Details
-
Bug
-
Resolution: Not a Bug
-
Major
-
6.6.0
-
Enterprise Edition 6.6.0 build 7883
-
Untriaged
-
-
1
-
Unknown
Description
Build: 6.6.0-7883
Scenario: (All doc_ops with doc_durability=MAJORITY)
- 2 node cluster, Couchbase bucket (replica=1)
- Load 250K docs into the bucket
- Rebalance-in 2 more nodes into the cluster. (4 node cluster with 250K docs)
- Start following doc update ops in a thread:
- Update with expiry=0 for doc indexes 0-125K
- Update with expiry=10 for doc indexes 127K-250K
- Rebalance-out one node from the cluster and rebalance-in the same node back
- Wait for doc_updates to complete and then sleep 12 secs for docs to expire
- Validate doc count=127K
Observation:
At the end of the above scenario, the doc_count is ~ 128K (Expected 127K)
Test log: http://qa.sc.couchbase.com/job/oel6-4node-rebalance_in_jython/1107/consoleText
SDK Client side pcap:https://cb-jira.s3.us-east-2.amazonaws.com/logs/expiry_docs/client.pcap
Cluster nodes pcaps: http://qa.sc.couchbase.com/job/oel6-4node-rebalance_in_jython/1107/artifact/logs/testrunner-20-Jul-21_02-01-13/test_1/