[MAPREDUCE-6110] JobHistoryServer CLI throws NullPointerException with job ids that do not exist

• Type: Bug
• Status: Resolved
• Priority: Minor
• Resolution: Fixed
• Fix Version/s: 2.9.0, 3.0.0-alpha1
• Component/s: jobhistoryserver
• Hadoop Flags: Reviewed
• Assignee: Kai
• Reporter: Li Lu
• Created: 25/Sep/14 21:11
• Updated: 30/Aug/16 01:19
• Resolved: 22/Mar/16 21:32

Description

When using JobHistoryServer CLI to query a job id that does not exist on the server, it may throw NullPointerException sometimes.

I tried "mapred job -events <some_wrong_jobId> 0 100", and the result was:

Exception in thread "main" java.lang.NullPointerException
at org.apache.hadoop.mapreduce.tools.CLI.listEvents(CLI.java:487)
at org.apache.hadoop.mapreduce.tools.CLI.run(CLI.java:316)
at org.apache.hadoop.util.ToolRunner.run(ToolRunner.java:70)
at org.apache.hadoop.util.ToolRunner.run(ToolRunner.java:84)
at org.apache.hadoop.mapred.JobClient.main(JobClient.java:1237)

Similar symptoms also appear with -list-attempt-ids, but were fine with -status and -set-priority.

I traced back to CLI.listEvents, and line 487 is:

TaskCompletionEvent[] events = job.
      getTaskCompletionEvents(fromEventId, numEvents);

The job object is obtained from JobID.forName(jobid)) (line 316), which will return null if the job does not exist on server.

Maybe we want to have some behaviors consistent with -status here, by simply reporting jobId does not exist?