hadoop-mapreduce-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Vrushali C (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (MAPREDUCE-5432) JobHistoryParser does not fetch clockSplits, cpuUsages, vMemKbytes, physMemKbytes from history file
Date Tue, 06 Aug 2013 18:43:56 GMT

    [ https://issues.apache.org/jira/browse/MAPREDUCE-5432?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13731078#comment-13731078
] 

Vrushali C commented on MAPREDUCE-5432:
---------------------------------------

A few minor comments:

+      printAttributes("CLOCK_SPLIT", clockSplits);
+      printAttributes("CPU_USAGE", cpuUsages);
Maybe print it as "CLOCK_SPLITS" AND "CPU_USAGES" so that it is consistent with the naming
of the attribute?


for the following function: 
+    void printAttributes(String tag, int[] attributes) {
+      for (int i=0;i<attributes.length;i++) {
+        System.out.println(tag + "[" + i + "]:" + attributes[i]);
+     }
+    }

I believe it would print it as: 
CLOCK_SPLIT[0]: 8691
CLOCK_SPLIT[1]: 129
CLOCK_SPLIT[2]: 128
CLOCK_SPLIT[3]: 129
CLOCK_SPLIT[4]: 128
... etc

Would it be slightly better to print it as: 
"clockSplits":[8691,129,128,129,128,129,128,129,128,129,128,129]
"cpuUsages":[28,28,29,28,28,29,28,28,29,28,28,29]



                
> JobHistoryParser does not fetch clockSplits, cpuUsages, vMemKbytes, physMemKbytes from
history file
> ---------------------------------------------------------------------------------------------------
>
>                 Key: MAPREDUCE-5432
>                 URL: https://issues.apache.org/jira/browse/MAPREDUCE-5432
>             Project: Hadoop Map/Reduce
>          Issue Type: Bug
>          Components: jobhistoryserver, mrv2
>    Affects Versions: 2.0.4-alpha
>            Reporter: Vrushali C
>            Assignee: Tsuyoshi OZAWA
>         Attachments: MAPREDUCE-5432.1.patch
>
>
> JobHistoryParser's handleMapAttemptFinishedEvent() function does not look at MapAttemptFinishedEvent's
  
>   int[] clockSplits;
>   int[] cpuUsages;
>   int[] vMemKbytes;
>   int[] physMemKbytes;
> JobHistoryParser's inner class TaskAttemptInfo also needs to be enhanced to have these
as members so that handleMapAttemptFinishedEvent() can get them and store them.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

Mime
View raw message