hive-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sankar Hariappan (JIRA)" <j...@apache.org>
Subject [jira] [Updated] (HIVE-16901) Distcp optimization - One distcp per CopyTask
Date Wed, 14 Jun 2017 18:51:00 GMT

     [ https://issues.apache.org/jira/browse/HIVE-16901?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Sankar Hariappan updated HIVE-16901:
------------------------------------
    Description: 
Currently, if a CopyTask is created to copy a list of files, then distcp is invoked for each
and every file. Instead, need to pass the list of source files to be copied to distcp tool
which basically copies the files in parallel and hence gets lot of performance gain.

If the copy of list of files fail, then traverse the destination directory to see which file
is missing and checksum mismatches, then trigger copy of those files one by one.

  was:
Currently, if a ReplCopyTask is created to copy a list of files, then distcp is invoked for
each and every file. Instead, need to pass the list of source files to be copied to distcp
tool which basically copies the files in parallel and hence gets lot of performance gain.

If the copy of list of files fail, then traverse the destination directory to see which file
is missing and checksum mismatches, then trigger copy of those files one by one.


> Distcp optimization - One distcp per CopyTask 
> ----------------------------------------------
>
>                 Key: HIVE-16901
>                 URL: https://issues.apache.org/jira/browse/HIVE-16901
>             Project: Hive
>          Issue Type: Sub-task
>          Components: Hive, repl
>    Affects Versions: 2.1.0
>            Reporter: Sankar Hariappan
>            Assignee: Sankar Hariappan
>              Labels: DR, replication
>             Fix For: 3.0.0
>
>
> Currently, if a CopyTask is created to copy a list of files, then distcp is invoked for
each and every file. Instead, need to pass the list of source files to be copied to distcp
tool which basically copies the files in parallel and hence gets lot of performance gain.
> If the copy of list of files fail, then traverse the destination directory to see which
file is missing and checksum mismatches, then trigger copy of those files one by one.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

Mime
View raw message