[MAHOUT-1863] cluster-syntheticcontrol.sh errors out with "Input path does not exist"

• Type: Bug
• Status: Closed
• Priority: Minor
• Resolution: Fixed
• Affects Version/s: 0.12.0
• Fix Version/s: 0.12.2
• Assignee: Andrew Palumbo
• Reporter: Albert Chu
• Created: 26/May/16 00:41
• Updated: 13/Jun/16 15:21
• Resolved: 26/May/16 23:56

Description

Running cluster-syntheticcontrol.sh on 0.12.0 resulted in this error:

Exception in thread "main" org.apache.hadoop.mapreduce.lib.input.InvalidInputException: Input path does not exist: hdfs://apex156:54310/user/achu/testdata
	at org.apache.hadoop.mapreduce.lib.input.FileInputFormat.singleThreadedListStatus(FileInputFormat.java:323)
	at org.apache.hadoop.mapreduce.lib.input.FileInputFormat.listStatus(FileInputFormat.java:265)
	at org.apache.hadoop.mapreduce.lib.input.FileInputFormat.getSplits(FileInputFormat.java:387)
	at org.apache.hadoop.mapreduce.JobSubmitter.writeNewSplits(JobSubmitter.java:301)
	at org.apache.hadoop.mapreduce.JobSubmitter.writeSplits(JobSubmitter.java:318)
	at org.apache.hadoop.mapreduce.JobSubmitter.submitJobInternal(JobSubmitter.java:196)
	at org.apache.hadoop.mapreduce.Job$10.run(Job.java:1290)
	at org.apache.hadoop.mapreduce.Job$10.run(Job.java:1287)
	at java.security.AccessController.doPrivileged(Native Method)
	at javax.security.auth.Subject.doAs(Subject.java:415)
	at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1657)
	at org.apache.hadoop.mapreduce.Job.submit(Job.java:1287)
	at org.apache.hadoop.mapreduce.Job.waitForCompletion(Job.java:1308)
	at org.apache.mahout.clustering.conversion.InputDriver.runJob(InputDriver.java:108)
	at org.apache.mahout.clustering.syntheticcontrol.fuzzykmeans.Job.run(Job.java:133)
	at org.apache.mahout.clustering.syntheticcontrol.fuzzykmeans.Job.main(Job.java:62)
	at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
	at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57)
	at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
	at java.lang.reflect.Method.invoke(Method.java:606)
	at org.apache.hadoop.util.ProgramDriver$ProgramDescription.invoke(ProgramDriver.java:71)
	at org.apache.hadoop.util.ProgramDriver.run(ProgramDriver.java:144)
	at org.apache.hadoop.util.ProgramDriver.driver(ProgramDriver.java:152)
	at org.apache.mahout.driver.MahoutDriver.main(MahoutDriver.java:195)
	at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
	at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57)
	at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
	at java.lang.reflect.Method.invoke(Method.java:606)
	at org.apache.hadoop.util.RunJar.run(RunJar.java:221)
	at org.apache.hadoop.util.RunJar.main(RunJar.java:136)

It appears cluster-syntheticcontrol.sh breaks under 0.12.0 due to patch

commit 23267a0bef064f3351fd879274724bcb02333c4a

one change in question

-    $DFS -mkdir testdata
+    $DFS -mkdir ${WORK_DIR}/testdata

now requires that the -p option be specified to -mkdir. This fix is simple.

Another change:

-    $DFS -put ${WORK_DIR}/synthetic_control.data testdata
+    $DFS -put ${WORK_DIR}/synthetic_control.data ${WORK_DIR}/testdata

appears to break the example b/c in:

examples/src/main/java/org/apache/mahout/clustering/syntheticcontrol/fuzzykmeans/Job.java
examples/src/main/java/org/apache/mahout/clustering/syntheticcontrol/kmeans/Job.java

the file 'testdata' is hard coded into the example as just 'testdata'. ${WORK_DIR}/testdata needs to be passed in as an option.

Reverting the lines listed above fixes the problem. However, the reverting presumably breaks the original problem listed in MAHOUT-1773.

I originally attempted to fix this by simply passing in the option "--input ${WORK_DIR}/testdata" into the command in the script. However, a number of other options are required if one option is specified.

I considered modifying the above Job.java files to take a minimal number of arguments and set the rest to some default, but that would have also required changes to DefaultOptionCreator.java to make required options non-optional, which I didn't want to go down the path of determining what other examples had requires/non-requires requirements.

So I just passed in every required option into cluster-syntheticcontrol.sh to fix this, using whatever defaults were hard coded into the Job.java files above.

I'm sure there's a better way to do this, and I'm happy to supply a patch, but thought I'd start with this.

Github pull request to be sent shortly.

Issue Links

• links to: GitHub Pull Request #235