[
https://issues.apache.org/jira/browse/HDDS-2850?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Aravindan Vijayan updated HDDS-2850:
------------------------------------
Summary: Handle Delete container use case in Recon. (was: Handle
Create/Delete container use case in Recon.)
> Handle Delete container use case in Recon.
> ------------------------------------------
>
> Key: HDDS-2850
> URL: https://issues.apache.org/jira/browse/HDDS-2850
> Project: Hadoop Distributed Data Store
> Issue Type: Sub-task
> Components: Ozone Recon
> Reporter: Aravindan Vijayan
> Assignee: Aravindan Vijayan
> Priority: Major
>
> CREATE/DELETE container needs to be handled differently in Recon since these
> actions are initiated in the SCM, and Recon does not know about that.
> It should not throw ContainerNotFoundException when it suddenly sees a new
> container. Similarly, it may mistakenly flag deleted containers as missing
> containers.
> A container is created in the Datanode through an SCM command, and is deleted
> in the Datanode when the SCM Replication Manager finds that it is over
> replicated. Currently, the Datanode does not send any ACK to the SCM when it
> has created/deleted a container. We can add an ACK command to the
> Create/Delete Container processing in the Datanode. The SCM will not consume
> the ACK command, but Recon can use it to update its internal container state.
> The same logic can be added for Pipeline Create as well. Pipeline destroy is
> inferred in SCM through node staleness. Hence, recon already takes care of
> that through its handlers.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
