Fix termination of sequential scans of sample profile hash table

When pg_wait_sampling sends a sample profile to caller backend it
initializes a new session of sequential scan for internal hash table.
This session is tracked by internal counter 'num_seq_scans' that is
bound above by some constant thresold. If caller interrupts process of
profile receiving from message queue collector doesn't decrement
internal counter of hash table that cause its threshold crossing.
Thus fix releases a sequencial scan session explicitely even if
abort of interaction process was appeared.

Author: maksm90

collector.c

@@ -266,6 +266,7 @@ send_profile(HTAB *profile_hash, shm_mq_handle *mqh)
 		mq_result = shm_mq_send(mqh, sizeof(ProfileItem), item, false);
 		if (mq_result == SHM_MQ_DETACHED)
 		{
+			hash_seq_term(status);
 			ereport(WARNING,
 					(errmsg("pg_wait_sampling collector: "
 							"receiver of message queue have been detached")));