Phil,

that is a very good question, and one I struggle with sometimes. On the one 
hand, when I was a system programmer (wel, don’t we keep being that always?) we 
had strict rules that nothing impacted customer’s work, and that implied always 
doing changes at night-time and in weekends. For production, we had batch 
windows and inter-application consistent moments (running DB2 quiesce a lot).  
But on the other hand, everything was limited to the current timezone, and when 
that stopped being so, there went the batch windows out of the door (or the 
window), there was always somebody needing the system somewhere, DB2 utilities 
got ‘online’ versions. Banks nowadays close systems with large differences 
because interfaces did not run at the right moment, and in the hope that 
tomorrow everything will be fine again - I remember the day when we needed to 
stay for the weekend because the ledger had a one cents discrepancy (It was 
COBOL rounding - and one programmer that had covered errors like that a decade 
before it happened). Because of no consistent points in time anymore, the 
discrepancies run into the millions sometimes and nobody really cares it seems 
(and sometimes it really turns out to be fraud).

Yes times have changed. But I think they need to change back. It will only take 
some great disasters before it will happen. When you are accustomed to RACF and 
structured security, it is a bit odd to see youths running with Windows Admin 
and Linux root for all small changes, and really nobody understanding how safe 
a system can be when one understands the difference between user- and 
supervisor mode. Not observing that, is another prominent reason for outages of 
the unplanned kind.

Best regards,

René.


> On 21 Jan 2022, at 06:15, Phil Smith III <[email protected]> wrote:
> 
> Shmuel wrote:
> 
>> Asking "what can possible go wrong?" is good. Believing that nothing can go 
>> wrong is suicidal.
> 
> 
> 
> Indeed. Yet so many "production" systems (non-Z) don't take that approach, 
> yet get away with it. Oh, that e-commerce website is down
> for a half-hour/day/week? That helpdesk is offline because someone pulled a 
> cable (to get back to Matt's post)? No big deal.
> 
> 
> 
> I don't get it. Are we wrong? Are they wrong? It's easy to be purist, but 
> have times changed?? I like to think not, but the evidence
> seems otherwise in so many cases.
> 
> 
> ----------------------------------------------------------------------
> For IBM-MAIN subscribe / signoff / archive access instructions,
> send email to [email protected] with the message: INFO IBM-MAIN

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to