"snapshot-latest" confusion on nexentastor 3.0.3 "dataset already exists"
|Status:||New||Start:||September 15, 2010|
|Assigned to:||Anatoly Legkodymov||% Done:||
[WARNINGS] link-id not found - sep 13 21:16:51 (zfs-auto-tier, 18099) rsync+ssh: /volumes/sr108-data => 192.168.7.3:/volumes/w360-data/backup ok (1 rsync invocations 5 sec)
Setup: I create local auto-snap snapshots of a folder every 4 hours. Additionally, every 4 hours I do a local-to-remote auto-tier replication of the same volume to two different destinations. This essentially works like a charm as expected with the following exception: The setup gets confused about the snapshots labeled "latest" and refuses to overwrite them. This leads to outdated snapshots bearing that particular labeling. It appears, that every auto-tier invocation triggers a snapshot generation and is confused when a file with such an ending already exists. This is not what I'd expect. latest should always be latest!
[ Sep 13 21:16:46 (zfs-auto-tier, 18099) from: localhost:/volumes/SR108-data, to: 192.168.7.3:/volumes/W360-data/backup, proto=rsync+ssh ] [ Sep 13 21:16:51 (zfs-auto-tier, 18099) rsync+ssh: /volumes/SR108-data => 192.168.7.3:/volumes/W360-data/backup OK (1 rsync invocations 5 sec) ] [ Sep 13 21:17:17 (zfs-auto-tier, 18099) Warning: could not create snapshot 'SR108-data@snap-hourly-4-latest' on 'localhost': com.nexenta.nms.SystemCallError: cannot create snapshot 'SR108-data/data@snap-hourly-4-latest': dataset already exists ] [ Sep 13 21:17:18 (zfs-auto-tier, 18099) WARNING: SR108-data@snap-hourly-4-2010-09-13-2116 OK but with some errors ]
Updated by Gernot Schmied over 2 years ago
Any comments on this behavior? This is rather important!