Mailing-List: contact mapreduce-issues-help@hadoop.apache.org; run by ezmlm
Precedence: bulk
Date: Thu, 7 Jul 2016 13:22:11 +0000 (UTC)
From: "mingleizhang (JIRA)" <jira@apache.org>
To: mapreduce-issues@hadoop.apache.org
Message-ID: <JIRA.12987095.1467796519000.47930.1467897731032@Atlassian.JIRA>
In-Reply-To: <JIRA.12987095.1467796519000@Atlassian.JIRA>
References: <JIRA.12987095.1467796519000@Atlassian.JIRA> <JIRA.12987095.1467796519973@arcas>
Subject: [jira] [Issue Comment Deleted] (MAPREDUCE-6729) Hitting performance
 and error when lots of files to write or read
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
archived-at: Thu, 07 Jul 2016 13:22:18 -0000


     [ https://issues.apache.org/jira/browse/MAPREDUCE-6729?page=3Dcom.atla=
ssian.jira.plugin.system.issuetabpanels:all-tabpanel ]

mingleizhang updated MAPREDUCE-6729:
------------------------------------
    Comment: was deleted

(was: Thanks Kai Zheng for review this code.)

> Hitting performance and error when lots of files to write or read
> -----------------------------------------------------------------
>
>                 Key: MAPREDUCE-6729
>                 URL: https://issues.apache.org/jira/browse/MAPREDUCE-6729
>             Project: Hadoop Map/Reduce
>          Issue Type: Improvement
>          Components: benchmarks, performance, test
>            Reporter: mingleizhang
>            Assignee: mingleizhang
>            Priority: Minor
>              Labels: performance, test
>         Attachments: MR-6729.txt
>
>
> When doing DFSIO test as a distributed i/o benchmark tool. Then especiall=
y writes plenty of files to disk or read from, both can cause performance i=
ssue and imprecise value in a way. The question is that existing practices =
needs to delete files when before running a job and that will cause extra t=
ime consumption and furthermore cause performance issue, statistical time e=
rror and imprecise throughput while the files are lots of. So we need to re=
place or improve this hack to prevent this from happening in the future.
> {code}
> public static void testWrite() throws Exception {
>     FileSystem fs =3D cluster.getFileSystem();
>     long tStart =3D System.currentTimeMillis();
>     bench.writeTest(fs); // this line of code will cause extra time consu=
mption because of fs.delete(*,*) by the writeTest method
>     long execTime =3D System.currentTimeMillis() - tStart;
>     bench.analyzeResult(fs, TestType.TEST_TYPE_WRITE, execTime);
>   }
> private void writeTest(FileSystem fs) throws IOException {
>   Path writeDir =3D getWriteDir(config);
>   fs.delete(getDataDir(config), true);
>   fs.delete(writeDir, true);   =20
>   runIOTest(WriteMapper.class, writeDir);
>   }
> {code}=E3=80=80
> [https://github.com/apache/hadoop/blob/trunk/hadoop-mapreduce-project/had=
oop-mapreduce-client/hadoop-mapreduce-client-jobclient/src/test/java/org/ap=
ache/hadoop/fs/TestDFSIO.java]


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

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