hadoop-common-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Todd Lipcon (JIRA)" <j...@apache.org>
Subject [jira] Reopened: (HADOOP-5623) Streaming: process provided status messages are overwritten every 10 seoncds
Date Sat, 05 Dec 2009 00:13:20 GMT

     [ https://issues.apache.org/jira/browse/HADOOP-5623?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel

Todd Lipcon reopened HADOOP-5623:

Reopening - this patch needs to be committed to branch-20 as well for 0.20.2. See MAPREDUCE-576.

I applied this patch to branch 20, and then ran the test case provided in that other JIRA
like this:

./bin/hadoop jar build/contrib/streaming/hadoop-0.20.2-dev-streaming.jar -jt local -Dfs.default.name=file:///
-mapper cat -reducer /tmp/test.py -input file:///etc/motd -output file:///tmp/foo

Before the patch, it failed as noted in that JIRA. After the patch, passed.

> Streaming: process provided status messages are overwritten every 10 seoncds
> ----------------------------------------------------------------------------
>                 Key: HADOOP-5623
>                 URL: https://issues.apache.org/jira/browse/HADOOP-5623
>             Project: Hadoop Common
>          Issue Type: Bug
>    Affects Versions: 0.19.0, 0.19.1, 0.20.0
>            Reporter: Rick Cox
>            Assignee: Rick Cox
>             Fix For: 0.20.2, 0.21.0
>         Attachments: HADOOP-5623-streaming-status.patch, HADOOP-5623-streaming-status.patch,
> Every 10 seconds (if the streaming process is producing output key/values on stdout),
PipeMapRed sets the task's status string to "Records R/W=N/N". This replaces any custom task
status that the streaming process may have specified using the "reporter:status:" stderr lines.

This message is automatically generated by JIRA.
You can reply to this email to add a comment to the issue online.

View raw message