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 "Arun C Murthy (JIRA)" <ji...@apache.org> on 2007/05/03 01:34:16 UTC
[jira] Created: (HADOOP-1322) Tasktracker blacklist leads to hung
jobs in single-node cluster
Tasktracker blacklist leads to hung jobs in single-node cluster
---------------------------------------------------------------
Key: HADOOP-1322
URL: https://issues.apache.org/jira/browse/HADOOP-1322
Project: Hadoop
Issue Type: Bug
Components: mapred
Reporter: Arun C Murthy
Assigned To: Arun C Murthy
Fix For: 0.13.0
Post HADOOP-1278, adding _the_ tasktracker to the blacklist in single-node clusters leads to a situation where the entire cluster is 'flaky' and no trackers are available to execute tasks.
The straight-forward fix to check before adding the tracker to the blacklist.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HADOOP-1322) Tasktracker blacklist leads to hung
jobs in single-node cluster
Posted by "Arun C Murthy (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-1322?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Arun C Murthy updated HADOOP-1322:
----------------------------------
Status: Patch Available (was: Open)
> Tasktracker blacklist leads to hung jobs in single-node cluster
> ---------------------------------------------------------------
>
> Key: HADOOP-1322
> URL: https://issues.apache.org/jira/browse/HADOOP-1322
> Project: Hadoop
> Issue Type: Bug
> Components: mapred
> Reporter: Arun C Murthy
> Assigned To: Arun C Murthy
> Fix For: 0.13.0
>
> Attachments: HADOOP-1322_20070503_1.patch
>
>
> Post HADOOP-1278, adding _the_ tasktracker to the blacklist in single-node clusters leads to a situation where the entire cluster is 'flaky' and no trackers are available to execute tasks.
> The straight-forward fix to check before adding the tracker to the blacklist.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HADOOP-1322) Tasktracker blacklist leads to hung
jobs in single-node cluster
Posted by "Hadoop QA (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-1322?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12493258 ]
Hadoop QA commented on HADOOP-1322:
-----------------------------------
+1
http://issues.apache.org/jira/secure/attachment/12356679/HADOOP-1322_20070503_2.patch applied and successfully tested against trunk revision r534624.
Test results: http://lucene.zones.apache.org:8080/hudson/job/Hadoop-Patch/107/testReport/
Console output: http://lucene.zones.apache.org:8080/hudson/job/Hadoop-Patch/107/console
> Tasktracker blacklist leads to hung jobs in single-node cluster
> ---------------------------------------------------------------
>
> Key: HADOOP-1322
> URL: https://issues.apache.org/jira/browse/HADOOP-1322
> Project: Hadoop
> Issue Type: Bug
> Components: mapred
> Reporter: Arun C Murthy
> Assigned To: Arun C Murthy
> Priority: Critical
> Fix For: 0.13.0
>
> Attachments: HADOOP-1322_20070503_1.patch, HADOOP-1322_20070503_2.patch
>
>
> Post HADOOP-1278, adding _the_ tasktracker to the blacklist in single-node clusters leads to a situation where the entire cluster is 'flaky' and no trackers are available to execute tasks.
> The straight-forward fix to check before adding the tracker to the blacklist.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HADOOP-1322) Tasktracker blacklist leads to hung
jobs in single-node cluster
Posted by "Doug Cutting (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-1322?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Doug Cutting updated HADOOP-1322:
---------------------------------
Resolution: Fixed
Status: Resolved (was: Patch Available)
I just committed this. Thanks, Arun!
> Tasktracker blacklist leads to hung jobs in single-node cluster
> ---------------------------------------------------------------
>
> Key: HADOOP-1322
> URL: https://issues.apache.org/jira/browse/HADOOP-1322
> Project: Hadoop
> Issue Type: Bug
> Components: mapred
> Reporter: Arun C Murthy
> Assigned To: Arun C Murthy
> Priority: Critical
> Fix For: 0.13.0
>
> Attachments: HADOOP-1322_20070503_1.patch, HADOOP-1322_20070503_2.patch
>
>
> Post HADOOP-1278, adding _the_ tasktracker to the blacklist in single-node clusters leads to a situation where the entire cluster is 'flaky' and no trackers are available to execute tasks.
> The straight-forward fix to check before adding the tracker to the blacklist.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HADOOP-1322) Tasktracker blacklist leads to hung
jobs in single-node cluster
Posted by "Doug Cutting (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-1322?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12493485 ]
Doug Cutting commented on HADOOP-1322:
--------------------------------------
I forgot to include the bug ID in the commit message. So, for the record, this is in revision 534973.
http://svn.apache.org/viewvc?view=rev&revision=534973
> Tasktracker blacklist leads to hung jobs in single-node cluster
> ---------------------------------------------------------------
>
> Key: HADOOP-1322
> URL: https://issues.apache.org/jira/browse/HADOOP-1322
> Project: Hadoop
> Issue Type: Bug
> Components: mapred
> Reporter: Arun C Murthy
> Assigned To: Arun C Murthy
> Priority: Critical
> Fix For: 0.13.0
>
> Attachments: HADOOP-1322_20070503_1.patch, HADOOP-1322_20070503_2.patch
>
>
> Post HADOOP-1278, adding _the_ tasktracker to the blacklist in single-node clusters leads to a situation where the entire cluster is 'flaky' and no trackers are available to execute tasks.
> The straight-forward fix to check before adding the tracker to the blacklist.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HADOOP-1322) Tasktracker blacklist leads to hung
jobs in single-node cluster
Posted by "Arun C Murthy (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-1322?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Arun C Murthy updated HADOOP-1322:
----------------------------------
Priority: Critical (was: Major)
> Tasktracker blacklist leads to hung jobs in single-node cluster
> ---------------------------------------------------------------
>
> Key: HADOOP-1322
> URL: https://issues.apache.org/jira/browse/HADOOP-1322
> Project: Hadoop
> Issue Type: Bug
> Components: mapred
> Reporter: Arun C Murthy
> Assigned To: Arun C Murthy
> Priority: Critical
> Fix For: 0.13.0
>
> Attachments: HADOOP-1322_20070503_1.patch
>
>
> Post HADOOP-1278, adding _the_ tasktracker to the blacklist in single-node clusters leads to a situation where the entire cluster is 'flaky' and no trackers are available to execute tasks.
> The straight-forward fix to check before adding the tracker to the blacklist.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HADOOP-1322) Tasktracker blacklist leads to hung
jobs in single-node cluster
Posted by "Arun C Murthy (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-1322?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Arun C Murthy updated HADOOP-1322:
----------------------------------
Attachment: HADOOP-1322_20070503_2.patch
This one should be it: check to ensure too many trackers aren't on the blacklist before rejecting the tracker.
> Tasktracker blacklist leads to hung jobs in single-node cluster
> ---------------------------------------------------------------
>
> Key: HADOOP-1322
> URL: https://issues.apache.org/jira/browse/HADOOP-1322
> Project: Hadoop
> Issue Type: Bug
> Components: mapred
> Reporter: Arun C Murthy
> Assigned To: Arun C Murthy
> Priority: Critical
> Fix For: 0.13.0
>
> Attachments: HADOOP-1322_20070503_1.patch, HADOOP-1322_20070503_2.patch
>
>
> Post HADOOP-1278, adding _the_ tasktracker to the blacklist in single-node clusters leads to a situation where the entire cluster is 'flaky' and no trackers are available to execute tasks.
> The straight-forward fix to check before adding the tracker to the blacklist.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HADOOP-1322) Tasktracker blacklist leads to hung
jobs in single-node cluster
Posted by "Arun C Murthy (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-1322?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Arun C Murthy updated HADOOP-1322:
----------------------------------
Status: Patch Available (was: Open)
> Tasktracker blacklist leads to hung jobs in single-node cluster
> ---------------------------------------------------------------
>
> Key: HADOOP-1322
> URL: https://issues.apache.org/jira/browse/HADOOP-1322
> Project: Hadoop
> Issue Type: Bug
> Components: mapred
> Reporter: Arun C Murthy
> Assigned To: Arun C Murthy
> Priority: Critical
> Fix For: 0.13.0
>
> Attachments: HADOOP-1322_20070503_1.patch, HADOOP-1322_20070503_2.patch
>
>
> Post HADOOP-1278, adding _the_ tasktracker to the blacklist in single-node clusters leads to a situation where the entire cluster is 'flaky' and no trackers are available to execute tasks.
> The straight-forward fix to check before adding the tracker to the blacklist.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HADOOP-1322) Tasktracker blacklist leads to hung
jobs in single-node cluster
Posted by "Hadoop QA (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-1322?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12493251 ]
Hadoop QA commented on HADOOP-1322:
-----------------------------------
+1
http://issues.apache.org/jira/secure/attachment/12356678/HADOOP-1322_20070503_1.patch applied and successfully tested against trunk revision r534624.
Test results: http://lucene.zones.apache.org:8080/hudson/job/Hadoop-Patch/106/testReport/
Console output: http://lucene.zones.apache.org:8080/hudson/job/Hadoop-Patch/106/console
> Tasktracker blacklist leads to hung jobs in single-node cluster
> ---------------------------------------------------------------
>
> Key: HADOOP-1322
> URL: https://issues.apache.org/jira/browse/HADOOP-1322
> Project: Hadoop
> Issue Type: Bug
> Components: mapred
> Reporter: Arun C Murthy
> Assigned To: Arun C Murthy
> Priority: Critical
> Fix For: 0.13.0
>
> Attachments: HADOOP-1322_20070503_1.patch, HADOOP-1322_20070503_2.patch
>
>
> Post HADOOP-1278, adding _the_ tasktracker to the blacklist in single-node clusters leads to a situation where the entire cluster is 'flaky' and no trackers are available to execute tasks.
> The straight-forward fix to check before adding the tracker to the blacklist.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HADOOP-1322) Tasktracker blacklist leads to hung
jobs in single-node cluster
Posted by "Arun C Murthy (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-1322?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Arun C Murthy updated HADOOP-1322:
----------------------------------
Status: Open (was: Patch Available)
Koji pointed out another corner case:
Two tasktrackers in the cluster. one is put to blacklist. and then another one goes down... this patch won't handle that.
> Tasktracker blacklist leads to hung jobs in single-node cluster
> ---------------------------------------------------------------
>
> Key: HADOOP-1322
> URL: https://issues.apache.org/jira/browse/HADOOP-1322
> Project: Hadoop
> Issue Type: Bug
> Components: mapred
> Reporter: Arun C Murthy
> Assigned To: Arun C Murthy
> Priority: Critical
> Fix For: 0.13.0
>
> Attachments: HADOOP-1322_20070503_1.patch
>
>
> Post HADOOP-1278, adding _the_ tasktracker to the blacklist in single-node clusters leads to a situation where the entire cluster is 'flaky' and no trackers are available to execute tasks.
> The straight-forward fix to check before adding the tracker to the blacklist.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HADOOP-1322) Tasktracker blacklist leads to hung
jobs in single-node cluster
Posted by "Arun C Murthy (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-1322?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Arun C Murthy updated HADOOP-1322:
----------------------------------
Attachment: HADOOP-1322_20070503_1.patch
Simple fix.
> Tasktracker blacklist leads to hung jobs in single-node cluster
> ---------------------------------------------------------------
>
> Key: HADOOP-1322
> URL: https://issues.apache.org/jira/browse/HADOOP-1322
> Project: Hadoop
> Issue Type: Bug
> Components: mapred
> Reporter: Arun C Murthy
> Assigned To: Arun C Murthy
> Fix For: 0.13.0
>
> Attachments: HADOOP-1322_20070503_1.patch
>
>
> Post HADOOP-1278, adding _the_ tasktracker to the blacklist in single-node clusters leads to a situation where the entire cluster is 'flaky' and no trackers are available to execute tasks.
> The straight-forward fix to check before adding the tracker to the blacklist.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.