-
Type: Task
-
Resolution: Fixed
-
Priority: Major - P3
-
Affects Version/s: None
-
Component/s: Sharding
-
Fully Compatible
-
Sharding EMEA 2021-08-09, Sharding EMEA 2021-08-23, Sharding EMEA 2021-09-06, Sharding EMEA 2021-09-20, Sharding EMEA 2021-10-04, Sharding EMEA 2021-10-18
The acquisition of the critical section should happen once the recipient has the up to date version of the documents of the chunk being migrated.
We should block write operations on the recipient shard.
This ticket only deals with the happy-path (i.e. non failover). The failover recovery will be addressed in a follow-up ticket (SERVER-60645).
- depends on
-
SERVER-59965 Distributed deadlock between renameCollection and multi-shard transaction
- Closed
- is depended on by
-
SERVER-58992 Force a refresh of the filtering metadata on the recipient shard during the commit phase of a moveChunk
- Closed
- is related to
-
SERVER-60415 No longer accept FailPointSetFailed errors on jstests that use migrationRecipientFailPostCommitRefresh failpoint
- Closed
-
SERVER-60982 Evaluate the performance of the new migration protocol
- Closed
- related to
-
SERVER-60645 Recoverability of recipient shard critical section of a moveChunk operation
- Closed
-
SERVER-60984 Report time in recipient critical section on serverStatus' shardingStatistics
- Closed