-
Type: Bug
-
Resolution: Fixed
-
Priority: Major - P3
-
Affects Version/s: None
-
Component/s: None
-
None
-
Fully Compatible
-
ALL
-
Sharding EMEA 2022-10-03
On step-up, range deletion tasks are recovered from disk by spawning a thread that is asynchronously taking care of registering tasks on the service in order to avoid blocking the transition to primary.
Since tasks registration requires the acquisition of the service lock, it's needed to synchronize the resuming with shutdown that acquire the same mutex and may deadlock when trying to join the executor the resume is scheduled on.
- causes
-
SERVER-70888 ScopedRangeDeleterLock might lead to a deadlock on stepdown
- Closed