-
Type: Bug
-
Resolution: Unresolved
-
Priority: Major - P3
-
None
-
Affects Version/s: None
-
Component/s: None
-
Storage Execution
Writes only wait for writeConcern after they finish. This means that large multi:true updates or deletes can replicate unthrottled until they finish.
This can engage flow control, but because it is quite agressive, this negatively impacts other writes in the system.
The solution may involve having long-running writes periodically wait for writeConcern if they have previously replicated writes.