[jira] [Created] (HDDS-6848) Datanode's replication queue fulled of timeout replication task

["Janus Chow (Jira)" <ji...@apache.org>]

Janus Chow created HDDS-6848:
--------------------------------

             Summary: Datanode's replication queue fulled of timeout replication task
                 Key: HDDS-6848
                 URL: https://issues.apache.org/jira/browse/HDDS-6848
             Project: Apache Ozone
          Issue Type: Improvement
            Reporter: Janus Chow
            Assignee: Janus Chow


If a datanode crashes, the blocks on the datanode will be added to inflightReplication to fulfill the replication factor. 

When the datanode is restarted, the containers don't need to be replicated anymore, but the replication commands have been sent to datanodes. If the crashed datanodes holds many containers, there will be many replication commands which will cause jams on Datanodes's replication queue, and the big replication traffic will also cause problems to the whole cluster.



--
This message was sent by Atlassian Jira
(v8.20.7#820007)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@ozone.apache.org
For additional commands, e-mail: issues-help@ozone.apache.org