-
Type: Bug
-
Resolution: Fixed
-
Priority: Major - P3
-
Affects Version/s: None
-
Component/s: None
-
Catalog and Routing
-
Fully Compatible
-
ALL
-
v8.0
-
CAR Team 2024-06-24, CAR Team 2024-07-08, CAR Team 2024-07-22
-
200
In testing, we have observed cases where a deleteMany running while a moveCollection commits returns a StaleConfig error which is retried by mongos, leading to an unexpected value for n in the response (because some of the deletes happened on the donor shard, then the rest of the deletes happened on the recipient shard after the retry, and n only captures the deletes that occurred on the recipient).
We are concerned that similar behavior exists for updateMany where a StaleConfig after a moveCollection to lead it to retry and update the same document twice (once on the donor originally, and once again on the recipient after the retry). However, at the time of writing, this has not been observed to happen in testing.
This ticket is to do the following:
1. Verify whether or not the scenario described above can occur for updateMany.
2. If it can, make updateMany fail with some error that mongos will not retry instead of StaleConfig.
- causes
-
SERVER-95244 Upsert statements which result in an insert may fail with tassert 9146500 when client connects directly to shard
- Closed
- has to be done after
-
SERVER-91568 Switch ShardRoleTest to use ShardServerTestFixture
- Closed
-
SERVER-91580 Create Concurrency Test to Verify Changes from SERVER-91465
- Closed
- is depended on by
-
SERVER-91634 Enable Disruptive States in FSM Tests Added By SERVER-91580
- Closed
- is related to
-
SERVER-96228 Fix QueryPlanKilled error in random_moveChunk_multi_update_delete_paused_migrations.js
- Closed
- related to
-
SERVER-92369 Fix QueryPlanKilled error in random_moveChunk_multi_update_delete_change_streams.js
- Closed
-
SERVER-92126 Consolidate functions used for iterating PlanExecutor in bulkWrite update statements and update command
- Closed