[ 
https://issues.apache.org/jira/browse/HDFS-14748?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16914931#comment-16914931
 ] 

Lisheng Sun commented on HDFS-14748:
------------------------------------

I confirmed the two UT are ok in my local. So the UT failures are unrelated to 
this patch.

> Make DataNodePeerMetrics#minOutlierDetectionSamples configurable
> ----------------------------------------------------------------
>
>                 Key: HDFS-14748
>                 URL: https://issues.apache.org/jira/browse/HDFS-14748
>             Project: Hadoop HDFS
>          Issue Type: Improvement
>    Affects Versions: 2.9.0, 3.0.0-alpha4
>            Reporter: Lisheng Sun
>            Assignee: Lisheng Sun
>            Priority: Major
>         Attachments: HDFS-14748.001.patch, HDFS-14748.002.patch, 
> HDFS-14748.003.patch
>
>
> Slow node monitoring is to transfer 1000 packets between DataNodes within 
> three hours before they are eligible to calculate and upload transmission 
> delays to the namenode.
> But if Write data is very small and number of packets is less than 1000, the 
> slow node will not be reported to NameNode, so make 
> DataNodePeerMetrics#minOutlierDetectionSamplesconfigurable.



--
This message was sent by Atlassian Jira
(v8.3.2#803003)

---------------------------------------------------------------------
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org

Reply via email to