checkForInterruptNoAssert calls during handleRequest for getMores that result in ClientDisconnect (being thrown and markKilled) cause an issue where a new killCursor opCtx is created and the old opCtx is not delisted. Find a way to change this if condition for this exact condition, possibly making a more clear distinction between thrown statuses and opCtx killStatus.
ServiceEntryPointImpl delist logic
- is duplicated by
-
SERVER-79758 Always release `opCtx` before cleaning up session
- Closed
-
SERVER-79959 Ensure that opCtx has been delisted before creating new opCtx during SessionWorkflow cleanup
- Closed