-
Type:
Bug
-
Resolution: Unresolved
-
Priority:
Major - P3
-
None
-
Affects Version/s: 8.0.10, 8.1.1
-
Component/s: Sharding
-
None
-
Catalog and Routing
-
ALL
-
CAR Team 2025-06-09, CAR Team 2025-06-23
-
200
-
None
-
3
-
TBD
-
🟥 DDL
-
None
-
None
-
None
-
None
-
None
-
None
-
0
SERVER-99440 added a new parameter to time out checkMetadataConsistency command after grabbing a database DDL lock. In order to check that this new parameter and maxTimeMS work, a test was added that sets the new added parameter with a high value, and maxTimeMS with a small value of 100 ms. We can totally set the new parameter to lower values in previous checks of the test, because the time starts to count once the DDL lock is taken, however, this is not the same for maxTimeMS, this parameter is tracked even in the router, so, it is totally possible that if for some reason the host is slower than usual, we can timeout checkMetadataConsistency before reaching the node, which is expecting for a failpoint to be hit.
We need to change this test so we take into account that in slow machines the cluster command might actually take a certain amount of time.
- is caused by
-
SERVER-99440 Add timeout parameter for check metadata consistency database operation
-
- Closed
-