Mailing-List: contact hdfs-issues-help@hadoop.apache.org; run by ezmlm
Precedence: bulk
Date: Fri, 29 Jul 2016 23:32:20 +0000 (UTC)
From: "Xiao Chen (JIRA)" <jira@apache.org>
To: hdfs-issues@hadoop.apache.org
Message-ID: <JIRA.12987770.1467980747000.173953.1469835140773@Atlassian.JIRA>
In-Reply-To: <JIRA.12987770.1467980747000@Atlassian.JIRA>
References: <JIRA.12987770.1467980747000@Atlassian.JIRA> <JIRA.12987770.1467980747715@arcas>
Subject: [jira] [Commented] (HDFS-10602) TestBalancer runs timeout
 intermittently
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit
archived-at: Fri, 29 Jul 2016 23:32:23 -0000


    [ https://issues.apache.org/jira/browse/HDFS-10602?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15400229#comment-15400229 ] 

Xiao Chen commented on HDFS-10602:
----------------------------------

Thanks for working on this, [~linyiqun].

From the provided log it looks like the rebalancing is bouncing between the 2 DNs ( 65160 -> 65164, then 65164 -> 65160 in the next iteration). The 3rd DN is excluded for the test case.

I don't fully understand the problem, and why changing the rack of the excluded DN would help in this case. Tried to track down the initial addition of this test case (HDFS-6798), which was trying to verify the fix of HDFS-6364. It seems the code has been refactored so no {{getBlockList}} exists in trunk now, and I would need to further look into this.

In the meantime, [~benoyantony] would you have time to take a look at this and share your thoughts? Thanks in advance.

> TestBalancer runs timeout intermittently
> ----------------------------------------
>
>                 Key: HDFS-10602
>                 URL: https://issues.apache.org/jira/browse/HDFS-10602
>             Project: Hadoop HDFS
>          Issue Type: Bug
>    Affects Versions: 3.0.0-alpha1
>            Reporter: Yiqun Lin
>            Assignee: Yiqun Lin
>         Attachments: HDFS-10602.001.patch
>
>
> As the jira HDFS-10336 has mentioned, the unit test {{TestBalancer#testBalancerWithKeytabs}} will runs too slowly sometimes and that leads the timeout. The test {{TestBalancer#testUnknownDatanodeSimple}}  will also has this problem. These two tests both use the method {{testUnknownDatanode}}. We can do some optimization for this method.


--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

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