-
Type: Bug
-
Resolution: Fixed
-
Priority: Major - P3
-
Affects Version/s: 6.0.6, 5.0.18, 7.0.0-rc4
-
Component/s: None
-
Sharding NYC
-
Fully Compatible
-
ALL
-
v7.1, v7.0, v6.0, v5.0
-
8
-
4
It is possible that a resharding donor/recipient service POS executor has shutdown on a primary node but has failed to step down due to no secondaries being caught up to be elected. While the secondaries catchup, the primary which has its services shutdown can receive a retry of shardSvrCommitReshardCollection command from an in-progress resharding operation and uassert on existence of resharding state documents (donor or recipient) here. This leads to a fatal error at the coordinator causing it to crash.
- depends on
-
SERVER-78108 POS interface should expose its shutdown state
- Closed
- is related to
-
SERVER-81115 ReplicaSetAwareService Can Be Shutdown While Node is Still Primary
- Open
- related to
-
SERVER-78839 Investigate changes needed in sharding POS services to use the new return value for instance lookup
- Backlog