[
https://issues.apache.org/jira/browse/HDDS-203?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16553385#comment-16553385
]
Tsz Wo Nicholas Sze commented on HDDS-203:
------------------------------------------
Thanks Shash for the new patch. Some thoughts:
- It seems quite expensive to parse and create the entire KeyData object in
order get the block length. How about either (1) storing the size as a field
in KeyData or (2) retrieving the chunk sizes without prasing KeyData object?
It seems that (1) is better although it needs more works.
- Let's have blockID in GetCommittedBlockLengthResponseProto.
> Add getCommittedBlockLength API in datanode
> -------------------------------------------
>
> Key: HDDS-203
> URL: https://issues.apache.org/jira/browse/HDDS-203
> Project: Hadoop Distributed Data Store
> Issue Type: Bug
> Components: Ozone Client, Ozone Datanode
> Reporter: Shashikant Banerjee
> Assignee: Shashikant Banerjee
> Priority: Major
> Fix For: 0.2.1
>
> Attachments: HDDS-203.00.patch, HDDS-203.01.patch, HDDS-203.02.patch,
> HDDS-203.03.patch, HDDS-203.04.patch
>
>
> When a container gets closed on the Datanode while the active Writes are
> happening by OzoneClient, Client Write requests will fail with
> ContainerClosedException. In such case, ozone Client needs to enquire the
> last committed block length from dataNodes and update the OzoneMaster with
> the updated length for the block. This Jira proposes to add to RPC call to
> get the last committed length of a block on a Datanode.
--
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
