Details
-
Bug
-
Resolution: Fixed
-
Major
-
7.6.0
-
Untriaged
-
0
-
Unknown
Description
Currently in shard rebalance, batching works such that we can theoretically schedule a `copy` and `move` for the same alternate shards in the same batch. Theoretically, This can lead to alternate shard being dropped (by `move` token state changes) before we finish the copy of the same.
Avoid the same
Attachments
Issue Links
- is duplicated by
-
MB-59621 [System Test] Rebalance failure - error locking shard - shardID not present
- Closed
For Gerrit Dashboard: MB-59602 | ||||||
---|---|---|---|---|---|---|
# | Subject | Branch | Project | Status | CR | V |
200952,5 | MB-59602 skip posting move tokens for shard to be copied | unstable | indexing | Status: MERGED | +2 | +1 |