FAQ
Stipulate main class in a job jar when using 'hadoop jar JARNAME'
-----------------------------------------------------------------

Key: HADOOP-31
URL: http://issues.apache.org/jira/browse/HADOOP-31
Project: Hadoop
Type: New Feature
Reporter: stack@archive.org
Priority: Minor


One use case I forsee is building one jar but using this one jar running multiple jobs: E.g. A single nutch job jar would now be used to do indexing job, later same jar is used to do dedup, etc. Currently, the recently added hadoop 'jar' option just takes the jar name then looks in the jar MANIFEST.MF for the Main-Class, failing if not present. This is grand but for the scenario above, it means I have to create a jar per job I want to run -- each with a different MANIFEST.MF#Main-Class entry.

Can we pass the Main-Class on the hadoop command-line as an (optional) argument to 'hadoop jar JAR_NAME'? (I can make a patch if wanted).

--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
http://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see:
http://www.atlassian.com/software/jira

Search Discussions

  • Owen O'Malley (JIRA) at Feb 10, 2006 at 1:04 am
    [ http://issues.apache.org/jira/browse/HADOOP-31?page=comments#action_12365807 ]

    Owen O'Malley commented on HADOOP-31:
    -------------------------------------

    I would argue that rather than creating more framework, you just use the arguments and dispatch on the first argument. If you need something generic you could do something like I did with the ExampleDriver.
    Stipulate main class in a job jar when using 'hadoop jar JARNAME'
    -----------------------------------------------------------------

    Key: HADOOP-31
    URL: http://issues.apache.org/jira/browse/HADOOP-31
    Project: Hadoop
    Type: New Feature
    Reporter: stack@archive.org
    Priority: Minor
    One use case I forsee is building one jar but using this one jar running multiple jobs: E.g. A single nutch job jar would now be used to do indexing job, later same jar is used to do dedup, etc. Currently, the recently added hadoop 'jar' option just takes the jar name then looks in the jar MANIFEST.MF for the Main-Class, failing if not present. This is grand but for the scenario above, it means I have to create a jar per job I want to run -- each with a different MANIFEST.MF#Main-Class entry.
    Can we pass the Main-Class on the hadoop command-line as an (optional) argument to 'hadoop jar JAR_NAME'? (I can make a patch if wanted).
    --
    This message is automatically generated by JIRA.
    -
    If you think it was sent incorrectly contact one of the administrators:
    http://issues.apache.org/jira/secure/Administrators.jspa
    -
    For more information on JIRA, see:
    http://www.atlassian.com/software/jira
  • stack@archive.org (JIRA) at Feb 10, 2006 at 1:15 am
    [ http://issues.apache.org/jira/browse/HADOOP-31?page=comments#action_12365810 ]

    stack@archive.org commented on HADOOP-31:
    -----------------------------------------

    That will work. Thanks.

    (I'd close this issue but I don't have sufficent permissions).
    Stipulate main class in a job jar when using 'hadoop jar JARNAME'
    -----------------------------------------------------------------

    Key: HADOOP-31
    URL: http://issues.apache.org/jira/browse/HADOOP-31
    Project: Hadoop
    Type: New Feature
    Reporter: stack@archive.org
    Priority: Minor
    One use case I forsee is building one jar but using this one jar running multiple jobs: E.g. A single nutch job jar would now be used to do indexing job, later same jar is used to do dedup, etc. Currently, the recently added hadoop 'jar' option just takes the jar name then looks in the jar MANIFEST.MF for the Main-Class, failing if not present. This is grand but for the scenario above, it means I have to create a jar per job I want to run -- each with a different MANIFEST.MF#Main-Class entry.
    Can we pass the Main-Class on the hadoop command-line as an (optional) argument to 'hadoop jar JAR_NAME'? (I can make a patch if wanted).
    --
    This message is automatically generated by JIRA.
    -
    If you think it was sent incorrectly contact one of the administrators:
    http://issues.apache.org/jira/secure/Administrators.jspa
    -
    For more information on JIRA, see:
    http://www.atlassian.com/software/jira
  • Doug Cutting (JIRA) at Feb 10, 2006 at 4:32 pm
    [ http://issues.apache.org/jira/browse/HADOOP-31?page=all ]

    Doug Cutting resolved HADOOP-31:
    --------------------------------

    Resolution: Won't Fix

    Closing at reporter's request.
    Stipulate main class in a job jar when using 'hadoop jar JARNAME'
    -----------------------------------------------------------------

    Key: HADOOP-31
    URL: http://issues.apache.org/jira/browse/HADOOP-31
    Project: Hadoop
    Type: New Feature
    Reporter: stack@archive.org
    Priority: Minor
    One use case I forsee is building one jar but using this one jar running multiple jobs: E.g. A single nutch job jar would now be used to do indexing job, later same jar is used to do dedup, etc. Currently, the recently added hadoop 'jar' option just takes the jar name then looks in the jar MANIFEST.MF for the Main-Class, failing if not present. This is grand but for the scenario above, it means I have to create a jar per job I want to run -- each with a different MANIFEST.MF#Main-Class entry.
    Can we pass the Main-Class on the hadoop command-line as an (optional) argument to 'hadoop jar JAR_NAME'? (I can make a patch if wanted).
    --
    This message is automatically generated by JIRA.
    -
    If you think it was sent incorrectly contact one of the administrators:
    http://issues.apache.org/jira/secure/Administrators.jspa
    -
    For more information on JIRA, see:
    http://www.atlassian.com/software/jira
  • Gordon Mohr (JIRA) at Feb 10, 2006 at 6:07 pm
    [ http://issues.apache.org/jira/browse/HADOOP-31?page=comments#action_12365935 ]

    Gordon Mohr commented on HADOOP-31:
    -----------------------------------

    Perhaps ExampleDriver could itself be renamed/moved to more general 'Launcher' that's documented as a framework convenience class for Hadoop users who have a single-Jar, many main()s usage?
    Stipulate main class in a job jar when using 'hadoop jar JARNAME'
    -----------------------------------------------------------------

    Key: HADOOP-31
    URL: http://issues.apache.org/jira/browse/HADOOP-31
    Project: Hadoop
    Type: New Feature
    Reporter: stack@archive.org
    Priority: Minor
    One use case I forsee is building one jar but using this one jar running multiple jobs: E.g. A single nutch job jar would now be used to do indexing job, later same jar is used to do dedup, etc. Currently, the recently added hadoop 'jar' option just takes the jar name then looks in the jar MANIFEST.MF for the Main-Class, failing if not present. This is grand but for the scenario above, it means I have to create a jar per job I want to run -- each with a different MANIFEST.MF#Main-Class entry.
    Can we pass the Main-Class on the hadoop command-line as an (optional) argument to 'hadoop jar JAR_NAME'? (I can make a patch if wanted).
    --
    This message is automatically generated by JIRA.
    -
    If you think it was sent incorrectly contact one of the administrators:
    http://issues.apache.org/jira/secure/Administrators.jspa
    -
    For more information on JIRA, see:
    http://www.atlassian.com/software/jira

Related Discussions

Discussion Navigation
viewthread | post
Discussion Overview
groupcommon-dev @
categorieshadoop
postedFeb 10, '06 at 12:17a
activeFeb 10, '06 at 6:07p
posts5
users1
websitehadoop.apache.org...
irc#hadoop

1 user in discussion

Gordon Mohr (JIRA): 5 posts

People

Translate

site design / logo © 2022 Grokbase