hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jing Zhao (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HDFS-5946) Webhdfs DN choosing code is flawed
Date Thu, 13 Feb 2014 18:49:21 GMT

    [ https://issues.apache.org/jira/browse/HDFS-5946?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13900592#comment-13900592

Jing Zhao commented on HDFS-5946:

I guess the DataNode list here is sorted? Thus all the decommissioned DNs are actually in
the end of the list. So the logic that throws an exception when the first DN is decommissioned
should be correct?

The original code before HDFS-5891 does not do random pick for all the cases. The random pick
is only for web UI. Thus I guess we do not need to worry about the scenario where too much
traffic is to be directed to the same DN. WebHdfs etc. always tries to use the first DN, and
this is consistent with the sorted DN list logic.

> Webhdfs DN choosing code is flawed
> ----------------------------------
>                 Key: HDFS-5946
>                 URL: https://issues.apache.org/jira/browse/HDFS-5946
>             Project: Hadoop HDFS
>          Issue Type: Bug
>          Components: namenode, webhdfs
>    Affects Versions: 3.0.0, 2.4.0
>            Reporter: Daryn Sharp
>            Priority: Critical
> HDFS-5891 improved the performance of redirecting webhdfs clients to a DN.  Instead of
attempting a connection with a 1-minute timeout, the NN skips decommissioned nodes.
> The logic appears flawed.  It finds the index of the first decommissioned node, if any,
> * Throws an exception if index = 0, even if other nodes later in the list are not decommissioned.
> * Else picks a random node prior to the index.  Let's say there are 10 replicas, 2nd
location is decommissioned.  All clients will be redirected to the first location even though
there are 8 other valid locations.

This message was sent by Atlassian JIRA

View raw message