[
https://issues.apache.org/jira/browse/HDDS-17101?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
ASF GitHub Bot updated HDDS-17101:
----------------------------------
Labels: pull-request-available (was: )
> Keep QUASI_CLOSED replicas of a CLOSED container on their origin datanode
> -------------------------------------------------------------------------
>
> Key: HDDS-17101
> URL: https://issues.apache.org/jira/browse/HDDS-17101
> Project: Apache Ozone
> Issue Type: Improvement
> Components: SCM
> Reporter: Siyao Meng
> Assignee: Siyao Meng
> Priority: Major
> Labels: pull-request-available
>
> When a QUASI_CLOSED Ratis container is force closed, its sequence ID is the
> highest one reported by any replica. After that, ReplicationManager counts
> every QUASI_CLOSED replica with a different sequence ID as unhealthy and
> deletes it once the container has enough CLOSED replicas. So which replicas
> survive depends on the report of a single replica. Origin IDs cannot tell the
> replicas apart either, because a copy made by replication reports the origin
> of the replica it was copied from. Separately, the container report handler
> still raises the sequence ID of a CLOSED container when a replica reports a
> higher one (it only logs an error).
> This change:
> * Keeps a QUASI_CLOSED replica of a CLOSED container while it is on its
> origin datanode, whatever its sequence ID. ReplicationManager does not select
> it for deletion and leaves it out of the replica counts, the placement check
> and the health state, so it does not cause healthy replicas to be removed or
> added and the container stays HEALTHY. Its datanode is excluded as a target
> for new replicas. An empty kept replica does not stop an empty container from
> being deleted. Copies on other datanodes are handled as before.
> * Stops the container report handler from raising the sequence ID of a CLOSED
> container.
> *Compatibility*
> * No protocol, layout or configuration change.
> * After a force close, replicas that lag behind stay on their origin
> datanodes instead of being deleted, so such a container can keep up to two
> extra replicas until it is deleted.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
