On Mon, 6 Mar 2006 16:43:01 -0700, Paul Gilmartin <[EMAIL PROTECTED]> wrote: >I work in a development lab. Backups are automated by the meager >scheduling facilities of HSM. That's all the admin I asked >offered; I elected not to badger her for more, but she offered >that our production shop uses more sophisticated commercial >schedulers. > >What scheduled maintenance should I expect in a development >lab? Perhaps scratching tramp data sets? Others? >
Not all that much is required for a system that is 100% development. It depends what you or capacity planning people might want to look at for planning purposes or perhaps to find out something that happened on the system at a previous date/time. Here are a few things I can think of: SMF data - May be dumped automatically via IEFU29 SMF exit after a switch or once a day via SMFDUMP program (see the archvies), but are there any scheduled jobs to combine all the daily dumps into weekly (or monthly). If there is a production LPAR on the same CEC, then perhaps the people that watch capacity only care about the total utilization of the LPAR, which is available in the other LPARs' SMF (RMF) data. SMF dump could go to GDG via IEFU29 with no combining of data or SMF can be set to not record at all (I had a P390 client that had their system set up that way to save cycles). SYSLOG - This is an easy one to handle with JES2 automatic command or automation command at a regular time each day to issue the WRITELOG command and submit a job or start an STC to copy the syslog to a GDG. Another option: the log just gets spun when IPLing. Then it stays in the spool until someone feels like purging it. LOGREC - Collection and archival, or at least "ZERO"-ing the logrec data set / logstream. Easy to handle the same way as syslog. LOGREC could be set to ignore or I suppose the clear job just submitted manually when someone sees "LOGREC FULL" message. I wouldn't want to set it to ignore on a flex or p390 because it contains software errors as well as hardware errors. There can be many others, but they are not required. For example, JES2 automatic commands to purge spool output older than n number of days. If you don't have that, do people just wait until the spool gets full and start purging output, or do the operators just issue the command to do so picking a number for N arbitrarily? It all depends on how many people really use your system. If you can just shout over a couple of cubicles and ask everyone to purge their output they don't need, that works also. Scratching temp data sets is another one. I can't remember if temp data sets are controlled via SMS on ADCD systems. But even if they are, there is a good chance there is junk on the temp work packs if the system has ever crashed. Find out where some of yours go and look at the volume via ISPF 3.4 and you may find lots of old temp data sets that need to be deleted. There are developers on this list the run flex or P390s still where they are the only user of the system or there are only a few users. I guess they can tell you what regular maintenance they run if any. Regards, Mark -- Mark Zelden Sr. Software and Systems Architect - z/OS Team Lead Zurich North America / Farmers Insurance Group mailto: [EMAIL PROTECTED] Systems Programming expert at http://expertanswercenter.com/ Mark's MVS Utilities: http://home.flash.net/~mzelden/mvsutil.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