Memcached losing connection intermittently fails with "bad match" error [Changed bucket password, and started load on the cluster]

Setup:

1.Setup a 2:2 node xdcr replication cluster.
2. Setup 2 unidirectional replications from cluster1-cluster2 and cluster2-cluster1
3. Load running on cluster1- replicating to cluster2
4. Remove password on bucket-2.
5. Start load on cluster2.

Output
On cluster2, losing memcached connection with "bad match" –

Control connection to memcached on 'ns_1@10.3.3.28' disconnected: {{badmatch,
{error,
closed}},
[

{mc_client_binary, cmd_binary_vocal_recv, 5}

,

{mc_client_binary, select_bucket, 2}

,

{ns_memcached, ensure_bucket, 2}

,

{ns_memcached, handle_info, 2}

,

{gen_server, handle_msg, 5}

,

{proc_lib, init_p_do_apply, 3}

]} (repeated 1 times)

-----------------

Port server memcached on node 'ns_1@10.3.3.28' exited with status 134. Restarting. Messages: TAP (Producer) eq_tapq:replication_ns_1@10.3.3.29 - Sending TAP_OPAQUE with command "close_backfill" and vbucket 432

Adding logs from /var/log here https://s3.amazonaws.com/bugdb/jira/bug-memcached/memcached-error.tar

We had seen a similar bug on 181 on bucket-password change.?