Client disconnect leads to stucked clusterExecutor shutdown (3.6.2)

[uladzimirFilipchanka]

I have two apps deployed on different VM's. First app connects to another as Hazelcast Client.
As soon second app goes down, client starts reconnection as expeced. Finally the client says

"Could not re-connect to cluster shutting down the client"

and goes to ClusterListenerSupport:233 to shutdown LifecycleService.

After some components shutdowns, called ClusterListenerSupport.shutdown
And the component tries to shutdown clusterExecutor and stucked for TERMINATE_TIMEOUT_SECONDS because the task is launched inside this clusterExecutor.

Here is screen from debug view

Log says:

13:34:41,955 INFO com.hazelcast.core.LifecycleService HazelcastClient[hz.client_50_ustaxi-shift][3.6.2] is SHUTTING_DOWN
13:35:11,957 WARNING com.hazelcast.client.spi.impl.ClusterListenerSupport cluster executor awaitTermination could not completed in 30 seconds
13:35:11,960 INFO com.hazelcast.core.LifecycleService HazelcastClient[hz.client_50_ustaxi-shift][3.6.2] is SHUTDOWN

This issue reproduces permanently and only after target instance goes down.