Write a "recovery" method on the MigrationManager for the case where it steps down (with a bunch of migrations going on) and then starts up again on a another primary. In this case the new MigrationManager will have to "inherit" the active migrations and wait for them to complete before doing another round.
*Write a JS test to test this, including querying the config.migrations collection at points to check it is correct.
- has to be done after
-
SERVER-25703 Enable dist lock unlock by sessionId AND _id (aka namespace)
-
- Closed
-
- has to be done before
-
-
- Closed
-
- is depended on by
-
SERVER-22935 Add JS/client function to retry moveChunk on mongod stepdown related errors
-
- Closed
-
- is related to
-
-
- Closed
-