hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Yongjun Zhang (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HDFS-8828) Utilize Snapshot diff report to build copy list in distcp
Date Thu, 13 Aug 2015 21:46:45 GMT

    [ https://issues.apache.org/jira/browse/HDFS-8828?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14696000#comment-14696000
] 

Yongjun Zhang commented on HDFS-8828:
-------------------------------------

Hi [~yufeigu],

Thanks for answering my question in person. So for newly created dir, there is indeed one
entry "CREATE" in the snapshot diff report, and no entries for new elements created below
this dir.

So please take care of my comment 1, 2 in my previous review, plus: 

3.  Suggest to change the {{getExcludeList}} method to {{getTraverseExcludeList}}  (hopefully
a better name) and with the following javadoc as we agreed.
{code}
This method returns a list of items to be excluded when recursively traversing newDir to build
the copy list.

Specifically, given a newly created directory newDir (a CREATE entry in the snapshot diff),
if a previously copied file/directory itemX is moved (a RENAME entry in the snapshot diff)
into newDir, itemX should be excluded when recursively traversing newDir in #traverseDirectory,
 so that it will not to be copied again.

If the same itemX also has a MODIFY entry in the snapshot diff report, meaning it was modified
after it was previously copied, it will still be added to the copy list (handled in the main
loop of doBuildListingWithSnapshotDiff).
{code}

4. Do refactoring to consolidate duplicated code in test code that we discussed.
 
Hi [~jingzhao], I had quite some side discussion with Yufei, I am +1 on the change after the
above comments are addressed. Would you please take a look at it if you wish? I'm targeting
at committing it next Monday.

Thanks much.



> Utilize Snapshot diff report to build copy list in distcp
> ---------------------------------------------------------
>
>                 Key: HDFS-8828
>                 URL: https://issues.apache.org/jira/browse/HDFS-8828
>             Project: Hadoop HDFS
>          Issue Type: Improvement
>          Components: distcp, snapshots
>            Reporter: Yufei Gu
>            Assignee: Yufei Gu
>         Attachments: HDFS-8828.001.patch, HDFS-8828.002.patch, HDFS-8828.003.patch, HDFS-8828.004.patch,
HDFS-8828.005.patch, HDFS-8828.006.patch
>
>
> Some users reported huge time cost to build file copy list in distcp. (30 hours for 1.6M
files). We can leverage snapshot diff report to build file copy list including files/dirs
which are changes only between two snapshots (or a snapshot and a normal dir). It speed up
the process in two folds: 1. less copy list building time. 2. less file copy MR jobs.
> HDFS snapshot diff report provide information about file/directory creation, deletion,
rename and modification between two snapshots or a snapshot and a normal directory. HDFS-7535
synchronize deletion and rename, then fallback to the default distcp. So it still relies on
default distcp to building complete list of files under the source dir. This patch only puts
creation and modification files into the copy list based on snapshot diff report. We can minimize
the number of files to copy. 



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

Mime
View raw message