[
https://issues.apache.org/jira/browse/HDFS-15549?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17261673#comment-17261673
]
Leon Gao commented on HDFS-15549:
---------------------------------
I have put a draft patch with some tests.
I tried to use hardlink to ensure durability when renaming the block files.
Directly using filesystem "rename" can cause data loss when only block/metadata
file is moved during datanode restart.
[~hexiaoqiao] [~jingzhao] [~jianghuazhu] Could you please take a look and let
me know your thoughts.
> Improve DISK/ARCHIVE movement if they are on same filesystem
> ------------------------------------------------------------
>
> Key: HDFS-15549
> URL: https://issues.apache.org/jira/browse/HDFS-15549
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Components: datanode
> Reporter: Leon Gao
> Assignee: Leon Gao
> Priority: Major
> Labels: pull-request-available
> Time Spent: 50m
> Remaining Estimate: 0h
>
> When moving blocks between DISK/ARCHIVE, we should prefer the volume on the
> same underlying filesystem and use "rename" instead of "copy" to save IO.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
