Error handling of processes

You can ask PyPPL to terminate, retry or ignore a job that fails to run. When a job finishes, it should generate a job.rc file containing the return code. When compare with the valid return codes pXXX.rc, the error triggered if it is not in pXXX.rc. pXXX.errhow determines what's next if errors happen.

  • "terminate": when errors happen, terminate the entire pipeline
  • "ignore" : ignore the errors, continuing run the next process
  • "retry" : re-submit and run the job again. pXXX.errntry defines how many time to retry.
  • "halt" : try to halt the entire pipeline if any job encounters error.