-
Type: Bug
-
Resolution: Fixed
-
Priority: Major - P3
-
Affects Version/s: None
-
Component/s: Sharding
-
Sharding EMEA
-
Fully Compatible
-
ALL
-
Sharding EMEA 2023-03-06, Sharding EMEA 2023-03-20, Sharding EMEA 2023-04-03
-
5
SERVER-70600 is a mitigation that prevents a shard to believe it is the primary database even after the termination of a drop database coordinator. However, we could still allow a write to sneak in after a stepdown, but after the commit was sent to the CSRS, allowing the creation of an unsharded collection that will be never deleted.
The purpose of this task is to revisit the dropDatabase DDL operation and change its implementation to start using the Recoverable Critical Section and prevent leaving unsharded collections in the shards that had executed a drop database.
- causes
-
SERVER-75642 Fixing condition to mitigate some issues with the database metadata during a dropDatabase op
- Closed
-
SERVER-75635 Pre 7.0 dropDatabase implementation releases the critical section too early
- Closed
- related to
-
SERVER-70600 dropDatabase DDL operation might complete with the old primary shard still believing that is the primary
- Closed
-
SERVER-74880 movePrimary participant should drain DDL ops over the same dbName before acquiring the critical section
- Closed
- split from
-
SERVER-70600 dropDatabase DDL operation might complete with the old primary shard still believing that is the primary
- Closed