menu

When the job does not execute properly

When a job does not execute properly, please confirm the following:

  • Can you try to run the same job again? Was the result correct?
  • Is this the first time to run this job or have you done it before?
  • Is it possible to execute in single execution (1 core 1 process, etc.)?
  • Can you run it on other systems such as PC or other computers?

Please contact the Research Center for Advanced Computing Infrastructure with the failed job ID, etc., if you can execute on another computer, and other information such as whether you were able to do the job successfully before, etc.

 

 

 

Troubles/Issues

 It may be not be possible to determine whether the job ended due to device failure, network trouble, or excessive use of the memory by the job itself.

In such a case, please refer to the following method to confirm:

1.Terminate the job and run it again
If the job takes a short time to run, the problem can be confirmed this way.


When investigating the cause of the problem while the program is executing,please consider the effects of the job execution using following confirmation method

2. (When job writes results sequentially)Confirm the processing part near the relevant time in the result file

3. (After job execution)Check PBS error output, error description in the standard output file
After the job has executed JOBNAME.eXXXX<job.number>, or JOBNAME.oXXXX<job.number> Please refer to the contents of the file written and check whether there is any difference from the contents normally exported.

If a process is affected, it usually indicates an error on the error output or standard output.