Mailing-List: contact mapreduce-issues-help@hadoop.apache.org; run by ezmlm
Precedence: bulk
Reply-To: mapreduce-issues@hadoop.apache.org
Date: Wed, 15 Apr 2015 05:52:59 +0000 (UTC)
From: "Peng Zhang (JIRA)" <jira@apache.org>
To: mapreduce-issues@hadoop.apache.org
Message-ID: <JIRA.12786594.1427689592000.4800.1429077179411@Atlassian.JIRA>
In-Reply-To: <JIRA.12786594.1427689592000@Atlassian.JIRA>
References: <JIRA.12786594.1427689592000@Atlassian.JIRA>
 <JIRA.12786594.1427689592782@arcas>
Subject: [jira] [Commented] (MAPREDUCE-6302) deadlock in a job between map
 and reduce cores allocation
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable


    [ https://issues.apache.org/jira/browse/MAPREDUCE-6302?page=3Dcom.atlas=
sian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=3D=
14495699#comment-14495699 ]=20

Peng Zhang commented on MAPREDUCE-6302:
---------------------------------------

in head log it shows=20
Number of reduces for job job_1427451307913_0012 =3D 380

in tail log it shows=20
headroom=3D<memory:333824, vCores:661> and totalResourceLimit:<memory:66764=
8, vCores:824>

It seems conflict with statement =E2=80=9Cas long as reduces of a job is mo=
re than max cores of the queue , deadlock always happened. =E2=80=9D

> deadlock in a job between map and reduce cores allocation=20
> ----------------------------------------------------------
>
>                 Key: MAPREDUCE-6302
>                 URL: https://issues.apache.org/jira/browse/MAPREDUCE-6302
>             Project: Hadoop Map/Reduce
>          Issue Type: Bug
>    Affects Versions: 2.6.0
>            Reporter: mai shurong
>            Assignee: Karthik Kambatla
>            Priority: Critical
>         Attachments: AM_log_head100000.txt.gz, AM_log_tail100000.txt.gz, =
queue_with_max163cores.png, queue_with_max263cores.png, queue_with_max333co=
res.png
>
>
> I submit a  big job, which has 500 maps and 350 reduce, to a queue(fairsc=
heduler) with 300 max cores. When the big mapreduce job is running 100% map=
s, the 300 reduces have occupied 300 max cores in the queue. And then, a ma=
p fails and retry, waiting for a core, while the 300 reduces are waiting fo=
r failed map to finish. So a deadlock occur. As a result, the job is blocke=
d, and the later job in the queue cannot run because no available cores in =
the queue.
> I think there is the similar issue for memory of a queue .


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