Details
-
Bug
-
Resolution: Won't Fix
-
Major - P3
-
None
-
3.4.6, 3.5.9
-
ALL
-
Sharding 2017-07-31
-
0
Description
Currently, the 'shardCollection' command's checking/refresh phase is not protected by a distributed lock.
This means that while that phase is performing a refresh, a concurrent drop can wipe out the chunks from underneath. Moving the distributed lock acquisition to happen inside the command itself should take care of this problem.
Attachments
Issue Links
- related to
-
SERVER-43487 Complete TODO listed in SERVER-30052
-
- Closed
-