[
https://issues.apache.org/jira/browse/HDDS-12723?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17941985#comment-17941985
]
Ethan Rose commented on HDDS-12723:
-----------------------------------
I've updates the Jira description to limit the scope of this change to just
total DB failure, which is safer than the original proposal because we will not
remove healthy copies of data that may have already been under-replicated.
> Volume scanner should fail volume if rocksDB is inaccessible
> ------------------------------------------------------------
>
> Key: HDDS-12723
> URL: https://issues.apache.org/jira/browse/HDDS-12723
> Project: Apache Ozone
> Issue Type: Sub-task
> Reporter: Sumit Agrawal
> Assignee: Peter Lee
> Priority: Major
>
> Volume DB is there per volume keeping all containers metadata in that volume.
> If the DB cannot be opened or read at all, we should fail the volume in the
> volume scan. This means none of the containers can be read. If only part of
> the DB is corrupted and some/most keys can be read, the container scanner
> will mark the corresponding containers as unhealthy. We should not drop the
> whole volume in this case because we may unnecessarily remove healthy copies.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@ozone.apache.org
For additional commands, e-mail: issues-h...@ozone.apache.org
