[
https://issues.apache.org/jira/browse/HDDS-9244?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17762508#comment-17762508
]
Ethan Rose commented on HDDS-9244:
----------------------------------
One issue is that datanodes do not persist information on whether a replica is
EC or Ratis, although we may be able to deduce this from other container
metadata. This info would determine whether the scanner marks the containers as
quasi-closed or closed on recovery.
> Allow container scanner to restore container health if health check passes
> --------------------------------------------------------------------------
>
> Key: HDDS-9244
> URL: https://issues.apache.org/jira/browse/HDDS-9244
> Project: Apache Ozone
> Issue Type: Sub-task
> Reporter: Ethan Rose
> Priority: Major
>
> Sometimes a container may be marked unhealthy on the write path due to
> pipeline errors that end up not affecting the integrity of its data. In this
> case, we could have unhealthy containers with all data intact that pass the
> scan. If so, we could allow the scanner to move these back to closed state
> (or quasi-closed for Ratis, which may be closed by SCM later). This replica
> info would be reported back to SCM.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
