We have some ETL jobs that are scheduled to run every day, and some that are scheduled to run every week via Control-M. These types of jobs tag data with the date the job was run and perform filter operations to get activity for that particular day or week. There's also a final step that involves taking the daily file, merging it into the existing archive of data, and then deleting the daily file.

When these types of jobs fail (or put on hold because third party inputs are bad, not available, etc) it requires lots of manual labor to reschedule all of these jobs to re-run again for their scheduled days. The jobs also need to be run in chronological order for the data output to be correct. I'm not a fan of these kinds of setups, as being a couple of days behind can suddenly mean spending 2 whole days queuing up 20 jobs to run for the days missed.

Are there better ways of managing these scenarios that reduce the manual work required? Control-M thankfully does keep the failed jobs in Monitoring so that they can be restarted where they left off, but sometimes this requires making manual edits to the job so that it doesn't accidentally read in raw input files it shouldn't read for the particular day/week it is running for.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.