Edited to only be about not blocking on the network while waiting for replication. The task should run on a separate executor (not the fixed executor) as part of SERVER-46075.
Original description:
It currently blocks on the network for:
- waiting for the last opTime to be majority committed
- sending _configsvrEnsureChunkVersionGreaterThan
- forcing a filtering metadata refresh
These sub-tasks should be made asynchronous.
- related to
-
SERVER-46075 Use asynchronous version of forceShardFilteringMetadataRefresh when submitting range deletion task
-
- Closed
-