Configure Data Transfer to and from the Other Batch System

Before running a job on the Other Batch System, you may need to transfer data from the local system to the Other Batch System as input to the job.

About this task

When the job completes, you may also need to transfer the standard output/error from the Other Batch System back to the local system.

To do this, you can configure submission options for input, output and error files, and create one local job in your flow definition that precedes the Other Batch job and another local job that follows the Other Batch job to handle input and output files.

1. In your submit.conf file, specify standard input, output, and error as customized options to the Job Definition.

   For example:

   #Label            #Environment Variable               #Required 
   "Input File"      JS_EE_SUBMISSION_STDIN              0
   "Output File"     JS_EE_SUBMISSION_STDOUT             0
   "Error File"      JS_EE_SUBMISSION_STDERR             0

2. Edit your submit.sh script file so that jobs are submitted with the appropriate input, output, and error file options for your specific batch system.

   For example, -i JS_EE_SUBMISSION_STDIN, –o $JS_EE_SUBMISSION_STDOUT, –e $JS_EE_SUBMISSION_ERR.

3. In Flow Editor, in your flow definition, define a local job immediately before and immediately after your Other Batch job.

   The local job that directly precedes your Other Batch job copies the local data to the Other Batch System.

   The local job that directly follows your Other Batch job copies the output and error data from the Other Batch System to the local system.

   For example:

   
4. In the Job Definition of the Other Batch job, ensure the correct path to the Other Batch System is specified.

   For example:

   • Input file: /home/usr1/input.dat
   • Output file: /home/usr1/output.dat
   • Error file: /home/usr1/error.dat