"and dedicated people called 'operators' who would terrorise misbehaving users and managers according to 'The Bastard Bible'."
Welcome back to Who, Me?, The Register's regular ramble into the dark recesses of readers' memories where we prod consciences with a long, sharp stick. Today's story, courtesy of "John", takes us back over 40 years to the end of the 1970s – a time of big iron and even bigger server rooms. John was working in a factory that …
That disk failure thing happened during my night shift as well. My co-shift wanted to move the disk to the third (out of four) disk unit we had. I strongly advised him against it, but he was very reluctant to call the systems engineer in the wee hours of the morning, so he did it anyway.
The unit failed, obviously, and now we had lost half our disk units. So the call was made, and I left that shift looking at the syseng trying to rebuild a system disk from scratch.
I didn't get any thanks out of that, but my co-shift had the balls to admit that I had advised against moving the disk, so there's that.
junior operators mounting dodgy disk-packs on multiple drive units causing mass-destruction
Had an actual DEC engineer do that at one place I worked. Loaded a diagnostics platter in one drive, wouldn't work, tried the other, game over.
His saving grace was he had two new head units delivered, installed, and whatever the original problem was resolved, before going home time.
forced run GTA IV on a lecture hall's computer, complete with projector connected.
"Computer" means a poor Intel Core 2 Quad with 2 GB of RAM (don't ask me why). It overheated, stuttered, and lagged.
No $$$$ disk packs were harmed, and the hardware is plebeian x86_64, so does that count? :-)
Meh. DOOM (the original one) at school, on one of the LCD-thingies you used to put onto a projector for transparencies for the light to shine through... doesn't count here, didn't damage anything. That time. (well, many a prank was played in the days of Windows for Workgroups 3.11 or whatever and complete lack of rights management).
Ah yes - at university in the early 90s, I'd somehow manage to wangle 24 hour access to the computer rooms (row upon row of PCs), which was unusual for a humanities undergrad. So of course, I made full use of said access by playing DOOM, in black and white, with the lights off, wired on coffee and Scotland's finest after the unions bar kicked out. After a while playing, every slight creak of the door, or footsteps outside would send my heart leaping out of chest like John Hurt's little friend... Fun times!
"unusual for a humanities undergrad"
Not necessarily that unusual. I was doing end-user support in a Humanities lab in the early 1990s. Watching the freshly minted students walk in stunned disbelief out of their bi-weekly 2 hour statistics for social scientists 'lab' was highly amusing. The sociology researchers at $work made _heavy_ use of regional supercomputers; more so than the scientists.
RM was great, reboot pc, hit f8, select service mode from there customised boot menu, get windows without the RM crap in the way, for more lols, be best mates the son the lady in charge of the schools computers and scare the crap out of her by remotely ejecting all the cd's out of the network attached cd rom array (used to share encarta and other multimejuh cd roms when elearning was one step up from watching bbc school programs), or scare the bible bashing teacher with netsend, proclaiming "I Satan have possessed this computer, your soul in now mine"
Are you referring to the security solution that closed all of the vulnerabilities in windows 9x? It disabled things such as "that other button on the mouse" which apparently had no legitimate use...
When I was at school their RM security solution thing was such a tempting challenge that I must have spent minutes working on the obscure kernel exploit I found to get around it. I discovered that if I triggered a certain interrupt it permitted a further exploit which I could use to fully bypass all protection mechansims. I called it the "CTRL+ALT+DEL - End Task" bug.
It was a great scheme. It meant when you came in on a Monday morning find your terminal was dead you could wonder into the computer room, notice the red light on the first drive and the pile of paper at the back of the console teletype. Power off the Vax.
And then for the good bit, you pulled the little plastic drive number light off the first drive, slotted it into the second drive and boot up to single user mode. Back the he root FS was small enough to keep a copy on all the drives so it was easy to get the machine up far enough to start reading the backup tapes.
As mentioned in the article DECUS was a good source of such things, we used to take a couple of spare tapes along to meetings "just in case".
On a more useful level, with PCs just starting to come into use, we acquired a copy of Kermit* for the VAX & PDPs at one of the user meetings. Very handy for transferring data between systems.
*For the youngsters it was an extremely useful cross-platform file transfer protocol.
100 upvotes for mention of Kermit.
it was an
extremely useful essential cross-platform file transfer protocol
Not just file transfer but terminal emulation as well. When all you had was a serial interface Kermit was the equivalent of FTP, Telnet, rcp, ssh or whatever your favourite weapon might be these days.
The last time I heard of Kermit was when uploading "programming" on HP48G's calculators.
The kind of programming that is considered "mobile gaming" these days, with an Italian-language Poker and... both Columns and Tetris. (God I miss that 48G keyboard.)
But sorry, nothing on mainframe level.
And I know of a story about Quake (or Doom?) being run every weekend, that was caught due to some .DLL being overwritten every night by the damaged version on game's install disk, but that wasn't my story. Anyway, DLL's were overwritten and gamers caught.
Yes, Kermit was a godsend in the days of dial-up.
For a while in the early '90s I had a remote office with a POTS connection to the main office. Most of my connectivity requirements were file transfer and email, for which I used UUCP - the modems at each end were Telebit Trailblazers with UUCP spoofing, so that made sense. But I built Kermit on machines at both ends and used it for a variety of ad hoc purposes because it was just so versatile.
Eventually we got a 56K dedicated line and I could just SLIP up an IP connection between the sites, but UUCP and Kermit helped me get a lot of work done before that.
cKermit was very nice, and the source was available, but the authors kept a tight control over it and eventaully it died because it wasn't nearly as useful any more. I've still got "Using C-Kermit Communication Software. Second Edition" on my shelf at work, since you never know when you might need to reference some esoterica from the bible of goodness when trying to manage some bogosity over a serial link that's been bodged together.
It was great stuff because it ran the same pretty much everywhere on any form of Unix and Windows in it's day. So you knew you could get data back and forth without compromise. Usually.
Haven't touched it in years, and can't for the life of me remember what I last used it for. But I'm sure it was to automate some stupid piece of junk serial controller hardware.
There were a number of documented cases where games had identified actual faults that the official diagnostics either did not detect or could not locate accurately. Not only were these well known to DEC engineers, but also to most customers.
In those days, the users were a relatively small number and talked amongst themselves.
[Speaking as a former designer of disk controllers, although not for DEC]
Back in the 70s I was working at a national daily paper, whose name included a synonym for "reflective surface"
Our pre-press system had two 60Mb washing machines. Quite apart from dust these are sensitive to vibration.
Colleague and I looked into the machine room through the window in the door.
Two carpenters. One using a drive as a saw bench to cut a 4x2 down. The other using the second drive as a rest to knock old nails out of another piece of wood.
We looked at each other and silently went elsewhere.
The drives survived.
They later survived rain getting in the electronics.
CDC made damn good hardware!
"CDC made decent kit when they actually bothered to make it. They were famously pushed into administration when an over-eager auditor checked that the vast stock of disk drives boxes actually contained drives only to discover they were full of bricks."
You're thinking of MiniScribe, I think.
Thanks for the link! Somehow the whole MiniScribe thing didn't hit my radar. I note that the article says "The company declared bankruptcy on 1 January 1990, and was quickly liquidated with Maxtor acquiring its assets in a $46 million cash and stock deal"
I wonder what Maxtor built with the bricks...
Wasn't just CDC. Miniscribe went down the same way back in the late 80's, but they did it with style. Hand packing 26,000 bricks into hard drive boxes and shipping them back and forth from Singapore to Colorado was just one of the shady things they did.
I worked on one mainframe installation where the machine failed its first inspection for 'saw marks on the upper side of the cabinet' It was a desk height machine and a carpenter had used it as a saw bench.
)This was not covered buy the warranty or hardware maintenance :)
Two carpenters. One using a drive as a saw bench to cut a 4x2 down. The other using the second drive as a rest to knock old nails out of another piece of wood. We looked at each other and silently went elsewhere.
People may be wondering why they didn't warn the carpenters that they could damage the disk drives and ask them politely to work elsewhere. Remember this was a national newspaper in the 1970s., and the carpenters were surely members of SOGAT or NATSOPA or the NGA. Any suggestion that there might be a better way of doing things would probably result in no paper the following morning.
(I worked at Times Newspapers in the 1970s, before the 12-month shutdown and the move to Wapping.)
Brrrppp - brrrrp. The sound of an industrial masonry drill cutting into the concrete ceiling in the mainframe computer room during prime shift. The workman was drilling vertically - standing on a stepladder lodged against the 8MB exchangeable disk units. The perspex covers of which were now buried under a thick layer of debris. The senior engineer interrupted him - and explained that even a fragment of cigarette ash was enough to damage a disk. "uh uh" Brrrp - brrrrpppppppp.
And if the whole team hadn't colluded and cooperated to keep the Junior out of the storm that would probably have been an option.
Cases like this is really a matter of "What management doesn't know can't hurt us (or them)". And since they care about being able to run the factory (and the team made that happen) and being able to show a shiny computer room (and they could), little details like how exactly 6 drives happen to fail on the same night don't really matter to PHBs.
Junior and the team got lucky he didn't try a good (production) disk pack in a failed drive after borking it though!
Yeah, unfortunately that’s often the case. Ruining someone’s who has just received a very expensive lesson career is just plain stupid. In this specific story, the junior operator taught himself to think twice when figuring out what’s wrong with a non-functioning device.
Fear of getting sacked for screwing up only leads to covering up tracks, which leads to even bigger problems, increased expenses, bad work moral and even mental ill-being. Fuck such management.
I have the same situation with newly designed circuit boards. If a prototype batch of boards comes back from the manufacturing process, and the first board tested doesn't work correctly when you power it up, do not, under any circumstances, keep trying other boards in the batch in the hope that they will work. If, for example, the bill of materials has the wrong resistor value for the feedback on your DC ⇾ DC switch mode power supplies*, you stand a good chance of destroying the whole lot. At least the modern pick and place machines test the discrete parts' values these days. Stops the problem when someone loads the wrong tape of resistors...
Oh, and make sure that the designers are the first to power up the boards, not the manufacturing team. Because they _will_ power up every damn board.
*Of course, the first thing you do on powering up the board is to measure the rails...
That reminded me of the time when an operator plugged a 3270 terminal in, switched it on and it started emitting smoke. I could just about forgive him for destroying a second terminal by plugging it in to the same socket (after all I'd probably have assumed it was the terminal that was faulty) but not for the third one.
It turned out some dodgy wiring in the factory had resulted in the neutral being connected to one of the phases on the 3-phase supply instead of to where it should have been,
We had an electrician swapping a phase with neutral in the three phase socket for our cooker. I did find the light in the oven blinding, but pans with water boiled in no time. That was overclocking
Electrician fixed it, and paid for a new fan for the oven
Mainframes went to the Systems Test floor for final assembly before being shipped. The area's three phase had been wired with the wrong colours arrangement - but there was never a point in the busy schedules when it could be rectified. Not a problem - all the engineers knew which colour was what.
Then one day the prototype mainframe received its 8MB exchangeable disk units. They were shipped in from another of the company's factories - and union pressure meant only that factory's people were allowed to wire them to the power. This was done overnight. Next day the drives were powered up and the disk OS primed from tape. It ran ok.
A few days later there was some head-scratching. The system disks had been moved to a second mainframe - and the OS would not even load. It took a little while before it was spotted that the new drives were rotating in the wrong direction. Fortunately the heads were symmetrical and flew ok in either direction. Why the disk didn't unscrew itself was a mystery.
Unlike another site where a misconnected card punch rotated backwards with much grinding and damage.
Finally (for now - out of an almost inexhaustible supply):
A customer had ordered the smallest mainframe of the newly developed range. Unfortunately the over-ambitious development cycle meant the range kept getting pruned of the lower models. That customer's order was upgraded several times for the original price - until they were to be given the current highest spec one. Unfortunately there weren't enough available to fulfil the customer's now long overdue order.
That model was a near-clone of a US partner's machine - so one of those was shipped from the USA. The engineers successfully commissioned it on the customer site. The customer was delighted with the bargain they were getting. The engineers left for the night - leaving the customer's electricians to incorporate the jury-rigged US/UK autotransformer into their structure.
Next morning the powered up the mainframe - and the autotransformer had now been wired-in the wrong way round.
Oh, what fun there was when some idiot installed a reel of Tantalum capacitors backwards in a robotic assembly machine. There were hundreds of boards built with the capacitors installed backwards. Oh, and Tantalum capacitors are polarized, so if they're installed backwards, they consume way too much current, overheat, and explode quite violently. The test room sounded like a small war was going on. There were literally hundreds of boards made that way before the first one made its way to the test facility. Whoopsie!
It's impossible to install a full reel backwards on a pick-and-place. For larger tape widths (16 or 20mm and up), it's possible to load cut tape backwards. If you build 100s of boards, you may likely have used cut tape instead of full reels.
I have seen cut tape spooled "backwards" when delivered. That's not surprising, since the distributor would have counted the tape off of a bigger reel (5K-20K parts).
Also posible the SMT programmer was too trusting of the input CAD data and didn't verify rotations of polarized parts. Every CAD package and every designer seems to have their own convention.
Mine's the one with the pickets full of OS/2 install disks for the old GSM.
"There were hundreds of boards built with the capacitors installed backwards. "
I built my Motorola 6800 evaluation kit board without realising that tantalum capacitors were polarised. It looked quite pretty when it was finished - all the capacitors aesthetically orientated in the same direction. I later discovered that by chance I had mounted the first one the electrical right way round.
What was a real fluke was the final two. They were positioned 90 degrees to the rest - and they were the right way round too.
Dry Tantalum capacitors can be satisfyingly pyrotechnic even when installed in the correct way. When they go through reflow, it can damage the dielectric (particularly for larger parts in case size C and bigger) which leaves a small short circuit in the device - if it is powered up from a low impedance source (such as a power supply), that short circuit becomes literally white hot and leads to thermal runaway.
That is known quite charmingly as an ignition failure and tends to leave a device where the solder terminals and outer case (sans middle area) are present along with a grey blob from what was the internal tantalum
Had that happen quite a few times in a new flight control computer (I won't say for what aircraft).
Wet tantalums are also pretty renowned for their rather impressive failure mode; there is a 'self-healing' mechanism within the device that is not so much self healing as building up a future failure as the device ends up with a lot of free hydrogen internally and the internal pressure builds up slowly over time to the point the internal tantalum cathode slug is forcefully ejected by a very hot internal event. (The internal construction includes dilute sulphuric acid.)
These have nickel leads spot welded to the terminals; one failure had vapourised the nickel - considering that nickel melts (let alone vapourise) at around 1500C it is not surprising that the internal wall of the aluminium box it was in when it failed had significant scorch marks.
Of course, the first thing you do on powering up the board is to measure the rails.
I worked for Xerox UK in the 1970's. We had a new test rig delivered. Some military type had replaced our tradition of using colour codes to indicate the use of a wire (red = +5, black=0V, green = -5, etc) with a system in which the colour represents the revision level (pink= original, while= revision 1, etc).
The new test rig (1.5m x 1.5m x 2m high) had a wiring error (the +5 wires and 240V AC wires were switched). When switched on, all the ICs in the pie-crust (died-cast) metal boxes started popping like popcorn.
At this very moment, the alarm went for the Friday fire practice. Xerox had a very good record for evacuating rapidly. Even before evacuation was complete, the smoke was filling the room. Someone called 999 and asked for "Fire". The switchboard explained that it was a practice - they had already been notified. The staff explained to the fire brigade that there really was a fire. <this exchange continued for some time>
Fortunately, the fire was confined to the test rig, which was 2 metres from anything else.
No buildings were destroyed in conducting this experiment.
Back in the early 70s, I was working as a videotape engineer at a big broadcasting company. The VT machines in those days were monstrous beasts that cost 3 times the price of a modest house, so only those with a degree or equivalent were allowed near them! These machines were incapable of playing slow motion, due to the way the information was recorded on tape, but the very clever engineers at Ampex (the major supplier of VT machines back then) had made a wonderful hybrid system. This basically took the signal system from a quadruplex VTR and married it to a computer hard disk system - not unlike the "washing machines" mentioned above! There were two disks (hence four surfaces) spinning at around 3000 rpm. The heads were controlled by stepper motors, and each recorded one field of TV information (two fields = 1 frame, due to interlacing). The heads would move inwards in sequence creating circular tracks of video as they went, and writing even numbered "circles" in one direction, and when they reached the edge of the disk, reverse direction and write the "odd" circles.
The disks were capable of storing around 30-something seconds of analogue video, and recorded continuously, until (say) a goal was scored! The studio director would then order the recording to stop and replay the goal, which the operator could do at any speed, forwards or backwards!
Very clever, and all built using "off the shelf" components!
The picture quality wasn't brilliant (analogue, remember) due to the heads "flying" a tiny fraction of an inch of the surface of the disk, but at the time it was the best that could be achieved, and it continued in use for many years.
There was, however, a minor design flaw that occasionally led to catastrophic failures!
The spinning disks were at the very top of a cabinet of electronics around 4 feet tall. They were heavy, and spinning very fast! This led to some interesting gyroscopic effects, especially as the whole thing was sitting on a not-very-rigid computer floor, with all the cabling underneath!
If any heavy-footed individual entered the room while the disks were spinning, this was likely to lead to the whole thing wobbling slightly, and crashing the heads into the disks! This was a not completely uncommon occurrence!
The disks themselves were quite striking in appearance, and the damage was often invisible to the naked eye. This made them highly prized as trophies, when no longer fit for original purpose. Every Christmas, engineers throughout the TV networks would compete for the "Christmas Tape" trophy, a sort of Eurovision Song contest for who could create the best out-take / humorous (and often obscene) tape! The trophy was a suitably engraved disk from a slo-mo machine!
Have a look here for more information on this rather unusual use of computer disk drives: http://www.vtoldboys.com/editingmuseum/hs100.htm
are you saying it was an analog signal being written to & read from the disks?
Back in the 1970s digitising that amount of data would have been rather difficult. I had always assumed (or maybe I read it somewhere once) that the clever thing was that the FM signal sent to the disc packs was the same as the FM signal recorded on the Quad video tape so in essence all you are doing is swapping a magnetic medium rotating one way for a magnetic medium rotating a different way.
Don't forget that the video signal (and the audio on earlier discs) on Laserdisc was also analogue. Today we associate optical discs with digital data, but while Laserdiscs were built in the same way as CDs, the pulses didn't produce a stream of digital data, they produced a PWM signal that was turned directly into analogue video.
Yes, that's absolutely correct. And that was the clever bit about Ampex' solution - they just bolted the electronics from a quad VTR to a spinning disk, and voila!
The FM signal recovered from disk was never quite as clean as it was from tape. On a tape machine, the heads were in physical contact with the tape, whereas on disk there was a gap, and this was why the quality never quite matched up! The tape machines were very sensitive to tape damage. The four heads were mounted on a 2" diameter drum, spinning at around 3000 rpm (in Europe). It was a bit like a circular saw, and the slightest damage to the edge of the tape would see it shredded! This occasionally happened when the machine was on the air, and was the cause of much frantic cleaning up and re-lacing to get it back to transmission!
Mercifully, it never happened to me, but I did witness it, and occasionally had to repair the tapes afterwards!
And around this time, I remember the research department (a load of boffins at an old country manor house) being delighted at having achieved the feat of successfully recording and playing back *half* of the test card from a digital tape deck!
How times have moved on!
As an aside, the quad videotape machines had a switch on the input marked "video" and "data". The word on the grapevine was that they were used by the CIA in their larger spyplanes (converted airliners) as they were the only things capable of handling the vast throughput of data at the time. It also explained why our cousins across the pond were paranoid about anyone behind the Iron Curtain getting their hands on one and copying it!
Of course, the Russians did eventually manage to copy them, but just as with Concordski, they missed some of the vital design details, and recordings made on Russian machines were incredibly difficult to get to playback properly, despite them using the SECAM colour system, which was much more robust in these situations!
Yeah, sure, if you ignore all the pre-work done since the 50s by the French and British. Plus the fact the basic outline of the aircraft had been on paper by 1961 already. Tu-144 development started in 1962.
So no, Tu-144 was NOT proposed before Concorde, it was NOT designed before Concorde. It did fly before Concorde (according to some reports mainly because it was rushed, not because the plane was actually ready). It was also a rather crude design that was noisy, uncomfortable and decidedly unsafe.
The undercarriage was certainly a problem for Concorde. They had to devise all sorts of solutions - to the problems of oscillations in the very long legs and the problem of stopping the 'plane without resorting to a parachute - Concorde was the first aircraft (the first anything?) with anti-lock brakes. I believe the Tu 144 used a parachute?
Concorde solved the problem of slowing the air into the engines not with simply a "very subtle design" but with an extremely clever set of vanes that operated automatically according to the airspeed, allowing air straight through in low-speed flight, but slowing (and thus compressing) the air in high speed flight. Concorde's engines were the first "fly by wire" engines too, I think, with many of the aircraft's systems controlled by a very clever analogue computer.
A lot of the development was done in parallel with the TSR-2 and you could argue that some of Concorde's cost over-runs were because the TSR-2 project was cancelled, thus depriving Concorde of a lot of valuable development work funded from defence budgets.
I have an excellent book (in a box somewhere due to being part-way through a house move), Celebrating Concorde by Reg Turnill who was the BBC's air and space correspondent for many years, including those of the development of Concorde. It's what might be termed a "contemporary retrospective", as it were. Published in 1994, it was close enough to the development of the aircraft that Reg's personal stories are all still vivid, but far enough into the commercial operation of the 'plane that there is some amount of distance and evaluation possible.
For an even more personal story, I can highly recommend Jim Lesurf's memories of chasing the 1973 solar eclipse in Concorde 001. In fact, all of Jim's personal website is a fascinating and informative read.
The Russians certainly gained access to a lot of Concorde's design details, but they missed some very important ones. And remember that the Brits had been working on delta winged supersonic aircraft during the 50s (The Fairey Delta 2 was the first aircraft to break 1000mph, in the mid 50s!)
IIRC (and its been a long time since I studied aerodynamics!), one of the problems of supersonic flight is slowing the air to sub-sonic speeds before it enters the engines. If you don't do this, the engines either stall, or become horrendously inefficient! Concorde had a very subtle design of its air intakes, that provided a passive method of slowing the air in supersonic flight. This mean that whilst Concorde needed afterburners (noisy and thirsty) to break the sound barrier, once it was through, the afterburners could be turned off.
The Russians missed this, and Concordski needed the afterburners running all the time it was supersonic. This had a severely detrimental effect on both its fuel consumption and its noise emissions!
The Russian 'plane was only ever used on a trans-Siberian route as a result, flying over a largely unpopulated region. Even government subsidies were unable to compensate for the horrendous fuel consumption, and it was withdrawn from service after a short period.
"The Russians certainly gained access to a lot of Concorde's design details, but they missed some very important ones"
AIUI it was realised the Russians were snooping and were deliberately fed a couple of wrong design details to put them off the scent - particularly related to the downward curve of the leading edge - which is why the TU144 required those ears
I was quickly scanning through posts, stopping at the eye-catching "TU-144" in yours, and started to read it from its beginning.
By that time, my brain had already processed "TU-nnn", "Russian...copy" and the article context together, and provided its analysis: DEC TU series 1/2" tape drive, unknown model number. Possibly Russian-made to go with the VAX 11/780 they copied.
I remember watching a football game and they were replaying a play in slow motion. As I watched horizontal bands of white noise appeared on the image. As they moved to successive frames the picture was good for a second and the same bands appeared. I was watching a head crash in real time!
"and the damage was often invisible to the naked eye."
In 1970 we had a mainframe 600MB behemoth - weighing over a tonne - and had water cooled bearings. The platters were each about 50cm in diameter. Initially head crashes were a too common occurrence. They would create a wide circular stripe of bright metal against the brown background.
Back (Oh, so far, far back!) in the day, our BBC Cymru effort was generally called, IIRC, the "Snow White Christmas" tape, and for a few years featured various bemused celebs and public figures holding a BBC polystyrene cup with the legend "VT Tea".
We even had Margaret Thatcher one year.
Happy (if somewhat anarchic) days!
Place I worked at many, many moons ago, there was one particular EDS-200 disk pack we abused the crap out of and the damned thing never failed. RAC033 - Director's pay-roll. Still remember it now,
And the error code for a failed (64Mb) disk - E11011 I seem to recall. Usually a snapped drive belt. Size of a large washing machine. The whole disk pack had to be pulled out and adjusted to get the belt tension right.
The engineers were great guys - taught us what we could get away with (like springing the lid on an EDS-200 and using a cloth to spin the disk down when the brakes failed - yes, these things had brakes)!
I had a room full of 80Mb five-platter CDC drives attached to a variety of PDP-11's, and another room with a somewhat more up to date VAX 11-750 and a bunch of MicroVAXs, with fixed disks. I spent most of my day feeding various paper stocks into various large printers, and responding to user requests to load pack X into drive Y, eventually awaiting the signal at 5pm that the day upstairs was done, and I could start the daily disk pack suffle that was the backup process.
Users felt they were in control, but I'd get the occasional stupid request. There was one system with two drives, and one head developer who used it remotely at 300 baud. Load pack Y4 into drive 2. Yeah sure, but no drive 2, only drives 0 & 1. So I dismounted and spun down drive 1, loaded Y4 in place of Y1 in it, spun it up, mounted it, andtold this prima-donna that it's available. Same request. So I reply 'in 1, no 2'. As he's using his phone line to remotely work, phone calls and the simple text messaging set up can't work at the same time. The phone rings, demanding to know why I have used drive 1 instead of the non-existent drive 2. Demanding also that I reload Y1 back into drive 1 immediately. So I do that. Next phone call, wheres Y4? I still need it. Look mate, make your mind up!. So he demands I put it in drive 0. Did Y4 have an operating system on it? Suffice it to say I didn't get any more requests of that system for the rest of the day while the system wondered where it's system disk went,and the user concerned learnt to count from 0, not 1.
It didn't end there, the next day, we then got an order to move a drive from the production system onto 'his' development system. Not a simple job because the system concerned only had controllers and/or cabling for two drives, and power supply was another issue. We sent the job back with those details, and waited for the go-ahead from the boss, which never came. Apparently the guy thrrew a hissy fit and said something unfortunate to the boss, who back in those days knew all the technical ins and outs of the system. Various things were said, and no further requests came from that user, ever again.
Never had this problem with the VAXen in the other room. The users, of which there were far more, generally only wanted the occasional tape mounted, and the users knew how to request volumes properly, and even made sure no two tapes had the same label. (Such levels of clue are unusual!) Requests came through OPCOM, and the system would assign devices to users and all I'd have to do is physically load the volume and bring the tape device concerned online, and the system would take care of the rest, with only the occasional need to resort to REPLY/PENDING=requestno "Where the heck is that tape?". There was clever stuff built in too, such as when you'd failed to find the tape the user wanted in MSA0 because it was already loaded into MSB0, you could REPLY/TO=requestno "SUBSITITUTE MSB0" and the user's request would complete, and if they'd followed the rules and specified a logical name when mount the volume, unless they were paying close attention, they'd be none the wiser. Somehow a different class of user. Different class of OS too, as we'd occasionally get requests to mount a TK50 cartridge in a large magtape drive, and we'd just use the same method to switch the system's attention to MUA0: instead of MSxx:, usually follwed by by a REPLY/USER=DIMWIT "Remember, your new system's data is on TK50's, please use MUA0 next time", and sometimes they did. Just shows that regardless of how clued your users seem to be, give them a choice, and they will make the wrong one.
Yeah, loved working on the VAX. Ours was an 11/780. VMS was a sweet joy after working on JCL and HASP.
We did have to satisfy one user who point blank refused to have a terminal. He only used punch cards. So I think we had one of the few VAXen with a card reader and card punch, and still had to retain a very noisy manual card punch.
Never had the fun of punched cards with the VAXen, although I note that the SimH installation I have on my laptop emulates a CR11 as a text file. That's one feature I've never bothered to explore.
When I was growing up, the kitchen drawers were full of boxes of 80 and 90 column cards which my mum mostly used for shopping lists. Some of them probably survive as separators for different batches of slides somewhere in my dads loft. We never ran out of drawing paper as kids, any amount of punched cards, boxes of greenbar printer paper discarded after someone send a binary file to the lineprinter, etc. were in plentiful supply.
"[...] with a card reader and card punch, [...]"
The bootstrap on our mainframe was a punched card. As it was binary the card was more holes than paper. In a high speed reader it was inevitable that there would be occasions when it jammed, concertinaed, and tore.
Not a problem as we had working duplicates - made from a submaster on a slow punched card data preparation machine. Once in a while a high speed reader misadjustment crumpled all the duplicates - and someone in a hurry used the submasters. When they were destroyed - they abused the master itself.
It then required a painstaking manual operation to transcribe the tattered mess onto a new card using a hand punch. Once proven it was immediately used to create new submasters and working copies.
Whilst I can't recount a story as good as this (although stopping every train in a 25 mile radius of York station for 15 minutes isn't a bad one) I did have the misfortune to visit a Cable TV franchise in Bristol back in the day when I was working on some less than reliable billing software. Had to work through the night on persuading the billing run run to complete - while the operator watched a non-stop stream of p*rn on one of the fully unlocked set top boxes that were scattered around the back offices.
He didn't try and engage in conversation at any point; I wasn't inclined to try. Fortunately (albeit somewhat surprisingly has he was the overnight sysop) there was no reason to interrupt him.
I used to work in a computer lab where a student did the same.. I encountered him while I was patrolling. In the corner, watching some pr0n, box of tissues on the table. This wasn't a small lab either (over 30 machines, a fair few of which were in use).
Think he got suspended.
I remember one year working for a company who built flight simulators (at the time mostly military, but some civilian). They had a big IBM mainframe that ran all sorts of stuff for the company, both technical applications as well as financial and management stuff. One day the shift operator went into the next room to find the tapes that had to be mounted in preparation for that night's backup; just after he went in the room there has a very loud "bang" followed by lots of noise as of broken glass bouncing off things. When he went back into the main server room he was greated by the sight of the mainframe crashing, and one of the "washing machine" disk drives with a hole in its side, with the metal bent outwards.
It turned out that the lower bearings on the disk spindle had failed, resulting in the spindle tearing free and (along with the disk drives) deciding to take a strolling around the server room. Odds are that had the operator been in the room at the time he would probably have caught some of the debris that was flying around at great velocity.
Took IBM 3 days to haul in a new disk drive unit, wire it up and then rebuild the system from scratch and backups.
Apart from electrocution and hearing loss it must be a while since a computer room was a hazardous environment. I remember having to unrack a Sun V880. Two of use tried to get a feel for it we could do it and decided not. Two burly lads from Estates were roped in and they started. It was most of the way out on its runners when they realised that they should have pulled the support legs out. No bones were broken but I shat a brick.
I heard a variation on that one where a field engineer was checking a drive. So cover open and a <crack> as the spindle broke.. And then the disk pack slowly rising up, then clipping the cover and spinning off to embed itself in the side of a mainframe cabinet.
My mum was an operator in the 1960s. At the time if you mounted a tape wobbly it could eventually work itself loose, throw itself across a room, and embed itself in a concrete wall.
Those IBM mass storage units? Magnetic drums loaded by a very fast robotic arm? They were enclosed in plastic walls after a South American operator lost his head by sticking it in the robot's path.
In the early 70s we had a Singer System 10 with split disks (model 41 & 42 ??). The top half (2 platters) unscrewed and could be replaced by a different top half, making it easy to move data around. The bottom half was also 2 platters. The top cover looked like a normal cake cover, but the top disk had a small pin that engaged with a small hole on the bottom disk.
When screwing a top half onto a bottom half, it was quite easy to misalign the pin and hole, resulting in a lop-sided set of platters which the heads didn't like at all.
Even though we used drums, it was a couple of years before my time, so I never saw one, but I heard one good story at a DECUS symposium, as told by Bruce Ellis or maybe some other VAX/VMS guru of the time (faded memory).
Seems that these big heavy drums that were normally spinning pretty fast could be physically locked so they couldn't move during transport.
The story was that some poor unfortunate managed to engage the lock while the drum was still spinning. At this point, sufficient force was available and was instantly transferred, sending the whole unit crashing through the cinder-block wall behind it. Believe it or not.
A couple of memories: Laser-Scan was frequently visited by VIPs (Prince Philip, Margaret Thatcher to name but two). The server room was cramped, but the dignitaries were always shown around: "VAX 11-780 with half a megabyte of RAM!". Running in shouting "which *** pushed the read-only switch on the drive" after someone had butt-pressed it could be a career limiting move.
Also, if you found a pack on the floor, you didn't just assume it was untidiness: it might have been left on top of the "washing machine" and the vibration could have shaken it off, disturbing the alignment of the platters. Mounting it (in the days when that was a physical action) would almost certainly lead to the destruction of the heads.
I worked their one summer whilst at Uni. Think that post-dated the VAX era, but doing fun things with DEC Alphas. Was a fascinating experience seeing how a company went from an idea to digitise particle traces to GIS and bank note design. And trying (but failing) to abuse their hologram creating room. Amazing what could be done with mirrors. And holograms of mirrors. One memorable moment was having to enter via the back door because a very important customer was busy poring over a huuuge plot of their new bank note design. And by huge, as in it took up most of the floorspace in the reception hall.
I've managed a somewhat similar error, but with modern 3.5" disks in my own PC.
I had a hotswap bay in the front of my case, containing some disks that were part of a RAID. I upgraded my motherboard, and assumed that everything would work just fine with the new SATA3 controller (old motherboard was SATA2).
As you've probably already guessed, it didn't behave, and started silently corrupting reads and writes. To try and diagnose the problem, I started swapping disks around ("hmm, maybe it's the cable?"), which made sure that every single disk in the array was equally f'ed up.
Fortunately I had a backup of everything that was irreplaceable, but it took the better part of a week to fix (and I had to throw an otherwise perfectly good hotswap bay out, a Zalman ZM-HDR1 for those keeping score at home).
I was working for a UK mainframe manufacturer when I received a call for an internal system. It wouldn't IPL after being patched.
I couldn't get a remote connection to the operating system so used the remote diagnostic computer inside the cabinet.
This basically told me that there was no fibre network disc controllers etc. almost as if all the connections were unplugged.
These were checked by the operator and confirmed that they were in place. 'Its software' he was screaming down the phone,' what have you done' I responded 'it cant be software as there are no disk drives connected'.
He finally confessed that they had needed to relocate a mainframe but didn't want the budget hit of the cost of an engineer and trollies so just used a forklift.
"Did at any point the cabinet tilt off the forklift" I asked. Sure enough not only had they decided to do a DIY move they had dropped the bloody thing. A colleague was a hardware engineer so he 'popped over the road' tot he labs and found that all the boards in the processor cabinet had been unseated and were piled on top of each other in the end of the cabinet.Having warned them that there was no way this thing would ever boot again we were amazed that having reseated all the board unbend ad refitted a few connectors it booted normally, I think we closed the call as an intermittent IPL error and nothing more was ever said.
An overseas customer received a long awaited memory upgrade. This was the 1970s when the nearly 2m high cabinet contained maybe 128KB of "compact" plated wire ferrite memory.
It was noticed that there was a long mark in the paint across one side. Opening the door - the board cage frame had turned from a rectangle into a slanted parallelogram. The frame's top bolts had sheared when presumably the case had toppled over and hit the edge of something - hence the mark. The customer had to wait a while longer for the replacement.
PDP 11/70's 1980 something (86? 87?). I was a Senior Operator at the time. We'd had a disk with suspected head crash on night shift - by suspected I mean it was definitely a head crash - you could see the scratches on the platters and the plasticy/metallically burning smell at head crash time was unmistakable. It was a bad one. There was a whole bunch of important work that couldn't be done but and - from the looks of it - the data was lost with no backup (if memory serves we could re-create the data but it would take a while and there was a procedure which us oppos where not privy to)
Anyhoo, we labelled up the disk and the 'washing machine' as : 'do not use, head crash' and told the boss in the morning during shift-handover time. He went spare, called us stupid, and proceeded to load the disk into each and every spare drive we had (some were system disks and couldn't be used) but Lewis - not his real name - was on a mission and could NOT be stopped. We tried pleading "Lewis, no, stop it", the day shift joined us in unison. "STOP IT!" I mean you could HEAR drive heads hitting platters like a high-pitched 'pertttthhhhhffff' sound but he trashed those disks like godzilla throwing a fit in downtown tokyo. Finally, with all the spare drives now knackered he stormed off without saying a word. It was left to us to call the engineers. DEC weren't happy and argued the toss for ages about wether this constituted missuse of equipmnt. I think we had to pay up, Lewis kept his job. happy days
Re: My boss did it
How could anyone's tenure survive such an insane reaction? We would have tied and gagged the boss (maybe put him under a tile somewhere) and waited for the nice men from the mental hospital to come and fit him with a new jacket. This and more would have been easily justifiable under the circumstances. Tackle the bugger before he crashes another!
["They had lost 10 drives to a zealous student operator trying to load a drive full of ASCII pr0n pictures."]
*Kid cannot afford pr0n mag*
*Kids causes many dollars of damage to try to see pixel pr0n*
A Nurd is what they called a tech nerd or geek back in the 70s.
I've had similar happen with floppies, and a friend had it happen with ZIP, but it usually was the media being destroyed by the drive. Randomly.
Even had it happen with CD's, when a drive had foreign matter in it, and started scratching everything.
And now I'm reminded of the anecdotal story of a modern day kid in a 2000's era car complaining the phone dock scratched his screen, after shoving the thing in the cassette deck.
Beer for what I have to drink when a 128Gb μSD card goes "read only", "write only", or "I think it fell down the sofa...".
Their malfunction report will read something like:
"As the drive span(sic) up, it stopped working."
The complete lack of awareness of cause and effect - demonstrated by moving the faulty disk pack from drive to drive - is almost as astounding as the idea the bean counters wouldn't go to General Quarters over the near simultaneous failure of several thousands of dollars worth of separate pieces of equipment.
All because Sysop Jr feared a tongue lashing for not having games ready to play? Gimme a break!
This tale smells suspiciously like Limburger Cheese gone bad.
I don't think I have ever actually changed one of those disc packs; but even I know from somewhere that a faulty one can damage the drive that's trying to read it.
In any case, how many very expensive pieces of equipment that definitely used to work up until a few moments ago need you try a removable thing in and find none of them work, before formulating the hypothesis that the removable thing -- which has been a common factor in all the failures -- might have been what's been killing them all?
We used a couple of Honeywell 'washing machines' in the early 80's. The factory was an iron foundry and *very* dirty due to the grinding of cast iron spreading graphite everywhere. Our procedure was to manually pull the heads up the loading ramps and clean the heads with alcohol every day. Woebetide him who pulled too far and the head all crashed together. This did happen but very rarely. The result was that we never suffered head crashed unlike a nearby site situated alongside lovely fields, no dust & no problems until they opened the windows when the rape seed was in flower, let in the pollen and overloaded the filters. To our amusement :-)
However there was a rule: never, never never put a disk in its case on the top of the disk drives. With the shaking in use it would fall off. Never, never, never means of course this did happen. The disk contained all the source code. There was a backup but not recent and the damaged" pack had a bent top platter a few mm out of true. "Please, please Mr Engineer can you help?". And he did, with us ready to press enter on a copy program while he manually loaded the heads. Sighs of relief !!
Not many talk about VAXes these days. Fond memories for me. People don't say "disk packs" too much anymore. Those washing machines were wonderful. Malfunctioning top-cover interlock switches that allowed you to open the lid while the platters were fully spun up. CDC disk pack inspection tools with their cool mirror-comb assemblies. When you set the tool case down and opened it, onlookers were amazed and intrigued. It was like opening some secret agent's or hit man's tool bag.
The best WM story? Someone at the firm had written a backup/restore program that was loaded into the PDP or TI computer (rows of mini-toggle switches and lights up at the front) from a compact audio cassette. One day, the program entered an infinate looping state. Smoke was seen to be emanating from the CDC drive (around 10 platters/300MB per disk pack). The heavy head movement solenoid was going from limit to limit once every second, and the whole thing had slowly "walked" across the raised floor until it reached the end of its tether (power, data cables), and luckily, did not fall off the edge onto the concrete floor below. Nowadays, everyone has portable disk drives. We were on the cutting edge.
In the 70s (and into the 80s before PCs and "winchester disks" become a thing that got stuffed into toy computers called PCs) I spent many happy hours looking after ICL and CDC disk packs. Maybe I was erm "lucky" but invariably a head crash would be preceded by a funny (but very distinctive) smell which, if caught early, would save the heads from actual destruction. There would be a small amount of oxide on the (usually) one head which could be cleaned with the special spatula and early form of wet wipes (they came dry and you wet them with isopropanol). But if you missed the smell, then the next stage emitted a characteristic "ping, ping ..." noise. There was still a 50/50 chance of saving the heads, although by this stage more than one would be affected. If all these clues were missed then that loud humming noise would tell you that the head assembly needed replacement after a tedious cleanup of the drive bay and filters and another trophy (or two) could be added to the "groovy disk" platter collection on the wall.
Not doing an adequate cleanup and then crashing the expensive drive alignment test disk was also erm... a strong learning experience!
Oh those old exchangeable disc drives:
40MB, 60MB, 80MB, and if you were very lucky 200MB!
What a revelation when the Fixed Drives came in with their whopping 320MB and 640MB capacity...
Now, as to the "start-up" on thse drives when a new pack had been inserted and someone hit the "go stud":
There was this pantomime on the EDS60's as a set of metal fingers swing in and back again before the read/write heads popped out and started their seek dance. These were originally meant to contain cleaning pads, but allegedly those were found to do more harm than good and everyone removed them, leaving their sad slow swing out and back again as a reminder. We used to call them "dust sprinklers:.
Now these drives, as the original author mentioned, were extremely susceptible to damage caused by stray dust and hair, and as a result all computer rooms were kept stringently clean and aseptic.
Who am I kidding?
The best modification I ever saw to one such drive was the addition of ash-trays to the side to the cabinet. This was because there was no way that the operators could be persuaded to stop smoking in the machine room, so the best that they had managed to do was to get them to take the cigs out of their mouths while they changed the disc packs.
Any guesses as to where this customer site was?
Oh go on, you'll never guess...
This was back in the 80's when as I recall the French tobacco industry was Government owned, and as a result smoking appeared to be compulsory for everyone over the age of 18 months. Although I understand kindergarten age children could be excused if they got a letter from the family Doctor.
It was, as has been said, a different time...
...ran from 7pm to 8am. Suits left and we were securely ensconced behind tightly secured doors.
All the operations staff drank, smoked and did drugs on the job
Crystal meth (biker crank, not Walter White's blue meth) felt like razor blades up your nose, but productivity soared.
1979 and punk, new wave or southern rock (Free Bird!) was blaring on the audio system in the computer room.
Tapes spinning, 40 hard drives rumbling like a 747 racing for takeoff, the laser printer the size of a station wagon, blinkin lights, teenagers and 20 somethings operating millions of dollars of data processing gear
I've never had a more fun job before or since.
Same here, the stuff that went on when the bosses weren't there...
The BEST laughs, the sort of job where you actually looked forward to going to work to be with your mates and have a laugh. It was literally work hard and play harder.
As you say, never had a more fun job before or since...
Glad I'm not on nights anymore though. Couldn't do it at my age. Miss those days
Night Shifts used to be the best as you were left to your own devices (no pun intended) and productivity soared as you did whatever you could to reduce the boredom. I recall many a night of machine room cricket, some of which had the unfortunate side effect of knocking non production disk units offline when a fantastic "six" managed to inadvertently flip a switch in the wrong direction. A quick call to a friendly engineer, who gave up some basic diagnostic tests (in return for not letting on where they hid their beer) help deflect from the real issue. Needless to say taking a drive offline was an instant "out" after the first occurrence.
Late 60's IBM 360 - same thing here. We woud have not considered a PDP a mainframe though.
Night shift was 11-7 curry first and a few beers for the ride - usually finished by 4 and playing cricked in the gaps between dist racks, As above - great times.
I watched one of thse drives fail the drop test! - hit every step on the way down!
As a salesman - selling 2 of those drives was a quarter's qota.
PDP-10s were mainframes? Never saw a DEC computer room that covered an entire NYC block or took up multiple floors of a large mid-town bulding, but saw plenty that were full of 370 and (later) 3033 mainframes. DEC was more like medium metal than big iron. If there are so many disk drives that they recede from view into the distance, that's a mainframe. If you could see everything you got in one room at a glance, you were only eligible for baby mainframe designation in the 70s, maybe. Have to be a big room though.
This brings back several memories from the early 80s as a new graduate.
I worked as a programmer for some typesetting software, written in assembler on a Data General Nova. It was mostly for local newspaper small ads but also for a company that produced annual financial reports for other businesses. I occasionally had the job of visiting to install new versions which meant I had to travel with one of those disc packs in it's plastic cover. I was allowed to get taxis rather than use the underground as the wisdom in those days was that the electric rail could corrupt the disc!
The users of the system we re-trained from the old hot-lead typesetting systems (The London Science museum has one, amazing mechanical technology) and peeved at what they saw as a de-skilling of their job. So it wasn't too surprising that the company had regular system crashes on a Friday afternoon, late enough that by the time the system had been rebooted and initialised there would be no point in continuing leading to the workers being sent home early. Took us ages to find the bug as any enquires like "what were you doing just before it crashed?" only ever got very vague answers!
I also have memories of the photo-typesetter, driven by paper-tape. It contained an opaque glass disc with transparent cutouts for the letters and numbers. The disc was spun to select the correct symbol and light shone through the "letter" onto photosensitive paper, the paper was moved and the next letter spun into place. To change font size lenses were moved back and forward to change the size of the image on the paper. To change font you inserted a different disc! It made a hell of a racket when in operation..
Ah, yes ... night shift. More modern time problem with support from a company we will just call Happy People. The nightshift operator at a bank I used to work for had a red light appear on a RAID-5 array on a server. She dutifully called the Happy People tech support, which told her to remove the drive and reinsert it. She did as told and voila! the drive went back online and all was well -- until the next time the red light came on. Without calling the Happy People tech support, she removed the drive and re-installed the drive. This happened repeatedly where the RAID array brought the drive back online and restriped properly each time, that is until it didn't. I have had the Happy People tech support tell me to do the exact same thing when a RAID-5 array drive failed. I refused that advice and insisted on them sending a replacement drive each time, which they reluctantly did.
Place i worked had a whole herd of them connected to a bunch of Unisys 494s (run in 490 mode). Also had some 1782s and (IIRC) VIIIC uniservos (tape drives). Eventually replaced with 1100 series and the "disk packs" of the day. More fun.
Prior to that i worked a place that had IBM 360/55 with 8 disk drives. We used to keep a card hanging on the front of the control unit to tell us which pack was in which slot. That worked until one guy failed to notice that the card had dropped into the drive and it crashed. And he moved it. An moved other packs. Disaster. Manglement would not allow us to purchase packs from anyone but IBM and it took quite a while to get the drives repaired and new packs ordered. UGH.
Think we had an IBM 3525 there (again, IIRC) - combination card reader and punch.
One place I worked back in the early part of this century our already struggling network started to die completely every Friday afternoon.
The cause was initially a mystery, an few people bothered because it gave us an excuse to bring beer o'clock forwards a few hours, but after a few weeks one of the network guys decided to investigate.
The cause was our MD and his brother playing Medal of Honour across the network, being the MD and the MD's Brother all their traffic was prioritised, so lowly work related traffic had to fit in the gaps between battles.
Eventually we installed a better switch and stuck their PCs on their own VLAN...
Biting the hand that feeds IT © 1998–2020