Its a particular case of SERVER-29987. As a result the scenario
coll.createIndex({x:1});
coll.explain().find({x:1 });
will may not the use the index if the explain goes to the secondary
Enabling support for readConcern will make the explain command wait for the clusterTime of the createIndex result and hence will be able to see the updated schema.
- duplicates
-
SERVER-29987 Make all non-write operations take readConcern
-
- Closed
-