-
Type:
Task
-
Resolution: Fixed
-
Priority:
Major - P3
-
Affects Version/s: None
-
Component/s: None
-
None
-
Replication
-
Fully Compatible
-
Repl 2026-08-31
-
None
-
None
-
None
-
None
-
None
-
None
-
None
There are 2 categories of fixes for DSC tests that this ticket will cover:
- Checking the tip of the oplog for a specific write
SERVER-131011andSERVER-133411made similar changes- If a background thread does a write before we check the oplog tip the test will fail, so apply this fix to other remaining places
- Fix is to filter the oplog on the entry we are expecting
- Pausing failpoint without confirming we hit it
SERVER-126321andSERVER-133455- Calling configureFailpoint just sets the failpoint without making sure we have entered it
- For the failpoint hangInPrimaryReadLoop we check the failpoint at the top of the loop, so if we set it mid loop and don't wait then we can end up sending another batch of entries to the log service which messes up the test
- Fix is to call wait() on the failpoint before proceeding to confirm we have entered the failpoint
- is related to
-
SERVER-131011 Fix race in disagg_storage/status.js between replSetGetStatus and oplog read caused by background writes
-
- Closed
-
-
SERVER-133455 primary_steps_down_from_heartbeat_higher_term.js is intolerant of old primary writes
-
- Closed
-
-
SERVER-133411 resume_after_against_oplog.js assumes no background oplog writes during awaitData
-
- Closed
-
-
SERVER-126321 Wait for failpoint on reconfig test
-
- Closed
-