[
https://issues.apache.org/jira/browse/OOZIE-2978?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16080477#comment-16080477
]
Peter Bacsko commented on OOZIE-2978:
-------------------------------------
Are we sure we don't need this part?
{code}
- // For embedded python or for version of pig lower than 0.8, pig stats
are not supported.
- // So retrieving hadoop Ids here
- File file = new File(System.getProperty(EXTERNAL_CHILD_IDS));
- if (!file.exists()) {
- writeExternalChildIDs(logFile, PIG_JOB_IDS_PATTERNS, "Pig");
- }
{code}
I'm concerned about the "embedded python" part.
> Remove code that handles Pig versions before 0.8
> -------------------------------------------------
>
> Key: OOZIE-2978
> URL: https://issues.apache.org/jira/browse/OOZIE-2978
> Project: Oozie
> Issue Type: Bug
> Reporter: Attila Sasvari
> Assignee: Attila Sasvari
> Attachments: OOZIE-2978-01.patch, OOZIE-2978-02.patch,
> OOZIE-2978-03.patch, OOZIE-2978-04.patch
>
>
> Oozie submits Pig jobs according to given Pig version. For example, Pig
> versions before 0.8 are treated in the following way:
>
> [Main.main(args)|https://github.com/apache/oozie/blob/678f1b2b2cc9a2456c7e3d9f693d2637f1f03257/sharelib/pig/src/main/java/org/apache/oozie/action/hadoop/PigMain.java#L382]
> However, we dropped support for *Hadoop 1*, *Hadoop 0.23* support via
> {{OOZIE-2316}}. I wonder whether submitting jobs using Pig version before 0.8
> will work at all:
> -
> https://pig.apache.org/releases.html#13+May%2C+2010%3A+release+0.7.0+available
> says "This release works with Hadoop 0.20". Pig *0.10.1* is the first pig
> release that works with Hadoop 2.
> - Looking at [PIG-2125|https://issues.apache.org/jira/browse/PIG-2125], I
> believe the answer is NO.
> Therefore we can safely delete corresponding parts from the codebase. There
> is also another class {{PigMainWithOldAPI}} that is only used by tests in
> {{TestPigMainWithOldAPI}}. So we can remove these too.
>
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
