Details
-
Bug
-
Resolution: Duplicate
-
Major
-
6.5.0
Description
- Create a cluster with 6 nodes
- Create a bucket with replicas=3
- Start a load generator with durability level = Majority or PersitToMajority
- Rebalance out node
- Rebalance in a node
Rebalance is getting struck intermittently.
This was found while running Jepsen tests. Attaching the logs from the run
While replicas=3 is not supported for durable writes, the rebalance should not hang irrespectively. The SDK throws durability impossible exception for the durable writes anyways.
Attachments
Issue Links
- duplicates
-
MB-35633 Aborts received by replica should not be ignored when matching prepare is missing
- Closed