You are viewing a plain text version of this content. The canonical link for it is here.
Posted to common-dev@hadoop.apache.org by "Christian Kunz (JIRA)" <ji...@apache.org> on 2008/07/23 00:27:31 UTC
[jira] Created: (HADOOP-3812) Job completion delayed
Job completion delayed
----------------------
Key: HADOOP-3812
URL: https://issues.apache.org/jira/browse/HADOOP-3812
Project: Hadoop Core
Issue Type: Bug
Components: mapred
Affects Versions: 0.17.1
Reporter: Christian Kunz
A job with all reduces completed but with some incomplete re-executed maps is not declared complete till all tasks are complete.
Couldn't the incomplete maps just be killed off?
Even if maps write to dfs directly, these incomplete maps must have been successful before, so there should be no reason to wait for them.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HADOOP-3812) Job completion delayed
Posted by "Christian Kunz (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-3812?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Christian Kunz updated HADOOP-3812:
-----------------------------------
Fix Version/s: 0.20.1
> Job completion delayed
> ----------------------
>
> Key: HADOOP-3812
> URL: https://issues.apache.org/jira/browse/HADOOP-3812
> Project: Hadoop Core
> Issue Type: Bug
> Components: mapred
> Affects Versions: 0.17.1
> Reporter: Christian Kunz
> Fix For: 0.20.1
>
>
> A job with all reduces completed but with some incomplete re-executed maps is not declared complete till all tasks are complete.
> Couldn't the incomplete maps just be killed off?
> Even if maps write to dfs directly, these incomplete maps must have been successful before, so there should be no reason to wait for them.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HADOOP-3812) Job completion delayed
Posted by "Christian Kunz (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-3812?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12720806#action_12720806 ]
Christian Kunz commented on HADOOP-3812:
----------------------------------------
Nobody else running into this issue. In our environment we encounter this quite frequently.
> Job completion delayed
> ----------------------
>
> Key: HADOOP-3812
> URL: https://issues.apache.org/jira/browse/HADOOP-3812
> Project: Hadoop Core
> Issue Type: Bug
> Components: mapred
> Affects Versions: 0.17.1
> Reporter: Christian Kunz
> Fix For: 0.20.1
>
>
> A job with all reduces completed but with some incomplete re-executed maps is not declared complete till all tasks are complete.
> Couldn't the incomplete maps just be killed off?
> Even if maps write to dfs directly, these incomplete maps must have been successful before, so there should be no reason to wait for them.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Issue Comment Edited: (HADOOP-3812) Job completion delayed
Posted by "Christian Kunz (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-3812?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12720806#action_12720806 ]
Christian Kunz edited comment on HADOOP-3812 at 6/17/09 11:15 AM:
------------------------------------------------------------------
Nobody else running into this issue? In our environment we encounter this quite frequently.
was (Author: ckunz):
Nobody else running into this issue. In our environment we encounter this quite frequently.
> Job completion delayed
> ----------------------
>
> Key: HADOOP-3812
> URL: https://issues.apache.org/jira/browse/HADOOP-3812
> Project: Hadoop Core
> Issue Type: Bug
> Components: mapred
> Affects Versions: 0.17.1
> Reporter: Christian Kunz
> Fix For: 0.20.1
>
>
> A job with all reduces completed but with some incomplete re-executed maps is not declared complete till all tasks are complete.
> Couldn't the incomplete maps just be killed off?
> Even if maps write to dfs directly, these incomplete maps must have been successful before, so there should be no reason to wait for them.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.