To celebrate Mainframe’s 50th birthday some of our Triton Consultants share their favourite mainframe stories:
My first contact with the mainframe was as a lowly graduate Trainee Programmer at a large chocolate manufacturer. All programs were stored on punch cards; partly because they were very old code and partly to give us youngsters a sense of history. As part of my training I was slated to do a few nights of ‘Ops’, including a graveyard shift. Much of the work was collecting vast sheets of green-lined print out from the laser printers and part of it was loading programs to be read in and executed on the mainframe. One particular program was something that had been written to calculate the potential cocoa bean crop. It was a big enough program that it had to be delivered on a trolley, from the card store, down to the mainframe room and then inserted into the reader in blocks about the size of a house brick.
Well, I managed to wheel the trolley down to the mainframe room, and then sat there for about an hour feeding cards into the reader. The mainframe digested this code and went off to calculate the results, a process that was going to take about another 4 hours, so I started to wheel the trolley back out of the data centre over to the long-term card-store, through the car park. Unfortunately, as I was making my way through the car park (it was about 4 a.m. and this was back in the days when I had a social life, so I probably wasn’t at my most alert), I put one wheel in a drain and tipped the whole cart over. Punch cards everywhere.
I’d like to say that I owned up, dragged the whole lot back inside and spent hours and hours putting all the cards back in the right order. What I in fact did, was stuff the whole lot back into boxes willy-nilly, on the grounds that it was going to be 12 months before the thing needed to be run again. I doubt if the mainframe made much sense of the program the next time it saw it but, if it did succeed in reading the thing, I’m pretty certain it came up with a wildly inaccurate forecast for cocoa bean production.
Anonymous, Triton Consultant!
When I started working with mainframes in 1985, my first employer had just moved away from the use of punched cards as a means of inputting programs and data to the IBM machines (an IBM S/370 3081 I think). This meant a large number of unused punched cards suddenly became surplus to requirements, and people found some very inventive ways of making use of them – the most popular being prompt cards for speakers to use when giving presentations. I remember bringing home a stack for my mum to use as recipe cards, and she still has a few of them today.
Julian Stuhler, Triton Director
As a relative newcomer to mainframes, having missed out on the first 20 years of mainframe history, one of my most vivid recollections is relatively recent, from back in the late 1980’s whilst I was working as Computer Operator in a small Data Processing department (IT didn’t exist back in the 1980’s. It was all Data Processing).
Having years earlier consolidated our Sperry Univac and Honeywell based applications onto an IBM 4381 MVS/JES2 complete with banks of tall, shiny, blue 3380 cabinets, we had just performed our first major IBM mainframe upgrade, moving to a 3090 MVS/ESA. All had gone well and availability, particularly compared to the days of Sperry Univac and Honeywell, had improved massively.
A few months on during an unremarkable night shift, whilst determining which takeaway establishment would be lucky enough to provide us with our evening meal (one of the most important responsibilities of a shift leader), there was a ring at the delivery door. No one rang the delivery doorbell other than local kids messing about. And it was a long walk down the dark, echoing corridor from the Computer Room to the delivery door.
“It’ll be kids. Just ignore them” advised one of my team urging me to make a decision on takeaway choice by shuffling a variety of menus in front of me.
Another, longer buzz from the delivery door.
“I’d better go. Even if just to warn the kids off”
I left the strangely soothing hum of the computer room to head down the eerily silent, dark, echoing corridor. To my surprise it wasn’t kids at the delivery door but a delivery man: –
“Parcel for Data Processing Computer Room. Sign here please.”
“But we haven’t ordered anything?”
“Well it’s for this address. From The Netherlands. IBM by the looks of it”
“IBM in The Netherlands? We haven’t called IBM for anything. I’ll sign anyway”
I wandered back down the long, dark, echoing corridor to the safe haven of the Computer Room, puzzling over the mysterious parcel. The unremarkable night shift continued without further excitement or event, other than plentiful amounts of pizza, until daylight broke about 07:00 and soon after another buzz at the delivery doorbell. Someone’s in early I thought. A quick jog down the long corridor and eventually opened the delivery door, but not to someone I recognised.
“Morning. How can I help?”
“Hi. I’m Jeff the IBM engineer. I’ve come to fix one of your 3880 disk controllers. I assume the new part turned up last night from The Netherlands?”
“There’s nothing wrong with our 3880s? And we certainly didn’t raise an issue with IBM, but yes a parcel from the Netherlands turned up last night.”
“Ahh I’ll explain”
Over a cup of tea, Jeff the IBM engineer explained that the new 3090 mainframe configuration also included self-diagnosis software. What had happened was that the self-diagnosis software had identified performance degradation with one of the 3880 disk controllers, realised which component needed replacing, contacted the relevant department to order a new 3880 part from The Netherlands, and scheduled an engineer visit. Self-diagnosis had proactively identified a potential issue, preventing a possible outage. These days the response would have been ‘so what’, but nearly 25 years ago this was the stuff of science fiction, given the internet was in its infancy and unheard of to the vast majority of people. Even email was fairly new.
Often labelled as ‘legacy’ and ‘prehistoric’, even a quarter of a century ago mainframe hardware and software was right at the cutting edge of technology with innovative ways of providing high availability. And self-diagnosis, followed by part ordering and engineer booking, is still something well beyond many other of today’s operating platforms.
Paul Stoker, Triton Director
I started work on mainframes in December 1983. Access was via a terminal called a 3277, which was green characters on a black background. I had access to 2 mainframes. One ran an operating system called VM and was used for supporting the email system. The second ran MVS, the forerunner of zOS. It was used to run a system called RETAIN which was IBM’s defect support system. 30 years ago RETAIN was already 24/7 and supporting data mirroring!
You could only be signed onto 1 of the systems at a time. To sign onto the other you had to log off, physically turn a switch and log onto the other.
Nick Smith – Triton Associate Consultant