If a shard's primary steps down, but the mongos can't connect to the new primary, the mongos continues thinking that the old primary is still the master. This will cause it to send setShardVersion commands to the former primary, which will fail and be retried many times.
- depends on
-
SERVER-5642 Make mongobridge work with ShardingTest to facilitate testing network partitions in a sharded system
- Closed
- is related to
-
SERVER-4661 Mongos doesn't detect primary change if old primary lost network connectivity
- Closed