AnsweredAssumed Answered

Can't open jobcache in /tmp/mapr-hadoop/mapred/local/taskTracker/... - Permission denied

Question asked by marek on Mar 5, 2012
Latest reply on Mar 5, 2012 by marek
Hello,

I am getting "permission denied" error when attempting to run a test job (see example below). 
The TestDFSIO job is kicked-off on the "mapr-client" (as user "marek"); the server/cluster is the M3-demo-VM ("mapr-desktop", 10.10.0.243).  

I my testing I found out that if I create a "mapr" user on the mapr-client (with same uid/gid as the mapr user already existing on the server), and run the job as mapr@mapr-client, then the job succeeds. However, running the job as another user ("marek" in my case) fails.

This is a new install that I am playing with (trying to learn things).  I bet there is some "missing step" I did not do (most likely related to permissions)... just not sure what. What is the proper procedure/steps to create a valid user? Please advice :) 

Thanks, 
Marek

--------
[marek@mapr-client ~]$ **hadoop jar $HADOOP_INSTALL/hadoop-*-test.jar TestDFSIO -write -nrFiles 13** 
  TestDFSIO.0.0.4 
  12/03/05 16:31:43 INFO fs.TestDFSIO: nrFiles = 13 
  12/03/05 16:31:43 INFO fs.TestDFSIO: fileSize (MB) = 1.0 
  12/03/05 16:31:43 INFO fs.TestDFSIO: bufferSize = 1000000 
  12/03/05 16:31:43 INFO fs.TestDFSIO: baseDir = /benchmarks/TestDFSIO 
  12/03/05 16:31:43 INFO fs.TestDFSIO: creating control file: 1048576 bytes, 13 files 
  12/03/05 16:31:43 INFO fs.TestDFSIO: created control files for: 13 files 
  12/03/05 16:31:44 INFO fs.JobTrackerWatcher: Current running JobTracker is: mapr-desktop/10.10.0.243:9001 
  12/03/05 16:31:44 INFO mapred.FileInputFormat: Total input paths to process : 13 
  12/03/05 16:31:44 INFO mapred.JobClient: Running job: job_201203021455_0012 
  12/03/05 16:31:45 INFO mapred.JobClient:  map 0% reduce 0% 
  12/03/05 16:31:47 INFO mapred.JobClient: Task Id : attempt_201203021455_0012_m_000000_0, Status : FAILED on node mapr-desktop 
  Error initializing attempt_201203021455_0012_m_000000_0: 
  java.io.IOException: Job initialization failed (255)org.apache.hadoop.util.Shell$ExitCodeException: . Output of LinuxTaskController initializeJob: Reading task controller config from /opt/mapr/hadoop/hadoop-0.20.2/conf/taskcontroller.cfg 
  main : command provided 0 
  main : user is marek 
  **Can't open jobcache in /tmp/mapr-hadoop/mapred/local/taskTracker/marek/jobcache/job_201203021455_0012 - Permission denied** 
  Did not create any job directories 
          at org.apache.hadoop.mapred.LinuxTaskController.initializeJob(LinuxTaskController.java:195) 
          at org.apache.hadoop.mapred.TaskTracker$4.run(TaskTracker.java:1576) 
          at java.security.AccessController.doPrivileged(Native Method) 
          at javax.security.auth.Subject.doAs(Subject.java:396) 
          at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1109) 
          at org.apache.hadoop.mapred.TaskTracker.initializeJob(TaskTracker.java:1551) 
          at org.apache.hadoop.mapred.TaskTracker.localizeJob(TaskTracker.java:1433) 
          at org.apache.hadoop.mapred.TaskTracker.startNewTask(TaskTracker.java:3718) 
          at org.apache.hadoop.mapred.TaskTracker$TaskLauncher.run(TaskTracker.java:3641) 
    ... 
    

Outcomes