-
Type:
Bug
-
Resolution: Fixed
-
Priority:
Major - P3
-
Affects Version/s: None
-
Component/s: None
-
None
-
Cluster Scalability
-
Fully Compatible
-
v8.0, v6.0
-
200
Resharding oplog application applies each delete operation in a transaction. Each one uses a brand new lsid and txnNumber=0.
While this is correct, using so many unique lsids can be problematic, as it makes the number of active sessions in the system grow without bounds. Instead, we could make each applier thread use it's own lsid, and use an incremented txnNumber to apply each one of the deletes.
- related to
-
SERVER-93586 [8.0] Tag concurrency tests that does moveChunk with resource_intensive
-
- Closed
-
-
SERVER-92435 Investigate if it is possible to make resharding generate fewer IncompleteTransactionHistory records
-
- Backlog
-
-
-
- Backlog
-