Things you never forget...

A few weeks ago I was interviewing MVS Systems Programmers. One of my favorite interview questions is:
What's the biggest mistake you've ever made? (As it relates to MVS Systems Programming, of course).

I love hearing the answers to this question. Like me, people never ever forget that biggest screw up. It's etched clearly on your mind. I may be naïve, but I believe that most people are inherently honest, and all of them quickly told me their stories, even though some might consider it a weakness. I consider it a building block of experience.

Winston Churchill said  “All men make mistakes, but only wise men learn from their mistakes"  One thing is for sure, whatever that big mistake was, I'll bet they never made that one again.

We are much less likely to make those kind of show-stopping mistakes these days, mostly because almost every shop has at least two lpars, the real one, and one that you can fix things from. We also have so many more functions that don't require an IPL to implement. I remember one shop who never did CLPAs unless absolutely necessary, because they never knew what was going to come in. Overall, life as an MVS systems programmer was much more risky for us back then. I sound like an old timer ‘Back in my day, we didn’t have systems to fix it from’. But that statement is analogous to ‘I walked to school every day, in the snow, no shoes, uphill, both ways’ that our parents used to say.

Back in the day I remember a program called SAFE, which came on a tape and was a StandAlone Fix Editor. It was an IPLable tape that gave you the opportunity to edit something in parmlib. I kept a 3480 at home with this on it, just in case. For whatever reason, it actually seems like it was called SAFFE to me, but that doesn't make any sense. If anyone recalls it, let me know.

All this reminds me why I've never met a backup I didn't like.

I guess it's only fair that I should tell you what I consider my biggest mistake. I moved an NCP loadlib. It was catalogued, but the VTAM proc had a volser on it for the dataset. Boom. No VTAM. Luckily I had another MVS that I had just installed, and was able to bring that up to remove the volser. Although I did have a fix, it took a few hours to get that other system up, find the problem, and fix it. Luckily, it was during a scheduled IPL.

So what was yours? Care to share? smile.JPG

Your link today is to my good friend Mark Zelden’s MVS Utilities Page:

Till next time,




A mistake I remember

September 25, 2012 12:40 PM by Mark Zelden

Since mary Anne plugged my web site, I guess I should SHARE a mistake I remember from some years ago. I'm not sure if it is the "biggest" mistake I ever made, but it is the biggest one I can remember right now. :-)

I cloned over a sysres set that was still in use by one production LPAR
in the sysplex running my client's SAP work.  I had forgotten that
the previous month's IPL window we couldn't IPL this particular
LPAR due to quarter end close and it was "stuck" on an old set that
I decided to reuse since the other available set was already in
use as well.   

Because we already had a PDSE in the LNKLST via PROGxx SYSLIB
ahead of SYS1.LINKLIB, I managed to copy enough modules in there
and shut down LLA and keep the system up during the day until that
evening when we were able to IPL.   Good thing that library in the
LNKLST was a PDSE where taking extents don't matter (a PDSE in
the LNKLST always looks like a single extent) because commands
like "F LLA,REFRESH" and "SETPROG LNKLST" (along with many others)
were abending, so it was a catch-22 to try and repair the LNKLST

Even though it was my mistake I got many thanks from management
that I kept the system up and running with no service outage
during the day.   I also got a lot of ribbing from co-workers that
took the opportunity to call me out on a mistake.  :-)

One Mistake of Many

September 25, 2012 01:00 PM by Adam Johanson

   I wish that I could think of a specific instance where I got that "all blood in body rushing to head" feeling, but none come to mind. Like where I saw the wrong eye-catcher for a module in dump of a production IMS indicating that I moved the wrong exit, maybe goofed a command, something like that. But one mistake does keep coming to mind...

   In my previous life as an IMS sysprog, one day a bunch of messages for 1 transaction started queuing. Something had happened to the JCL for the message processing regions (MPRs) for those transactions, and it just wasn't there. So in the interest of fast return to service, rather than creating JCL for them, I thought the quickest thing would be just to assign those transactions to a class that had functioning MPRs. Well... the class that I assigned them to had a "special little STEPLIB" in their MPRs that I didn't notice and it essentially caused all messages for that transaction to be "eaten" (taken of the queue, but nothing done with them).

   Luckily, it wasn't "that important" of a transaction, just held some state-type data for the user, and the affect on the users (only green screen users) was just that they had to hit Enter an extra time. The error wasn't reported for a day or so, but after I found the problem and told my coworkers, they got quite the chuckle out of it. And to my relief, I heard a couple of times, "Yeah, listen to what I did this one time..."

Recent Stories
What Would You Do?

Happy 60th Birthday SHARE!

SHARE in Seattle Sights