[
https://issues.apache.org/jira/browse/HDDS-7785?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17679554#comment-17679554
]
Siddhant Sangwan commented on HDDS-7785:
----------------------------------------
I'm actually not satisfied with these terms: healthy, mismatched, and
unhealthy. For someone who doesn't know their current definitions, it's easy to
assume unhealthy replicas are a subset of mismatched replicas. Another
assumption could be that mismatched replicas are healthy because they're not
UNHEALTHY (corruption etc).
We should either define them everywhere or just let the code define them.
Some other terms I was thinking of:
Nonreplicable/Unreplicable instead of mismatched: Since OPEN or CLOSING
replicas should not be replicated.
> Improve Handling of Unhealthy Container Replicas in the new RM
> --------------------------------------------------------------
>
> Key: HDDS-7785
> URL: https://issues.apache.org/jira/browse/HDDS-7785
> Project: Apache Ozone
> Issue Type: Sub-task
> Components: SCM
> Affects Versions: 1.3.0
> Reporter: Siddhant Sangwan
> Assignee: Siddhant Sangwan
> Priority: Major
> Fix For: 1.4.0
>
>
> HDDS-6447 changed the handling of unhealthy replicas (replicas whose state
> doesn't match with the container state) in the legacy RM. This jira adds
> these changes to the new RM.
> Should we call "replicas whose state doesn't match with the container state"
> {{mismatched}} replicas and leave the {{unhealthy}} term to replicas whose
> ContainerReplicaProto.State is UNHEALTHY (meaning corrupted, checksum
> mismatch etc)?
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
