Details
-
Bug
-
Status: Closed
-
Major - P3
-
Resolution: Fixed
-
3.1.7
-
None
-
Fully Compatible
-
ALL
-
-
Quint 9 09/18/15
Description
It appears the ReplicaSetMonitor that the shell is using to connect to replica sets is being cached, so that if that replica set is stopped, and a new replica set is started with the same name, it cannot connect to the set.
This came up while working on SERVER-18272, where moving away from the hard-coded 31000 start port made it impossible to start two sharded clusters with the default name. But I believe the steps to reproduce hit the same issue.
Attachments
Issue Links
- is depended on by
-
SERVER-18268 Support running sharding and replication tests in parallel
-
- Closed
-