U29 is definitely the way to go as long as you're using exits. You can only end up with all datasets full if your datasets are too small or some process is logging SMF records faster than your SMFDUMP task can unload them. The MPF exit seems to wait until it's critical to dump - at which point, if the dump doesn't happen, you can lose SMF data when the buffers fill.
The other thing we do to help speed it up (and I'm sure this is not unique) is to dump to GDG (+1) on disk to avoid any tape mount delays in the dump process; we have a daily process to pick up all the generations and filter them to the various collections of SMF data we need. Tim Hare Senior Systems Programmer Florida Department of Transportation (850) 414-4209 IBM Mainframe Discussion List <[email protected]> wrote on 09/28/2006 05:30:19 PM: > On Thu, 2006-09-28 at 13:43 -0500, Matthew Stitt wrote: > > You might want to switch to using the SMFDUMP program. It dumps all the SMF > > datasets automagically evertime it runs. I also have an MPF exit which > > traps the "ALL DATASETS FULL" message and automatically starts the SMFDUMP > > procedure. > > > > I've got a copy which works up to 1.7. I'm sure others do too. > > This is the answer to the underlying issue. Check the cbt. > > If you use U29, you can detect the SMF switch for each dataset and avoid > the MPF requirement mentioned. Has the benefit that there is always > (hopefully) at least one empty SMF dataset available. > > Shane ... > > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO > Search the archives at http://bama.ua.edu/archives/ibm-main.html ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
