[jira] Resolved: (HADOOP-1052) Map tasks not getting killed

["Srikanth Kakani (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/HADOOP-1052?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Srikanth Kakani resolved HADOOP-1052.
-------------------------------------

    Resolution: Cannot Reproduce

In release 0.12.2 I have seen the long standing map jobs getting killed consistently when they exceed the time.




> Map tasks not getting killed
> ----------------------------
>
>                 Key: HADOOP-1052
>                 URL: https://issues.apache.org/jira/browse/HADOOP-1052
>             Project: Hadoop
>          Issue Type: Bug
>          Components: mapred
>    Affects Versions: 0.11.2
>         Environment: Redhat, 8 node cluster
>            Reporter: Srikanth Kakani
>
> -----------Steps to reproduce---------
> 1. Create and compile a map reduce application with 
> public void map(WritableComparable key,
>                     Writable value,
>                     OutputCollector output, Reporter reporter) throws IOException {
>      reporter.setStatus("Started");
>     Thread.sleep(3600000);
>     report.setStatus("Should have been killed"); 
> }
> 2. Run the task with the default hadoop configuration: 
> <property>
>   <name>mapred.task.timeout</name>
>   <value>1200000</value>
>   <description>The number of milliseconds before a task will be
>   terminated if it neither reads an input, writes an output, nor
>   updates its status string.
>   </description>
> </property>
> ----------------Result------------------
> The task does not get killed, and the job is successful
> ----------------Expected Result------------------
> The map tasks get killed and the job should fail.
> -Srikanth
>   

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