Continuous NotMyVBucketException on get ops when rebalancing in nodes

During a situational test doing Kv ops, and rebalancing in 2 nodes, the throughput of operations drops significantly.

http://sdkqe-testresults.couchbase.com.s3.amazonaws.com/SDK-SDK/CB-7.1.0-2549/Rb2In-KV/06-10-22/052111/52819971efd250979a1b6506b96372a4-MC.html

Logs: SdkdConsole.log.zip

Looking at the logs, it seems that the get operations fail due to Couchbase.KeyValue.NotMyVBucketException, and are retried many times. And set operations also are failing with NMVB, and the server is returning VBucketBelongsToAnotherServer.

It seems that only rebalancing in causes this issue, and other rabalance in, or rebalance swap operations are ok and the kv op throughput returns to base levels.

Tested with SHA: a3611b3c8f78c6bbee0d112ede11cc115dd9fa31
Cluster version : 7.1.0, and observed the same behaviour on 7.0, 6.6, 6.0.