MVSFORUMS.com

[RajSanj]

Hi,

The place where I work we have systems where most of the jobs are run manually from there personal libraries instead of using a scheduler. We get the feed from the customers & we run our jobs after we get these input files. There are lot of redundant jobs running due to absence of scheduler in place. Also most of these jobs generate huge intermediate files which the user stores to avoid re-running the job again.

Few points which I thought could be implemented are:

[calspach]

Well, the scheduler is obviously the first step. That will allow you to quit running the duplicates. Get everyone together and set up the schedule so somebody knows what everyone is doing.

As for the file space issue, I wouldn't worry about it. Storage is cheap, processing is expensive. If it's really an issue, you can use the same file each night so it will overwrite itself, but I really wouldn't recommend using temp datasets that won't be there if you do need to restart.