@snap914 0B - 27K - vpool1/fs<xx>#snap914 - - - - ==> this is a bookmark using '#' while snapshots use'@' IBM Spectrum Protect Versions Affected: IBM Spectrum Protect Plus 10.1.4 and 10.1.5 Initial Impact: Medium Additional Keywords: SPP, SPPLUS, TS003530631, replication, bookmark, snapshot, delete, expiration
Local fix
1. Identify with the above maintenance job log message the correct bookmark reference. In the above example, it is 'vpool1/fs<xx>#snap<yyyy>' 2. Manually delete the offending bookmark on the replication source vSnap host with the command : sudo zfs destroy "vpool1/fs<xx>#snap<yyyy>" 3. Let the next maintenance job run proceed with the snapshot deletion.
Problem summary
**************************************************************** * USERS AFFECTED: * * IBM Spectrum Protect Plus level 10.1.5 * **************************************************************** * PROBLEM DESCRIPTION: * * See Error Description * **************************************************************** * RECOMMENDATION: * * Apply fixing level when available. This problem is currently * * projected to be fixed in IBM Spectrum Protect Plus level * * 10.1.6. Note that this is subject to change at the * * discretion of IBM. * ****************************************************************
Problem conclusion
When a snapshot is expired by the Maintenance job and that snapshot is the most recent replicated one on the volume, an alternate bookmark is created so that future replication operations can use the bookmark to perform incremental replication of later snapshots. After creating the bookmark if the snapshot expiration fails for any other reason, the bookmark is left over. The next time SPP Maintenance tries to expire the same snapshot, vSnap tries the create the bookmark again but this fails because the previous bookmark already exists. The problem has been resolved by detecting and using existing bookmarks, if present, before trying to create new ones.
Temporary fix
Comments
APAR Information
APAR number
IT32541
Reported component name
SP PLUS
Reported component ID
5737SPLUS
Reported release
A15
Status
CLOSED PER
PE
NoPE
HIPER
NoHIPER
Special Attention
NoSpecatt / Xsystem
Submitted date
2020-04-14
Closed date
2020-04-29
Last modified date
2020-04-29
APAR is sysrouted FROM one or more of the following:
APAR is sysrouted TO one or more of the following:
Fix information
Fixed component name
SP PLUS
Fixed component ID
5737SPLUS
Applicable component levels
[{"Business Unit":{"code":"BU058","label":"IBM Infrastructure w\/TPS"},"Product":{"code":"SSNQFQ","label":"IBM Spectrum Protect Plus"},"Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"A15","Line of Business":{"code":"LOB26","label":"Storage"}}]
Document Information
Modified date:
31 January 2024