-
Type: Task
-
Resolution: Unresolved
-
Priority: Major - P3
-
None
-
Affects Version/s: None
-
Component/s: Sharding, Testing Infrastructure
-
Labels:None
-
Cluster Scalability
-
6
-
1
The lack of synchronization between createIndexes and dropIndexes is known to result in cases where shards may permanently have differing indexes. FSM workloads, such as snapshot_read_catalog_operations.js, which run as part of the concurrency framework are capable of executing concurrent createIndexes and dropIndexes commands on the same collection and can trigger this type of inconsistency. We should remove the validation from the CheckClusterIndexConsistency hook until the guarantees provided by sharded clusters are strengthened.
- is depended on by
-
SERVER-71510 Undenylist snapshot_read_catalog_operations.js from sharded initial sync suite
- Backlog