Seeing few items not drained to disk on destination cluster ( Seeing 6 items in disk-write queue for over an hour).

1. Setup a XDCR replication between 2 clusters.
2. Load 2 M items on source cluster
3. 2M items are replicated to the destination cluster. - Works as expected.
4. Seeing last 6 items on the disk-write-queue for over an hour. Looks like these items are not being drained to the disk.

Stats from the nodes
ketaki@ubu-2513:~$ /opt/couchbase/bin/cbstats localhost:11210 -b bucket1 dispatcher
dispatcher
runtime: 655ms
state: dispatcher_running
status: running
task: IdleTask (sleeping)
nio_dispatcher
runtime: 2s
state: dispatcher_running
status: running
task: IdleTask (sleeping)
ro_dispatcher
runtime: 649ms
state: dispatcher_running
status: running
task: IdleTask (sleeping)
tap_dispatcher
state: dispatcher_runni

ketaki@ubu-2513:~$ /opt/couchbase/bin/cbstats 10.3.3.36:11210 -b bucket1 dispatcher
dispatcher
runtime: 556ms
state: dispatcher_running
status: running
task: IdleTask (sleeping)
nio_dispatcher
runtime: 485ms
state: dispatcher_running
status: running
task: IdleTask (sleeping)
ro_dispatcher
runtime: 17ms
state: dispatcher_running
status: running
task: IdleTask (sleeping)
tap_dispatcher
state: dispatcher_running
status: idle
ketaki@ubu-2513:~$ /opt/couchbase/bin/cbstats 10.3.3.35:11210 -b bucket1 dispatcher
dispatcher
runtime: 402ms
state: dispatcher_running
status: running
task: IdleTask (sleeping)
nio_dispatcher
runtime: 1s
state: dispatcher_running
status: running
task: IdleTask (sleeping)
ro_dispatcher
runtime: 325ms
state: dispatcher_running
status: running
task: IdleTask (sleeping)
tap_dispatcher
state: dispatcher_running
status: idle

Let me know if you need any other stats/information.