Details
-
Bug
-
Resolution: Not a Bug
-
Major
-
6.6.5
-
None
-
[6.6.5-153183 -> 6.6.5-15322]
-
Untriaged
-
Centos 64-bit
-
1
-
Unknown
Description
Setup was on the toy 6.6.5-153183 , When we got an another toy with fix 6.6.5-15322, tried upgrading the cluster using offline upgrade. Auto failover was enabled and set to smaller value. So, we see rollbackAllToZero in indexer logs.
Upgrade order -> 3 kv nodes(172.23.100.34, 172.23.105.37, 172.23.106.156) , followed by 2 indexer nodes(172.23.106.159, 172.23.106.163) and a n1ql(172.23.106.204) node in a serial fashion
172.23.106.159 : index
-bash-4.2# zgrep rollbackAllToZero indexer.log*
|
indexer.log:2022-08-02T22:41:48.967-07:00 [Info] StorageMgr::rollbackAllToZero MAINT_STREAM bucket3
|
indexer.log.2.gz:2022-08-02T22:37:25.063-07:00 [Info] StorageMgr::rollbackAllToZero MAINT_STREAM bucket2
|
-bash-4.2# zgrep 'Unable to find a snapshot older than last used Snapshot' indexer.log.*
|
indexer.log.2.gz:2022-08-02T22:37:24.987-07:00 [Info] StorageMgr::handleRollback 3648848136011536914 Unable to find a snapshot older than last used Snapshot SnapshotInfo: count:525000 committed:false. Use nil snapshot.
|
172.23.106.163 : index
-bash-4.2# zgrep rollbackAllToZero indexer.log*
|
indexer.log.1.gz:2022-08-02T22:41:04.369-07:00 [Info] StorageMgr::rollbackAllToZero MAINT_STREAM bucket3
|
indexer.log.2.gz:2022-08-02T22:37:16.012-07:00 [Info] StorageMgr::rollbackAllToZero MAINT_STREAM bucket2
|
-bash-4.2# zgrep 'Unable to find a snapshot older than last used Snapshot' indexer.log.*
|
indexer.log.1.gz:2022-08-02T22:41:04.258-07:00 [Info] StorageMgr::handleRollback 12772435220397978300 Unable to find a snapshot older than last used Snapshot SnapshotInfo: count:560000 committed:false. Use nil snapshot.
|
Cbcollect was failing as couchbase-cli broke post upgrade. So, had to take cbcollect almost 4 hours after the fact.