[
https://issues.apache.org/jira/browse/HDFS-7411?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14268306#comment-14268306
]
Andrew Wang commented on HDFS-7411:
-----------------------------------
Thanks for the reviews Colin and Ming, new patch up.
* Rebased on trunk
* Updated for new ChunkedArrayList with a working size function
* Updated the logging calls to use new log level helper methods, consolidated
some duplicated logging
* Did the rename and spacing fixes that Colin recommended. Log level change was
necessary since slf4j doesn't have fatal.
* Ming, agree with everything you pointed out. The blocks limit enforcement is
intentionally inexact; I played with iterating based on both DN+block rather
than just DN, but it seemed more complex for a small gain.
As to the limit configuration, I didn't touch it for now until we agree on a
solution. The idea here was to be compatible with the old config option, yet
also provide a way of migrating to the new one. Ming's proposal seems
reasonable, but the override makes configuration more complex. I feel that
computing the limit based on runtime information could also lead to surprises.
> Refactor and improve decommissioning logic into DecommissionManager
> -------------------------------------------------------------------
>
> Key: HDFS-7411
> URL: https://issues.apache.org/jira/browse/HDFS-7411
> Project: Hadoop HDFS
> Issue Type: Improvement
> Affects Versions: 2.5.1
> Reporter: Andrew Wang
> Assignee: Andrew Wang
> Attachments: hdfs-7411.001.patch, hdfs-7411.002.patch,
> hdfs-7411.003.patch, hdfs-7411.004.patch, hdfs-7411.005.patch,
> hdfs-7411.006.patch
>
>
> Would be nice to split out decommission logic from DatanodeManager to
> DecommissionManager.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
