Details
-
Bug
-
Resolution: Fixed
-
Major
-
5.5.0
-
Untriaged
-
Unknown
Description
This is due to race condition:
1) On a 3 node cluster, create partition index with 1 replica. Build index could be delayed until all replica definition is finished.
2) Immediately rebalance out one of the nodes
3) Rebalance could jump in before build index happens. The bug is that this could hang rebalance.
Attachments
Issue Links
- relates to
-
MB-28618 fix hang in rebalance partitioned index
- Closed