Read request times out(no response from server) when memcached is stopped on replica and durable crud is issued

1. Create a 2 node cluster
2. Create a default bucket with 1 replica
3. Create a doc
4. Stop memcahced on replica
5. Delete Doc. Observe that it is deleted from the bucket in UI. Additionally, read through Java SDK threw "RequestTimeoutException"
6. Resume memcached on replica node and observe that the document is not deleted from replica.
7. Delete Bucket - It went to Orphan state and never comes out of it
8. Rebalance out node and rebalance failed.

Rebalance exited with reason {buckets_shutdown_wait_failed,

[{'ns_1@10.112.180.103',

{'EXIT',

{old_buckets_shutdown_wait_failed,

["default"]}}}]}