MapOutputBuffer allocates 4x as much space to record capacity as intended

• Key: HADOOP-3475
• Type: Bug
• Status: Resolved
• Resolution: Fixed
• Priority: Major
• Assignee: Chris Douglas
• Reporter: Chris Douglas
• Affects Version/s: 0.17.0
• Fix Version/s: 0.17.1

Description

When setting up its accounting tables, MapTask.MapOutputBuffer calculates the size of each of its buffers as:

Where #total_bytes is $io.sort.mb * 2^20 and #accounting_bytes is $io.sort.record.percent * #total_bytes
kvbuffer: #total_bytes - #accounting_bytes
kvoffsets: 1/4 of #accounting_bytes
kvindices: 3/4 of #accounting_bytes

kvoffsets and kvindices are ints, not bytes; the size of both arrays is incorrectly calculated.

Comments

[Chris Douglas]

Corrects accounting and puts allocations in terms of #bytes required per record. No unit tests are included, as allocations internal to MapOutputBuffer aren't easily queried or observed.

[Chris Douglas]

Updated to include fewer embarrassing errors...

[Hadoop QA]

-1 overall. Here are the results of testing the latest attachment
http://issues.apache.org/jira/secure/attachment/12383158/3475-1.patch
against trunk revision 661918.

+1 @author. The patch does not contain any @author tags.

-1 tests included. The patch doesn't appear to include any new or modified tests.
Please justify why no tests are needed for this patch.

+1 javadoc. The javadoc tool did not generate any warning messages.

+1 javac. The applied patch does not increase the total number of javac compiler warnings.

+1 findbugs. The patch does not introduce any new Findbugs warnings.

+1 release audit. The applied patch does not increase the total number of release audit warnings.

-1 core tests. The patch failed core unit tests.

+1 contrib tests. The patch passed contrib unit tests.

Test results: http://hudson.zones.apache.org/hudson/job/Hadoop-Patch/2530/testReport/
Findbugs warnings: http://hudson.zones.apache.org/hudson/job/Hadoop-Patch/2530/artifact/trunk/build/test/findbugs/newPatchFindbugsWarnings.html
Checkstyle results: http://hudson.zones.apache.org/hudson/job/Hadoop-Patch/2530/artifact/trunk/build/test/checkstyle-errors.html
Console output: http://hudson.zones.apache.org/hudson/job/Hadoop-Patch/2530/console

This message is automatically generated.

[Tsz Wo (Nicholas), SZE]

To be consistent, (Integer.SIZE / 8) should be either (Integer.SIZE / Byte.SIZE) or just 4. Otherwise, patch looks good.

[Chris Douglas]

Incorporated Nicholas's feedback

[Chris Douglas]

I just committed this.

[Hudson]

Integrated in Hadoop-trunk #511 (See http://hudson.zones.apache.org/hudson/job/Hadoop-trunk/511/)

[Chris Douglas]

Reopening this, as it is necessary for 0.17.1

[Chris Douglas]

I just committed this to the 0.17 branch.

[Hudson]

Integrated in Hadoop-trunk #523 (See http://hudson.zones.apache.org/hudson/job/Hadoop-trunk/523/)