Client not dispatching doc operation requests to active vbucket after memcached got restarted and getting 'RequestTimeoutException' for undispatched docs

Server build: 6.5.0-3821

Client build:

core-io-2.0.0-alpha.5.jar
java-client-3.0.0-alpha.5.jar

Scenario:

1. 4 node cluster, Couchbase bucket (replicas=2)
2. Start loading documents targeting the active vbuckets persent on node-1 (10.112.191.101)
3. Kill Memcached when doc_loading is running in background

Observation:

Memcached on node-1 recovers successfully. But new doc_create requests are not dispatched from clients's end (Verified using doc_req.pcapng).

Seeing 'com.couchbase.client.core.error.RequestTimeoutException' for un-dispatched requests from client end.

Expected behavior:

Doc create request should be resumed once the memcached restarted successfully in cb-server.

Server logs:
https://s3.amazonaws.com/cb-engineering/jira/request_dispatched_post_active_mc_restart/collectinfo-2019-07-22T072137-ns_1%4010.112.191.101.zip
https://s3.amazonaws.com/cb-engineering/jira/request_dispatched_post_active_mc_restart/collectinfo-2019-07-22T072137-ns_1%4010.112.191.102.zip
https://s3.amazonaws.com/cb-engineering/jira/request_dispatched_post_active_mc_restart/collectinfo-2019-07-22T072137-ns_1%4010.112.191.103.zip
https://s3.amazonaws.com/cb-engineering/jira/request_dispatched_post_active_mc_restart/collectinfo-2019-07-22T072137-ns_1%4010.112.191.104.zip
 
TAF testcase:

crash_test.crash_process.CrashTest.test_crash_process,nodes_init=4,replicas=2,num_items=10000,process=memcached,service=memcached,sig_type=sigkill,target_node=active,durability=MAJORITY