[
https://issues.apache.org/jira/browse/HDFS-12370?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16161098#comment-16161098
]
Yiqun Lin commented on HDFS-12370:
----------------------------------
Thanks for the review and commit, [~cheersyang].I noticed you mentioned one
point that current deletion behaviour is really slow.
bq. I think that was because the throttle logic at KSM or SCM side were
non-optimal
Now we did the throttle logic across many phases. Can we throttle it only in
source place by default? I mean that we just do this in KSM since deletion
operation is invoked fistly in KSM side. Since the source place is throttled, I
think there will also not be much pending deletion blocks that to be deleted
afterward. Users can enable throttle in SCM if they want.
> Ozone: Implement TopN container choosing policy for BlockDeletionService
> ------------------------------------------------------------------------
>
> Key: HDFS-12370
> URL: https://issues.apache.org/jira/browse/HDFS-12370
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Components: ozone
> Affects Versions: HDFS-7240
> Reporter: Yiqun Lin
> Assignee: Yiqun Lin
> Labels: ozoneMerge
> Fix For: HDFS-7240
>
> Attachments: HDFS-12370-HDFS-7240.001.patch,
> HDFS-12370-HDFS-7240.002.patch, HDFS-12370-HDFS-7240.003.patch
>
>
> Implement TopN container choosing policy for BlockDeletionService. This is
> discussed from HDFS-12354.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
