Improve FsShell -du/-dus and FileSystem.getContentSummary efficiency

• Key: HADOOP-4339
• Type: Improvement
• Status: Closed
• Resolution: Fixed
• Priority: Major
• Assignee: David Phillips
• Reporter: David Phillips
• Component/s: fs
• Affects Version/s: 0.18.1
• Fix Version/s: 0.20.0

Description

FsShell.du has two inefficiencies:

• calling getContentSummary twice for each top-level item rather than calling it once and saving the result
• calling getContentSummary for files rather than using the size it already has in FileStatus

getContentSummary has one:

• calling itself for files rather than using the length it already has in FileStatus

Every call to getContentSummary results in a call to getFileStatus, which may be expensive (e.g. NativeS3FileSystem has both network latency and actual monetary cost).

The simple solution:

• FsShell.du calls once per item and saves the ContentSummary
• FsShell.du uses FileStatus.getLen for files
• getContentSummary only calls itself for directories

Another solution, rather than adding special casing to callers, is to add a getContentSummary that takes a FileStatus.

Comments

[David Phillips]

Patch for the simple solution. This reduces calls to getFileStatus to one per directory for -du and -dus [1], as opposed to the previous one (-dus) or two (-dus) per file and directory.

[1] -dus still has an extra call for the base directory due to the initial call to globStatus.

[Hadoop QA]

-1 overall. Here are the results of testing the latest attachment
http://issues.apache.org/jira/secure/attachment/12392279/hadoop-fsshell-du-simple.patch
against trunk revision 712102.

+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 Eclipse classpath. The patch retains Eclipse classpath integrity.

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

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

Test results: http://hudson.zones.apache.org/hudson/job/Hadoop-Patch/3549/testReport/
Findbugs warnings: http://hudson.zones.apache.org/hudson/job/Hadoop-Patch/3549/artifact/trunk/build/test/findbugs/newPatchFindbugsWarnings.html
Checkstyle results: http://hudson.zones.apache.org/hudson/job/Hadoop-Patch/3549/artifact/trunk/build/test/checkstyle-errors.html
Console output: http://hudson.zones.apache.org/hudson/job/Hadoop-Patch/3549/console

This message is automatically generated.

[Chris Douglas]

In FsShell, it makes more sense to save the length instead of the ContentSummary. The FileSystem change looks good.

[David Phillips]

Good point, Chris. Patch updated.

[Chris Douglas]

+1 Looks good

[Chris Douglas]

I just committed this. Thanks, David.

[Hudson]

Integrated in Hadoop-trunk #670 (See http://hudson.zones.apache.org/hudson/job/Hadoop-trunk/670/)
. Remove redundant calls from FileSystem/FsShell when
generating/processing ContentSummary. Contributed by David Phillips.