hadoop-mapreduce-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Akira AJISAKA (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (MAPREDUCE-6499) Add More Informations Of Jobs In JobHistory Main Page
Date Wed, 28 Oct 2015 02:43:27 GMT

    [ https://issues.apache.org/jira/browse/MAPREDUCE-6499?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14977612#comment-14977612
] 

Akira AJISAKA commented on MAPREDUCE-6499:
------------------------------------------

Thank you for updating the patch, [~linyiqun].
bq. But there is a problem when the history job is larger enough, it will be load much more
job infos and so slowly to open jobhistory pages.
Good point. If you really need to sort by average map time and other information, we need
to add them into index not to load full job information (See JobIndexInfo.java). In addition,
we need to choose what information to add very carefully considering performance. 
IMO, adding elapsed time to the main page is a good idea because
* Elapsed time can be calculated without loading full job info or editing index file.
* Users can easily detect long-running-jobs by sorting by elapsed time

{code}
-        th().input("search_init").$type(InputType.text).$name("start_time").$value("Job ID")._()._().
-        th().input("search_init").$type(InputType.text).$name("start_time").$value("Name")._()._().
-        th().input("search_init").$type(InputType.text).$name("start_time").$value("User")._()._().
-        th().input("search_init").$type(InputType.text).$name("start_time").$value("Queue")._()._().
-        th().input("search_init").$type(InputType.text).$name("start_time").$value("State")._()._().
-        th().input("search_init").$type(InputType.text).$name("start_time").$value("Maps
Total")._()._().
-        th().input("search_init").$type(InputType.text).$name("start_time").$value("Maps
Completed")._()._().
-        th().input("search_init").$type(InputType.text).$name("start_time").$value("Reduces
Total")._()._().
-        th().input("search_init").$type(InputType.text).$name("start_time").$value("Reduces
Completed")._()._()
(snip)
+        th().input("search_init").$type(InputType.text)
+          .$name("job_id").$value("Job ID")._()._().
+        th().input("search_init").$type(InputType.text)
+          .$name("name").$value("Name")._()._().
+        th().input("search_init").$type(InputType.text)
+          .$name("user").$value("User")._()._().
+        th().input("search_init").$type(InputType.text)
+          .$name("queue").$value("Queue")._()._().
+        th().input("search_init").$type(InputType.text)
+          .$name("state").$value("State")._()._().
+        th().input("search_init").$type(InputType.text)
+          .$name("maps_total").$value("Maps Total")._()._().
+        th().input("search_init").$type(InputType.text).
+          $name("maps_completed").$value("Maps Completed")._()._().
+        th().input("search_init").$type(InputType.text).
+          $name("reduces_total").$value("Reduces Total")._()._().
+        th().input("search_init").$type(InputType.text).
+          $name("reduces_completed").$value("Reduces Completed")._()._().
{code}
Good catch. We need to fix it.

> Add More Informations Of Jobs In JobHistory Main Page
> -----------------------------------------------------
>
>                 Key: MAPREDUCE-6499
>                 URL: https://issues.apache.org/jira/browse/MAPREDUCE-6499
>             Project: Hadoop Map/Reduce
>          Issue Type: Improvement
>    Affects Versions: 2.7.1
>            Reporter: Lin Yiqun
>            Assignee: Lin Yiqun
>            Priority: Trivial
>         Attachments: HADOOP-MAPREDUCE-6499.2.patch, HADOOP-MAPREDUCE-6499.3.patch, HADOOP-MAPREDUCE-6499.4.patch,
HADOOP-MAPREDUCE-6499.5.patch, HADOOP-MAPREDUCE-6499.patch
>
>
> Now in  JobHistory Main Page show too little information about finished jobs,even don't
have the job's running total time,only startTime,finishedTime.So,in jobHistory main page,we
can add more informations about jobs, that we can better analyze jobs.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message