[
https://issues.apache.org/jira/browse/HDDS-252?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16556387#comment-16556387
]
Hanisha Koneru commented on HDDS-252:
-------------------------------------
Thanks for working on this [~bharatviswa].
In VolumeSet, we are iterating over all the version files twice. Once to check
dnUuid is consistent and again to intialize the volumes. These two operations
can be combined into one. We can have something similar to
{{VolumSet#checkAndSetClusterID()}} for datanodeUUID also.
If there are no version files, then after initializing volumes, VolumeSet can
assign itself a new UUID. When the volumes are formatted after VersionEndPoint
response from SCM, this new UUID will be persisted on disk.
We do not need to store the {{DatanodeDetails}} in each HddsVolume.
DatanodeUUID would suffice.
> Eliminate the datanode ID file
> ------------------------------
>
> Key: HDDS-252
> URL: https://issues.apache.org/jira/browse/HDDS-252
> Project: Hadoop Distributed Data Store
> Issue Type: Improvement
> Reporter: Bharat Viswanadham
> Assignee: Bharat Viswanadham
> Priority: Major
> Fix For: 0.2.1
>
> Attachments: HDDS-252.00.patch, HDDS-252.01.patch, HDDS-252.02.patch,
> HDDS-252.03.patch, HDDS-252.04.patch
>
>
> This Jira is to remove the datanodeID file. After ContainerIO work (HDDS-48
> branch) is merged, we have a version file in each Volume which stores
> datanodeUuid and some additional fields in that file.
> And also if this disk containing datanodeId path is removed, that DN will now
> be unusable with current code.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
