[ 
https://issues.apache.org/jira/browse/MAPREDUCE-1925?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12889758#action_12889758
 ] 

Ravi Gummadi commented on MAPREDUCE-1925:
-----------------------------------------

    *  Invert the order of closing the input stream and returning the the 
decompressor in PossiblyDecompressedInputStream. Rationale: 
CompressionInputStream.close() may still access the decompressor.

OK. Will do this.

    * Explicitly make the buffer size and mark limit consistent. E.g. "static 
final int BUFSIZE = 10000; BufferedInputStream bis = new 
BufferedInputStream(inputLogStream, BUFSIZE); bis.mark(BUFSIZE);"

I didn't get why these 2 should be same. As default BUFSIZE is 8K, we just need 
some number > 8k and the testcase is using 10000. Isn't this OK ? Any issues ? 
Just to be on the safer side, if you want it to be increased to a bigger value 
like 32K(from 10000), I can do so. Do you see any gain because of having these 
2 same ?

    * Move this "Hadoop20JHParser parser = new Hadoop20JHParser(bis);" after 
bis.reset(). Rationale: in case in the future this constructor may read a few 
bytes from the stream.

This was already incorporated in my latest patch 1925.v1.1.patch.


> TestRumenJobTraces fails in trunk
> ---------------------------------
>
>                 Key: MAPREDUCE-1925
>                 URL: https://issues.apache.org/jira/browse/MAPREDUCE-1925
>             Project: Hadoop Map/Reduce
>          Issue Type: Bug
>          Components: tools/rumen
>    Affects Versions: 0.22.0
>            Reporter: Amareshwari Sriramadasu
>            Assignee: Ravi Gummadi
>             Fix For: 0.22.0
>
>         Attachments: 1925.patch, 1925.v1.1.patch, 1925.v1.patch
>
>
> TestRumenJobTraces failed with following error:
> Error Message
> the gold file contains more text at line 1 expected:<56> but was:<0>
> Stacktrace
>       at 
> org.apache.hadoop.tools.rumen.TestRumenJobTraces.testHadoop20JHParser(TestRumenJobTraces.java:294)
> Full log of the failure is available at 
> http://hudson.zones.apache.org/hudson/job/Mapreduce-Patch-h4.grid.sp2.yahoo.net/292/testReport/org.apache.hadoop.tools.rumen/TestRumenJobTraces/testHadoop20JHParser/

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.

Reply via email to