[
https://issues.apache.org/jira/browse/HDDS-1016?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16755136#comment-16755136
]
Yiqun Lin commented on HDDS-1016:
---------------------------------
The patch almost looks good to me. Besides the failed unit test, some minor
comments from me, [~arpitagarwal]:
*KeyValueContainer.java*
Line30: Unused import.
*KeyValueHandler.java*
Line588: {{actualState}} is not used.
*TestKeyValueHandlerWithUnhealthyContainer.java*
Some unsed imports.
{quote}Add check for unhealthy container to KeyValueHandler read operations
{quote}
I have some questions for this, actually here we also do the unhealthy check
for a write request ({{DeleteChunk)}} not all read operations as mentioned. Why
not unhealthy check make sense for all requests? And what's the meaning of one
unhealthy container? Does that mean we find one volume is failed, then all the
containers inside are unhealthy state?
> Allow marking containers as unhealthy
> -------------------------------------
>
> Key: HDDS-1016
> URL: https://issues.apache.org/jira/browse/HDDS-1016
> Project: Hadoop Distributed Data Store
> Issue Type: Improvement
> Components: Ozone Datanode
> Reporter: Arpit Agarwal
> Assignee: Arpit Agarwal
> Priority: Major
> Attachments: HDDS-1016.01.patch
>
>
> Containers support an unhealthy state but currently the Container interface
> on the DataNodes does not expose a way to mark containers as unhealthy.
> -We can also make a few locking improvements to the KeyValueContainer class.-
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
