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 "Amar Kamat (JIRA)" <ji...@apache.org> on 2008/08/25 20:37:44 UTC
[jira] Created: (HADOOP-4017) Unify the way job history filename
is parsed
Unify the way job history filename is parsed
---------------------------------------------
Key: HADOOP-4017
URL: https://issues.apache.org/jira/browse/HADOOP-4017
Project: Hadoop Core
Issue Type: Bug
Components: mapred
Reporter: Amar Kamat
Job history filename has the following meta-info :
- jobtracker's hostname
- job id
- username
- jobname
{{HistoryViewer.java}} and {{jobhistory.jsp}} are required to parse the history filename to extract the meta-info. It makes more sense to provide a common utility in {{JobHistory}} to do it.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HADOOP-4017) Unify the way job history filename
is parsed
Posted by "Amar Kamat (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-4017?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12627934#action_12627934 ]
Amar Kamat commented on HADOOP-4017:
------------------------------------
The way job details are recovered from the jobhistory filename is buggy. Every piece of code that tries to infer this information does something like
{code}
1. split on '_'
2) username = split[5].
{code}
But the username can also have a '_' and this trick might no longer work. Same is true for jobname.
> Unify the way job history filename is parsed
> ---------------------------------------------
>
> Key: HADOOP-4017
> URL: https://issues.apache.org/jira/browse/HADOOP-4017
> Project: Hadoop Core
> Issue Type: Bug
> Components: mapred
> Reporter: Amar Kamat
>
> Job history filename has the following meta-info :
> - jobtracker's hostname
> - job id
> - username
> - jobname
> {{HistoryViewer.java}} and {{jobhistory.jsp}} are required to parse the history filename to extract the meta-info. It makes more sense to provide a common utility in {{JobHistory}} to do it.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HADOOP-4017) Unify the way job history filename
is parsed
Posted by "Amar Kamat (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-4017?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12642413#action_12642413 ]
Amar Kamat commented on HADOOP-4017:
------------------------------------
Two things missing
1) The case where the _username_/_jobname_ contains a '_'.
2) Test case.
> Unify the way job history filename is parsed
> ---------------------------------------------
>
> Key: HADOOP-4017
> URL: https://issues.apache.org/jira/browse/HADOOP-4017
> Project: Hadoop Core
> Issue Type: Bug
> Components: mapred
> Reporter: Amar Kamat
> Assignee: Amar Kamat
> Attachments: HADOOP-4017-v1.1.patch, HADOOP-4017-v1.patch
>
>
> Job history filename has the following meta-info :
> - jobtracker's hostname
> - job id
> - username
> - jobname
> {{HistoryViewer.java}} and {{jobhistory.jsp}} are required to parse the history filename to extract the meta-info. It makes more sense to provide a common utility in {{JobHistory}} to do it.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HADOOP-4017) Unify the way job history filename
is parsed
Posted by "Amar Kamat (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-4017?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Amar Kamat updated HADOOP-4017:
-------------------------------
Attachment: HADOOP-4017-v1.patch
Attaching a patch the implements the basic idea. I have found out 2 places where the job-history filename was hacked
1) jobhistory.jsp
2) HistoryViewer.java
Am I missing something?
> Unify the way job history filename is parsed
> ---------------------------------------------
>
> Key: HADOOP-4017
> URL: https://issues.apache.org/jira/browse/HADOOP-4017
> Project: Hadoop Core
> Issue Type: Bug
> Components: mapred
> Reporter: Amar Kamat
> Attachments: HADOOP-4017-v1.patch
>
>
> Job history filename has the following meta-info :
> - jobtracker's hostname
> - job id
> - username
> - jobname
> {{HistoryViewer.java}} and {{jobhistory.jsp}} are required to parse the history filename to extract the meta-info. It makes more sense to provide a common utility in {{JobHistory}} to do it.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HADOOP-4017) Unify the way job history filename
is parsed
Posted by "Amar Kamat (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-4017?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Amar Kamat updated HADOOP-4017:
-------------------------------
Attachment: HADOOP-4017-v1.1.patch
A small bug fixed.
> Unify the way job history filename is parsed
> ---------------------------------------------
>
> Key: HADOOP-4017
> URL: https://issues.apache.org/jira/browse/HADOOP-4017
> Project: Hadoop Core
> Issue Type: Bug
> Components: mapred
> Reporter: Amar Kamat
> Assignee: Amar Kamat
> Attachments: HADOOP-4017-v1.1.patch, HADOOP-4017-v1.patch
>
>
> Job history filename has the following meta-info :
> - jobtracker's hostname
> - job id
> - username
> - jobname
> {{HistoryViewer.java}} and {{jobhistory.jsp}} are required to parse the history filename to extract the meta-info. It makes more sense to provide a common utility in {{JobHistory}} to do it.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Assigned: (HADOOP-4017) Unify the way job history filename
is parsed
Posted by "Amar Kamat (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-4017?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Amar Kamat reassigned HADOOP-4017:
----------------------------------
Assignee: Amar Kamat
> Unify the way job history filename is parsed
> ---------------------------------------------
>
> Key: HADOOP-4017
> URL: https://issues.apache.org/jira/browse/HADOOP-4017
> Project: Hadoop Core
> Issue Type: Bug
> Components: mapred
> Reporter: Amar Kamat
> Assignee: Amar Kamat
> Attachments: HADOOP-4017-v1.patch
>
>
> Job history filename has the following meta-info :
> - jobtracker's hostname
> - job id
> - username
> - jobname
> {{HistoryViewer.java}} and {{jobhistory.jsp}} are required to parse the history filename to extract the meta-info. It makes more sense to provide a common utility in {{JobHistory}} to do it.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.