Details
-
Bug
-
Resolution: Fixed
-
Major
-
5.5.4
-
Enterprise Edition 5.5.4 build 4340
-
Untriaged
-
Centos 64-bit
-
-
No
Description
Build: 5.5.4 build 4340
Scenario:
- Create 4node cluster
- Create "default" couchbase bucket
- Load initial data 1K
- Perform failover of node-2 using `stop couchbase server`
- Perform update operation on 1K of docs again
- Validate vbucket_seqno on each node
Observation:
Few vbuckets in node-3 has last_persisted_snap_start > last_persisted_seqno.
vb_356:abs_high_seqno: 0
|
vb_356:high_seqno: 0
|
vb_356:last_persisted_seqno: 0
|
vb_356:last_persisted_snap_end: 1
|
vb_356:last_persisted_snap_start: 1
|
vb_356:purge_seqno: 0
|
Test case used:
From testrunner:
|
failover.failovertests.FailoverTests.test_failover_stop_server,replicas=1,graceful=False,num_failed_nodes=1,load_ratio=10,flusher_batch_split_trigger=3,GROUP=P1
|
Attachments
For Gerrit Dashboard: MB-34329 | ||||||
---|---|---|---|---|---|---|
# | Subject | Branch | Project | Status | CR | V |
109983,2 | MB-34329: reset snapshot range to be a valid range | vulcan | kv_engine | Status: MERGED | +2 | +1 |
110106,1 | Merge couchbase/vulcan into couchbase/alice | alice | kv_engine | Status: MERGED | +2 | +1 |
110143,2 | Merge couchbase/alice into couchbase/master | master | kv_engine | Status: MERGED | +2 | +1 |
110577,2 | Revert "Pin kv_engine manifest for 6.0.2" | master | manifest | Status: MERGED | +2 | +1 |