The Novell NetWare box keeps rebooting over and over again yet no one has touched it? We're going on a stakeout A seemingly innocent uplighter looms large in today's On Call tale of NetWare nefariousness in which a reader takes observability to new lengths. "Dan", for that is not his name, was working as departmental support for a UK local authority "more famous for its wild animal park" than its IT prowess. This was back in the 1990s, …
Working for a small software house that used Apricot hardware* we had a laser printer** on the end of an overlong*** parallel cable. Every morning when the fluorescent lights were turned on the printer fired up and spat out a page. It was somewhat disconcerting the first few times until we figured the cause and got used to it.
*Which dates how long ago it was!
**Can't remember the make/model but it was Canon engined.
***About 7m, due to the way the office was laid out the server and printer were at opposite corners of the room.
I had the pleasure of meeting a lady from the far east many years ago. She was over on businness with the company I worked for and a few of us took her out for a drink after work. She was very nice (bought a round or two on expenses) and told us a story about her offices. Before they'd first moved in they'd done up the building with an interior designer 'helping'. As a result there was a disgusting multicoloured carpet on each floor. The designer had a specified a different yucky pattern/colour scheme on each floor so you knew you were on your floor. The office design had been done with feng Shui so the Qi energy flowed and other such mumbo jumbo. Only problem was that hadn't helped with the actual energy as the static the carpet generated was considerable. If you touched for example a metal desk after walking across the room you got a sizeable shock. She said it played havoc with the computers too and they had to have the carpets removed and replaced. The replacements were a more sober single colour and generated far less static. Sadly we never got to see any pictures of this multicoloured hell.
When the Barbican Centre first opened in London all the handrails were brass or bronze, screwed to the concrete the place is made from. As you walked around and touched a handrail to go up or down then many short flights of stairs you got a shock. As a student next door we used to wander round to see how much of a spark you could generate if you didn't touch anything for a while. You could get it to jump about 3/8" (1cm to all you newbies). You had to stay on the carpet as the woodblock flooring reduced the charge slowly.
Ahh Mixing desks.. I used to support a small video studio as part of my work, and part of the equipment was a midi controlled mixing desk. One day, I was helping some students with setting up a camera. They hadn't done anything involving audio yet, so didn't even think about the audio mixer.
While I was watching them do something, I subtly flicked the switch on the socket powering the mixer off and on. The Mixer went through it's normal self test/startup routine which involved moving all the faders individually.
The students all stopped, looking terrified, and one of them asked me if the mixer was haunted.
I did get a bit of a telling off from their lecturer when he heard about it, but my boss just laughed.
OH God. The carpet! The shock!
In our (council) office I got such a conditioned reflex I couldn't even go near a radiator or door knob without great mental effort. It was a flinch moment going through one doorway that was close to a radiator.
God that sudden ZAP if I got too near some metal. Decades later it can come back to haunt me, like getting out of the car, slight Zap of static and I'm back not wanting to touch metal for days.
The key to reducing the shock is a "key". At least when getting out of a car. I learned this after getting a new winter coat that imagined itself to be a Van de Graaff generator.
Hold the key in your hand and touch it to non-painted metal. You don't want to risk blasting holes in your shinny paint. The key spreads the shock out to a larger area of your skin so you barely feel it. Don't wear gloves or this may not work. And don't touch anywhere near the fuel door.
Also this only works for metal keys. Don't put static sparks into the newfangled electronic "keys" that you can leave in your pocket.
The office where I first worked had metal partitions, with glass from above around chest height.
The new carpet arrived, and the shocks started.
In those far-off days, we programmers filled in coding sheets with pencils. They are the ideal tool to touch each door before risking skin contact. Non-programmers soon adopted the same solution.
I worked in an office like that for a while. Unfortunately the door was also metal and the door handle was made of brass, which made it impossible to avoid getting a shock every time the door was opened from the inside. I put a plastic bag over the door handle to alleviate the problem.
"Hold the key in your hand and touch it to non-painted metal. "
Funnily enough, I had to use the "key" trick multiple times per day, at a certain site, years ago.
This wasn't a carpet thing, but a work seat thing. They were apparently designed to generate hundreds of volts of static electricity, no matter what clothing you were wearing !
I remember sparks of 3+ mm when I was approaching the key to any grounded object !
PS: yes, the key trick works very well, you don't feel the shock. Only the key feels it :)
Some decades ago my employer developed an online parking-and-admission system for an entertainment venue. The place was newly remodeled, new uniforms for the tellers, new booths, new stools to sit on, cold weather. Cue Opening Day, the terminals were found to reboot randomly.
The stools had metal legs, rubber crutch tips and nonconductive plastic seats. The uniforms were polyester. Every time a teller shifted her derriere on the stool, static electricity would be generated and sparks would fly.
The solution was to remove the insulating crutch tips from the stool legs, so the legs would have some sort of path to ground through the stone-on-slab floor, in conjunction with sourcing and applying conductive plastic seat covers to the stools.
It's not just carpets. I've experienced static shocks from walking across an epoxy-painted concrete floor whilst wearing rubber-soled shoes, then touching something grounded. I learned to hold a key in my hand for use as a discharge electrode when I walked in that room; the discharge hurt a lot less that way.
At 6'5" doorways have always caused me trouble but nothing like when leaving one room where you could generate a good 1" spark and the lingering smell of burning hair. or on more than one occasion an involuntary jerk showering hot tea in the air. My forehead still prickles when lightning is about!
Not as bad as the time I slipped on the golf course and got a wet backside and then sliced it nearly out of bounds and, being delighted to find I had an easy chip on to the green failed to notice the wire behind me was held up on insulators addressed the ball and got 8KV through my coccyx and couldn't control my legs for the next 3 holes.
In the office where we used to work, one of the directors had his office at the end of ours. I was sat in the corner near the door to his office. On the rare occasions he would come in (he was mostly out with clients). He would go into his office and turn on his PC, something in there would make a "clunk" sound and my monitor would turn off and on again from whatever EM pulse it was sending out. I was sat a good 5 metres away, behind a stud wall. I have no idea what it was that was causing it, but it seemed to be only my monitor (a cheapo Acer flat screen) that was affected. I'm glad there was nobody nearby with a pacemaker.
Sounds like it was doing a degaussing step at power on, that was a loud clunk and a wobbly screen for a couple of seconds whilst it sorted its self out.
I had a 21 inch Philips* CRT that did this too, it took up much of the desk space and I only learned afterwards about the fields behind them.
* One I acquired as I'd never actually buy anything from Philips, except light bulbs and even those were unreliable. This followed on from a 2x CDROM that I payed a lot for when they first came out. 2 days out of warranty, it drew a really nice pattern on the disc and never worked again. Customer service was non existent, so they got what they deserved and I've blacklisted them since.
In the early days of digital TV, I had a Philips set-top box (It was quite small, and I reckon could have been made smaller if the back of it didn't need to have a SCART socket on it). It was slow, the software buggy, despite the fact that it only had to manage to display the EPG and let you pick a channel, and it died after about a year of use. I used to have an old Philips LCD monitor that someone gave me as an emergency spare, kept in storage before it got taken to the local recycling centre. Whenever it was turned on, it would emit a loud high-pitched whine (I'm led to believe this is due to cheap electrolytic capacitors). I think it was only a couple of years old when I got it.
So, yeah. Blacklisted.
We have a 'Philips' 1080p 24" telly in the kitchen. Takes about an hour to start up for some unfathomable reason. Not really a Philips I think...
...but I also have a 17" CRT computer monitor by Philips, the old 'Acorn AKF85". Probably the best monitor I've owned, though my first Iiyama at work comes close. My second - and more expensive - Iiyama at work is very much a disappointment with an awfully inconsistent backlight leading to very noticeable darker and lighter parts of the screen.
The AKF85 is a beast. Recently reminded of this as it has just reappeared from the attic in preparation for a house move. Shame I don't have the desk space to see how it compares with my current crop of LCDs.
M.
This brings back the memory of an early Phillips colour TV when I was a student, earning a few quid at weekends in a TV shop on deliveries and repairs. This set was a behemoth, taking two or three of us to shift it. It was a dual standard model, able to show the older B&W 405 line system and the newly introduced 625 line standard. I don’t remember the screen size, but the CRT seemed enormous (this was before the wider scan ones were available, so was as long as it was high; we counted over 30 further valves on its boards. In those days, one of the the first checks was for faulty valves - so a deep breath before starting. I don’t remember what the fault was but we found several faulty valves; they couldn’t all have failed together so a call to the customer. He only watched on 625 as that was colour and had another TV for B&W. After some trial and error, we found we could actually leave out around a dozen valves and it still worked on 625. In appeared that there were almost no common parts between 405 and 625 - basically two TVs in one box, sharing the CRT, speaker and some of the buttons. It probably shipped with the 405 side not working. We hadn’t sold it and it had come to us to fix because the original shop hadn’t a clue - a box-shifter that took the cash (as much as they could squeeze) and the customer was on his own.
Early 80's I bought a Philips colour tv, I think 19" maybe 21". After a year it started to play up. Had it fixed, a few months later more problems; fixed, a few months later... anyway, you get it. I eventually got a chatty engineer who told me Philips tv's always suffered from dry joints and the fix was to just reflow them every so often.
Much later I had a Philips VCR and Hifi (I was working for someone who was sponsored by Philips CD...). The hifi fell apart, literally, after a few weeks. The VCR lasted a while though. Shit software.
Absolutely avoided Philips ever since.
Back in my younger years I worked for my dad in his electronics repair business. We *loved* Phillips products. They're just expensive enough that people want to fix them rather than replace them, and they're crap enough to ensure steady business.
Since Phillips themselves never fixed anything (even under warranty they farmed it out to third parties, AFAIK) I have to assume that this was not by design, because where's the profit to them?
Easy to fix, too. With Phillips, it was always the solder joints, and if it wasn't the solder joints, it was the electrolytics, both of which are cheap and quick to repair/replace.
Hmm. Guess who Austin Rover bought their OEM in-car kit from?
Guess what always goes wrong first on any Austin Rover / Rover Group car...
Later high-end models have Alpine OEM kit common to contemporary BMWs. That's far more reliable, but the sound quality makes the Philips stuff sound good. This is such a low bar to fail to clear that I have to suspect that Alpine is staffed by limbo dancers.
My Philips TV broke 11 months into its warrantee period. I had tilted it forward to plug something in and the sound bar in front of it poked a hole in the screen :(
The only other Philips thing I had issues with was a food processor, which gave me shocks when it was UNPLUGGED. Philips diagnosed a blown resistor which was stopping a capacitor from discharging, and swapped it out for a new top of the range non-shocky food processor.
Hang on there, is this me from a alternate reality? Your comment is word for word identical to what I would have written. Even your name dwarf has some important meaning to me. Degaussing coils, tick, never buy anything from Philips tick, (you didn't once work for them or many of the companies they gobble up?)
If you have an old fluorescent fitting, make sure you change the ballast from magnetic (basically an inductor and a bi-metallic strip) to electronic.
https://www.lampshoponline.com/advice/ballasts-explained-the-complete-guide#_Swapping_magnetic_ballasts
There's a myriad benefits. Faster turn on time, even when it's cold. More efficient. Less EMI. No flicker. Longer lamp life. No hum. Etc...
p.s. Disadvantages. "There is a risk of electrocution when changing a ballast so if you’re unsure ask an electrician to do the job for you." :-)
A friend once worked in an office building. The agency were in the midst of an energy-saving conversion from magnetic to electronic ballasts. There was a power line glitch (might have been a lightning strike, I forget exactly.) The electronic ballasts let out their magic smoke, the magnetic ones just shrugged and kept working.
I suppose in this case it would be termed a darkening strike, as opposed to a lightening strike.
My optimistic side hopes that there was some sort of warranty, so the taxpayers were not on the hook for the failure.
If you have an old fluorescent fitting just replace it with LEDs when the bulbs need replaced.
You can get LED tubes that fit in place of the fluorescent tubes. I plan to use the LED tubes that are made to be direct wired to AC without the ballasts. Some of the LED tubes are made to work with the fluorescent ballasts, but then you need to have a working ballast.
Actually, the LED tubes require you to bypass the old florry ballast entirely.
Fortunately it's pretty easy to do, and they normally have very clear instructions.
However, I'd very much recommend replacing the whole fitting with a nice LED one. It's about the same price, possibly a little cheaper and they're far nicer yo look at.
they started replacing the fluorescent tubes with LED tubes at work but they were having to be replaced quicker than the old ones. they are now going through a slow programme (budget!) of replacing the whole fittings with LED panels. not only do they appear to last, they are also much brighter (although that may just be that they're ordering a higher spec)
If they had their budgeting right, they'd be _speeding up_ the replacement programme.
This is why local authorities have been replacing even comparatively new sodium street lights with LED. The payback period is mind-bogglingly short.
Even led to Kent aborting their "turn street lights of at midnight" programme, to the relief of anyone (but especially females) out after midnight.
"This is why local authorities have been replacing even comparatively new sodium street lights with LED."
Sometimes that requires changing the bowl/lens of the luminaire too - in the past that would have been wiped clean at the time of relamping, but now it has to be self-cleaning (i.e. so the rain washes the dust away) because the lamp replacement period is so much longer.
The output of LEDs tends to reduce over time, so the better LED lamps are designed with excess capacity and initially dimmed and then over a period of years they ramp up to full power, resulting in a constant output.
Some manufacturers use standard PCBs which can be populated with emitters to suit the project requirements - makes it easy to adjust the shape of the bundle of light and avoid extraneous light.
"The output of LEDs tends to reduce over time"
ONLY if they're overdriven in the first place.
If you drive a 1W LED at 750mw there's virtually no dropoff at all, their efficiency improves by 20% and the lifespan is effectively infinity (not 15,000 hours, more like 150,000+++)
Even better, if you fit a 75p microwave body sensor on the pole, they can dim to 7% power (which is only about half visible brightness) until something moves under them - result: even more power savings AND the effect that n'er-do-wells skulking around the neighbourhood can be tracked by where the streetlights are brightening
You can get hallway luminaires for this application too
"Even led to Kent aborting their "turn street lights of at midnight" programme, "
If you want to have fun with your councils over this one
Ask to see the costings and query whwther they've budgetted for having to replace tubes twice as often given that Flourescent tube are rated for ~1500 on/off cycles and they've doubled the cycling rate
(I wasn't popular when I did this.... and my projecion that increased maintenance costs would be more than double the power savings turned out to be spot on - it's not at all uncommon for people to fixate on "power savings" without realizing they might be killing a £5 lamp to save 20p in electricity)
"[...] change the ballast from magnetic (basically an inductor and a bi-metallic strip) to electronic [...]"
Don't then replace the fluorescent tube with a more efficient LED one. I did that and it was hesitant about lighting up properly. Soon followed by a "crack!" from the vicinity of the holder's ballast. The LED suppliers pointed me at their sales very small print footnote about asking their advice before fitting it to a holder that had electronic ballast.
Apparently the LED tube still works - just need to replace the ballast with a wire - a conversion normally achieved with a supplied dummy starter for an inductive ballast. I didn't twig the significance of the electronic ballast. No instructions. The LED's dummy starter was later discovered hidden inside some bubble-wrap cushioning. I had presumed the suppliers knew what they were doing with their LED design.
The issue isn't the power draw, it's the extremely reactive load
If you're billed for your power factor, you want to get rid of the ballasts wherever possible
This is where electronuc ballasts could reduce costs in a large building without changing kWh draw at all - going from phi of 0.65-0.7 to 0.95 made power companies happy and reduced charge penalties
I call BS: First something well engineered from a big brand vendor is not going to be affected by EMI from a puny light fitting even if it was right on top of it. Second, those uplighters popular in the 90s were always halogen, often 500W or 1kW. They had a tubular bulb with a metal cap on each end so might have looked a little bit like a fluorescent but were most certainly incandescent and had no ballast.
I'd put my money on dodgy mains wiring to the socket meaning the surge from switching on the light caused a brown out. It probably should have been on a separate circuit anyway if it was critical infrastructure.
No, it was to give more compatibility with the previous hardware - the ACT Sirius 1, known in the US as the Victor 9000.
Of course once you've pointed the ship in one direction, it takes a huge effort to steer it somewhere else.
The Victor had some interesting features. Storing 1.2MB on the same sort of floppy that IBM managed 360KB was only one of them.
Perhaps there is a relation with the fact that in the '80 and later there was supposed to be an embargo to South Africa for anything related to computers? (supposed, because at the time I read a number of tech papers about CTOS systems coming out of SA...)
And the one smuggled didn't have the local power cord but the one from their originating country.
In the UK we had different plugs for our servers - the earth pin was round, like on the older mains plugs. The Server power sockets had matching receplatcles, meaning that the servers would only plug into a surge protected power supply/UPS and that vacuum cleaners etc could not.
[...] all the pins are rotated by 90º [...]
Apparently you need more than one variant where there are different mains phases in the same room. UK computer rooms often used a British 13 amp plug - with just the thick rectangular earth pin slightly rotated. There were several different angles available.
Nope. That's an urban myth. Phase separation was required for a short time in early versions of the 15th edition of the wiring regs. It simply meant sockets on different phases had to be >6 feet apart.
This was very quickly removed from the next version of the 15th Ed, yet somehow remains in the minds of some to this day.
There are several types of 13A plug though, which are explicitly designed to be incompatible.
Most common is probably the T-shaped earth pin, but there's several variants in case you need to properly confuse people.
"There are several types of 13A plug though, which are explicitly designed to be incompatible."
NONE of them are BS1363 compliant, walsall and friends are all proprietary
https://www.flameport.com/electric_museum/plugs_13A_non_standard/ - using these means "vendor lock in"
AS/NZS3112's variants are standardised meaning you can hopefully buy plugs from a different supplier in 10 years' time
https://www.morvantrading.co.uk/technical-as-nzs3112 (and the wikipedia page on the same thing, but this page has clearer diagrams plus better explanation on use)
Trivia: The AS/NZS3112 plug/socket is the oldest nationally standardised connector (predates the UK standards) and has been in continual use since the early 1930s (it's based on a proprietary NEMA Hubell connector from 1916, but NEMA plugs weren't standardised until the 1940s and most countries didn't write up national standard on connectors until the late 1940s)
"meaning that the servers would only plug into a surge protected power supply/UPS and that vacuum cleaners etc could not."
A fact that the cleaner would discover only after unplugging the server, right? At least it wouldn't become a habit.
This post has been deleted by its author
"Sounds like a great way to take out a data centre to me."
The English Electric Deuce had mercury delay lines for data storage. They looked like large mushrooms with an electric blanket shrouding to maintain the working temperature. The shroud's mains cable dangled down the stalk and was plugged into a mains floor socket near the base. Yep - the cleaners found it very convenient.
This post has been deleted by its author
“ one for servers/PCs/etc, the other for vacuums ”
L5-15 or L5-20 twist lock usually does the trick.
Vacuum cleaners happen a lot. If you have the pull during the build out, place the plugs about a foot above head height. What we call a server closet, others call a storage room and a well places box will rip a standard height plug out of the wall.
It's an urban legend, but one rooted in practice:
At least city council one server room I dealt with in New Zealand had fitted BS1363 outlets on all the UPS fed circuits instead of AS/NZS 3112 outlets for this exact reason.
They got told off by electrical inspectors (not legal for permanent wiring) and forced to refit with NZS3112 "round earth pin" sockets instead (these won't take flat pin plugs, so job's a good-un)
These days you'd just use PDUs with C13 sockets
"how often is fluorescent lighting plugged into a plug socket." I find it to be quite common, at least in the US.
Task lighting at every cubicle desk since 2005 -- and some labs prior -- has been 2-foot or mini-bulb fluorescent*. Usually two modules per desk, one (fat) cord each, together happily filling up a duplex US/NA outlet whether they're on or off. Good thing there was usually another outlet around to power the important stuff (PC, monitor(s), etc.).
* Including a work trip from Michigan to South Carolina this past September, going from WFH (no work papers = no task light needed) to a small array of cubicles set up inside a factory/warehouse area. But look, the same fluorescent task lights, natch, and everything dusty as hell. (Plus a crappy old LCD monitor with VGA cord with less resolution than the laptop screen. So glad to return home to my beloved 32" with HDMI.)
Which also explains the lack of EMC.
The USA has only just begun to understand the importance of regulations for interference. While not quite a free-for-all, it's been pretty close for many years.
The conducted emissions from some US florry ballasts... oh my.
We had a server that kept shutting down every morning shortly after my boss arrived. Since we weren't allowed in his office none of us could figure out what he might have done to cause the constant FacePlants. Until one morning I'm passing his office with a cuppa & watch him hang up his coat. There's a soft click, his personal ionizer thingy comes on, & I hear the beep of the server going thud.
I asked him to change the outlet it was plugged into, he asked why, I pointed out that the plug on the other side of the wall was the server that wasn't appreciating his ionizer, so he magnimously lets me move the plug. All well & good, right? Not bloody likely.
He comes in the next morning, hangs up his coat, I'm watching out in the hall, & the server goes thud again. WTF? Had he moved the ionizer back? Nope, still on the far wall where I left it. So WTF is causing the problem?
We had to call in a sparky to do his little leccy dance & determine that the boss' coat was essentially a giant static discharge waiting to happen. Which it did every time he hung it up on the coat hook over the switchplate that the ionizer was plugged into. It hadn't been the ionizer, it was his jacket hitting the metal mounting screw, discharging through it, & the unearthed bits taking out the server.
Sparky fixed the grounding issue, the server didn't go thud when the boss came in anymore, but I couldn't figure out how/why such an issue hadn't been found at a much earlier stage - like when the plate had been installed in the first place. FekIfIKnow, not my baliwick, so I'll just be happy that I've convinced the boss to pay for a UPS that can handle the full load of the server & hopefully prevent any further FacePlants...
Fuck me & my wishful thinking.
Dispite being on a different wall, the UPS somehow wound up on the same circuit. The ionizer that had spent months getting zapped by the boss' leccy spewing coat. An ionizer that now played merry hell with the UPS batteries.
I have no idea how, no idea why, nor do I give a flying fuck. I left that job two weeks later when my notice kicked in so I could work for a company closer to home. I was never so glad to cut my commute in half as when leaving that Twilight Zone of a company...
However, the server in my case was (supposedly) protected by an UPS (Smart UPS 600VA). The UPS did a good enough job protecting against power failures (I tested it several times by unplugging its power lead - no trouble). However, when the server monitor (CRT, remember those?) was powered on, it would *sometimes* cause a reboot. Since the console was rarely used, the monitor was kept switched off and only powered when actually needing to use the console. Luckily for us it was a clear direct cause. Increasing the UPS sensitivity (the voltage threshold) did cause the UPS to switch to battery when monitor was switched on, but the reboots still happened.
Powering the monitor from a socket immediately upstrem from the UPS wouldn't help, while powering from a completely different socket would apparently make the problem go away (but introduce a tripping hazard in the form of an extension cord). In the end, it was simpler to instruct everyone to avoid switching the monitor off.
I used to have 2 21-inch Samsung CRT monitors at the office and learned the hard way not to power them on or off at the same time. The few times I did that I managed to trip the whole floor. And each of those times several angry looks followed my walk of shame from my post to the electrical board to switch it on again.
Long ago in another profession, a space heater under a desk was found to make the CRT display have an epileptic fit in a temporary trailer on a construction site. For the first hour or so, the display was crazy, once the trailer warmed, the space heater shut off... After lunch, door opened and closed a lot, cue space heater!
I have to think the quality of electricity wasn't great either!
In a place I worked at (my first proper job up in Yorkshire), a guy had a heater under his desk.
One day there was a horrible small, and when we looked under the desk, the crap quality multi-socket extension that he had the heater plugged into was actually on fire. There were flames coming out of it.
Fairly bowel loosening seeing that. :(
A place we rented on holiday once in Oxford (pre-AirBnB but similar deal) had dimmers for the sitting room lights. One evening, sitting there quietly, we smelt burning plastic and noticed smoke pouring out of one of the dimmers. The landlord seemed less worried than we were; he didn't have to sleep in the next room. (Although actually, the daily invasions of parades of ants were a lot more annoying.)
Our Phone system was declared non Y2K compliant so we needed a newer one.
The powers that be decided on a Centrix phone system. the biggest POS I have ever used, with the worst sales engineer / project manager it has ever been my misfortune to meet.
Anyway there was an issue with a phone hanging up mid call.
This went on for many months with phones, cables being swapped.
After all that it turned out to be the chair.
Apparently the phones were very susceptible to static discharge (tested in a false teeth factory I was told) and whenever someone stood up / sat down on the chair it created a static charge and killed the call.
One chair meet skip later all was solved.
Of fingers and telephones...
One site I worked at had a Siemens phone system, instead of the more usual button and microswitch receiver detection, they used an IR light beam. One of my colleagues was on a call one day and getting bored, he looked at the receiver recess and noticed that dirt was gathering so he used his finger to clear the fluff ball, cue a broken IR beam and dropped call....
In a council another part of the world, at the time of Netware 3.x we had a 4 node VAX cluster (they fell over so ofter that 2/3 nodes wasn't enough to guarantee the system would be up) and Solaris (we saved money on the uptime counter by using a single digit rather than risk permanantly etching a 0 into the tens column) and SCO Unix.
Aside from needing a seperate Network server to run Arcserve backups (x86, early 90's SCSI and Arcserve didn't make it easy to have stability AND working backups), the Netware environment was pretty stable.
We eventually got them stable, but it was a combination of hardware fixes, upgrades and patches sent on CD's via snail mail so took a year or two.
A colleague brought in his home made desktop-sized Tesla coil to show us all how clever he was. It was very sparky and pretty and wiped out our floor, the floor above and the floor below. Fortunately the coffee machine was made of sterner stuff, which made the otherwise quiet recovery time more bearable.
Had the same thing happen that destroyed an inventory database which was hosted on said Netware server. A user plugged in a kettle to the only spare socket they could find which happened to be on a multiblock that the tower server was plugged into that was then plugged into a small APC UPS. The UPS didn't take kindly to having to supply enough current to boil a kettle and keep tower server going and went bang spectacularly. The server was located behind a desk in a pantry, the database was the inventory of the pantry for stock control.
Years ago (mid 90's) I was working at a site where the servers were in a secure server room (as in you needed lots of very official paperwork to get in to the room). We were working in a less secure 'console room' where there were some consoles and switches to let us connect to the subset of Sun servers we were permitted to access.
There was a bit of a strange situation. Many of us had recently come into possession of modern GSM phones (think Motorola Micro TAC 7500 and similar), which we would ostenta... err ... casually leave lying on the benches near the console. When we were connected to one particular server, if the phone rang it would spontaneously init 6.
We made some enquiries about why that might be the case and were forecfully told a) there was absolutely nothing different about that server and how it was configured, connected and spec'ed b) no we couldn't have a look and c) there was nowhere else on site we could connect from (not even to try out some hypotheses).
So mobile phones stayed off the bench, the server stayed running but we never worked out why :(
Any theories from the wiser heads of the RegHive?
I could tell when calls were coming in on my old Motorola AMPS phones because they'd click a few times before starting to ring
After moving to GSM, a few people were apparently convinced that I was psychic because I'd pull my phone out of a pocket and look at it for a second before it started ringing. I never had the heart to tell them It was set to vibrate 6 seconds before audible ringing commenced
Put one next to a live speaker with unshielded inputs. Then ring it and listen......
Just under a second before the phone rings you'll hear a buzz as (I think) the handset responds to the base station. Older phones had much more transmit power than modern ones do.....
Same happens just before a text lands but a bit less so.
Can still hear same if you stand near the door in our local chemists, as the call patient intercom speaker is above the door.
Crappy radio in my old van used to go berserk a bit before incoming calls to a Nokia.
Didn't have a mobile phone 'til about 1995 for tech support. Console "room" oddly enough had no reception given half the racks were stacked with telco kit for an Aussie mobile operator.
The rats, shit, and exploding capacitors can wait for another story.
I'm sure Dan knows that. Haven't all of us at some time worked for someone who isn't known for their IT prowess, doing IT? I have (note to employer, I'm not doing so now). All we can do in that situation is try not to make the reputation of the IT prowess worse; dreaming of improving it is often a wasted effort. Still, it doesn't necessarily mean that it's our fault. Put the wrong person in charge, have them ignore the people below, and you can take skilled techs and still make a mess.
When I were but knee-high to a grasshopper I was PMing a roll-out of a new networked in-store terminal system to a now-defunct electrical retailer. The sort of thing the salesy people would use to tell you it was out of stock until next year. The logistics went fine but the customer was not happy as the whole thing would collapse in a heap of springs and cogs at random points in the day. Great effort was spent trying to debug/diagnose everything to no avail. So, similar to this, I got an engineer to sit in a store all day and watch.... after a couple of hours, sure enough, it fell over... the culprit... every time they got a new delivery of a telly or a fridge they opened the rear electric wind-up door to move it in, the motor thing created huge spikes on the mains at which point the server had a wobbly and went tits-up.
In the CEGB where I was working we had a PDP 11/70 with 2 LP25 line printers. One of the printers worked flawlessly, the other glitched every 10 minutes or so. The DEC engineer turned up and tried to fix the problem - nothing showed on diagnostics - all voltages etc were correct. In desperation he tried swapping parts between the two printers to see which one caused the fault to move. He (and I) were astounded that the fault moved with the printer lid !! The lid was a large plastic part that did not come close to the electronics or any moving parts and both lids appeared identical to the naked eye !! Once a replacement lid was fitted, both printers stayed in good working order until the whole system was disposed of.
I still do not know what caused the problem.
Icon for what the lid did to the printer =========>
Nah, big printers often have a 'tinsel strip' to discharge the static build up caused by moving paper at high speed, when the tinsel strip wears out or gets disconnected/removed you get static problems.
Possibly someone had remover it, it was worn out or, as occasionally happened with some record players which suffered with *loud* static pops, some well intending person had replaced the lid with a piece of random acrylic which wasn't static passivated.
And you thought that the imminent demise of fluorescent lighting would solve all problems? Guess again.
Apparently those economical, not to mention wonderfully programmable, LED bulbs have their own emissions that cause trouble. At least that's the case if the installation instructions on a certain brand of - here's the USA bit - garage door opener with wireless remote control are to be believed (if you think it's a luxury, consider living somewhere that gets snow by the foot on a weekly basis and whose winter temperatures only sound tolerable because they are in Fahrenheit and rarely go negative).
So said brand recommends its own version of the LED bulb for that light that comes on when the door opens, so you don't run over Junior's electric buggy. The "interference free" recommended bulb, of course, is a bit more pricey than the one at the home centre. Those of us who can spot a scam a mile away just use the bog standard LED bulb. Still, it makes you think.
So said brand recommends its own version of the LED bulb for that light that comes on when the door opens, so you don't run over Junior's electric buggy. The "interference free" recommended bulb, of course, is a bit more pricey than the one at the home centre.
I suppose its quite likely that there are differences in the quality of the drivers in the bulb. As an example most cheaper basic bulbs are not-dimmable (ie might take offense of tried to be dimmed with triac). A cheap bulb is likely to have fairly basic led driver which could potentially (not necessarily) cause issues/interference elsewhere in the mains circuit. A mains powered LED "bulb" is not just a bulb like old fashioned incandescent mains bulb.
Suggesting their own brand and own brand only does sound bit like money making exercise.
So far the very non-dimmable 1OOWeqv bulb I installed is not giving problems. The same could not be said of the generic doorbell button I tried to use to trigger the door from outside. It turned out that the new open/close circuit, still a two-wire affair like the older version but with added features, requires a clean break in the circuit to function properly, and the trickle current through the bulb lighting up the button was giving it fits. Removing the bulb fixed the issue.
"interference free" recommended bulb
That is not actually that wrong. I've had problems with a no name cheap mains powered LED bulbs causing RF interference. Mate had problems with loss of signal on a digital TV that only occurred at night. It turns out it was the bulb in the upstairs hall.
Replaced it with a better brand name and no more interference.
When Netware 386 3.0 came out, I had a nightmare server on a customer's site in London (my office was in West Sussex).
To cut a long story short...
Regular trips to site to try and catch the issue. Very stress-inducing customer.
Replaced components until the server was effectively swapped out - apart from the case.
Novell took an interest on the basis that it might be a software glitch...they found an issue with the 'network cable disconnected' code...but the server still crashed maybe 2-3 times a week. Nothing else in the server room was playing up.
I tried a mains analyser - nothing stood out except perhaps the odd 'blip' at crash time - but it was barely more than a bit of noise.
Widening the search for a fix, I asked about the room, which belonged to another company, behind the server room...it was their kitchen, and hard up against the wall behind the server was their dishwasher. It turned out that the dishwasher was switched on when full..a couple of times a week!
Hey - EMP from the some motor or pump in the dishwasher knocking over the server!
Fix: Move the server to against another wall.
Plugging the light into a different socket may have cured the symptom, but did not remove the root cause, and so the setup remained vulnerable.
The fact that a momentary mains spike reset the PC would suggest that the PC's PSU was not all that robust (the PC I am using right now can survive as much as a 500mS power outage without rebooting, though my VDU does a reboot during the same event). A better solution would therefore have been to upgrade the PSU - or perhaps run that PC from a dedicated UPS (which after all was being used as a server). Not that I am a big fan of small UPS's because they are more likely to fail than the mains supply - at least in my case. At the very least it should have been fitted with an inline mains filter and then tested to ensure that switching on the light no longer resulted in a reboot.
You should try working somewhere that shares their mains supply with a hospital trust. Once a month at least they create a fake power cut by tripping the substation. Lots of spikes and surges when they restore the supply. Chews through switch-mode PSUs like it was at an all-you-can-eat buffet.
My parents rented a shop and flat in a small suburb of Rugby, very near to the radio station (long gone now). We never had a power cut, even while the rest of town was in the dark, had our circuit been interrupted and Rugby transmitter failed, it could have initiated WW3 because that was where the Polaris submarines were controlled from. There was also another unexpected benefit, there was so much EMF from the transmitters that fluorescent tubes would stay bright for several minutes, or even half an hour, after being switched off, and you could often hear Morse Code emanating from the cooker rings in the kitchen if you put a damp saucepan onto the stove without drying its bottom first. Rugby Radio Station's 1000ft masts were a "Welcome Home" as one approached southbound on the M6, it was a shame when they came down, we missed those lovely red lights in the night sky.
ALTAIAGFFA. A colleague was running an unofficial, but necessary server at his desk that would go dark in the evening. He was convinced someone was turning it off in some misplaced desire to do some tree-hugging. Denials that anything like this was happening were immediate and stern, so he set up a camera on his machine (they were a little bit rare at the time). He was treated to a fish-eye close-up of once of the guards as he leaned around the back to apparently unplug and plug the machine back in. I'm sure it's no surprise my colleague got a good chewing out for using surveillance.
We really think he was trying to be "green". When the PC was unplugged and plugged back in, it was not configured to reboot. Therefore the guard was saving the planet by powering off the machine. Just our best guess. We were never told that was why; we just got that we should never setup a sneaky camera like that...even in our own cubes.
As I recall with 3.x, which used dos to do the initial boot, you needed to write an autoexec.bat if you wanted the server to come up without stopping at a dos prompt: the install didn't create one for you. I recall writing a little assembler utility that slowly wrote a row of dots on the screen which was press any key to abort processing autoexec so servers would come up unattended, but it was easy to abort.
A KDF9 was running its on-site acceptance tests. One test involved reading a full reel of paper-tape at high speed. The system kept crashing partway through - at almost exactly the same place. Eventually someone standing next to the reader heard a sharp crack! from its vicinity as the system crashed. The metal bin collecting the paper-tape had become disconnected from its earthing cable. The sound of the reader had stopped anyone hearing it before.
Not strictly computer, but certainly related! Many decades ago, between leaving school and starting a "proper" job, I filled the summer holidays working as a van driver and salesman for a local electrical retailer. Colour TV had just started to reach a few lucky locations within our area, and one of our wealthier customers had bought a magnificent 25" Decca dual standard colour TV. These were the Rolls-Royce of colour sets in their day, so it came as something of a shock when the customer complained - less than a week later - that the TV had gone all psychedelic on him!
A service engineer was quickly despatched and spent a few hours degaussing it, then re-doing the purity and convergence. This was a BIG job on those old sets! Eventually, the set was as good as new, and he left a satisfied customer behind him.
A week later, and the customer is back! The TV has gone psychedelic again! Rinse wash and repeat!
After the 3rd occurrence, the service engineer decided to spend Thursday at the customer's house (it was always Thursday when the set went wonky!).
About 1130, the maid comes in with an enormous, pre-war, upright Hoover and proceeds to shove it under the TV before the engineer can stop her! The stray magnetic fields from the massive motor in this device played absolute havoc with the (fluorescent!) vacuum tube that was the CRT!
It only took him a couple of hours to sort it this last time, because of the practice he had now had, but the poor maid was left with strict instructions NOT to Hoover anywhere near the TV in future!
Slightly off-topic but I read a story about a cat that liked to curl up under the tv. The power lead was a long extension lead and the owner had coiled the excess cable up and placed it neatly under the tv. Electrics 101 coiling mains cable like that is not a good idea as it causes heat to be generated, which attracted the cat to treat it as a toasty basket. There have been stories of situations where, after many years of operation in coiled mode, cable can not be uncoiled because the insulation between the coils fuses together.
At one place I worked certain staff needed conductive footwear so as to avoid things a lot louder than the crack! of a static discharge.
Prospective suppliers would have their sample shoes tested. Virtually all would fail. They carefully made the soles out of a conductive compound, but then spoiled things by using a standard, non-conductive, insole.
The test rig was ingenious. The shoe was placed on a metal plate connected to one terminal of a meter, and the other terminal was connected to the inside of the shoe by a braided lead. The interior of the shoe was then filled with ball bearings to complete the circuit (though mostly not).
I too was in local government at the time. It was strict policy that every server has a UPS and was backed up to tape nightly. The network was not up to remote backups so it's a shame there were no fitness trackers in those days, the daily tape swapping routine involved daily visits to remote server locations.
Another thing though. Why set up a remote server so it needed a logon at the console to function? Especially if it didn't have a UPS.
And what sort of support person needed to stake out the server? Logs would show it power cycled. So you investigate the power supply. Response #1 why does this server not have a ups? #2 why is the server on the same supply as dirty devices?
We had a policy back in the day that connecting unauthorized devices to the same ring as a server was prohibited with sanctions up to and including disciplinary action. Every socket in every buildings was labelled clean or dirty. Somebody got a written warning for connecting a fan heater to a clearly labelled clean supply. It wasn't so much the noise on the supply that got them a warning. It was tripping the supply for several rooms.
Dead middle of the 90s and I was supporting several offices, each of which had their own web server running on a desktop PC, located physically in the office somewhere because they hadn't yet let me build a proper data centre.
One such office team had their web service running on a desktop that sat beneath their desks, which backed on to each other. The web site would fail every morning, then come back no more than 30 minutes later. It was very much a work in progress - imagine those UNDER CONSTRUCTION images with an animated gif of a workman with a shovel and lots of <BLINK> tags. Yes, that kind of web site. So I wasn't too fussed until one of them asked me what was wrong. I asked if anybody was kicking a power lead out, given its location and the spontanous power-down that seemed to be the problem. Most definitely not, they assured me.
After some weeks of this, I happened to be in that office at 10:30 doing some generic desktop support, and one of the office ladies kindly offered me a cup of tea. Thank you, yes please, I responded, then watched as she pulled a power lead from a wall socket above the desk to plug in the kettle.
...I always specified an APC UPS and Powerchute software with the Netware servers I used to install. A friend of mine who still runs the business we started nearly 30 years a go recently de-commisioned the last Netware server we had running (3.12 IIRC). It was showing an uptime of almost 6 years before it was finally laid to rest.
There are a few places left that combine office and R&D space with a manufacturing floor. Anyone who works there or visits knows that static control is extremely important because it not only gives you annoying shocks but it also can cause damage to products, damage that won't necessarily show up in product testing. On the manufacturing floor you'll see conductive flooring, anti-static smocks and personal kit like wrist and ankle straps. People will not be allowed on the floor without personal anti-static equipment.
The offices are equipped with anti-static carpeting. This stuff doesn't last for ever but you can mitigate even bad static using a fine spray of dilute fabric conditioner.
In this kind of facility you often find two sorts of outlets, generic and clean. Modern computers are a lot less sensistive to power dips than they used to but back in the day (20-30) years ago you'd plug your computers into the clean supply and the office lights into the generic.
Anyway, the point I'm making is that if you're used to working with electronic equipment then you're also used to taking precautions against static. Modern components are a lot less sensitive than they used to be -- early CMOS, for example, you could kill by just 'looking at it in a funny way'.
One of my first IT jobs had a Netware 4 server and a laser printer. Out of misplaced frugality they would power of the printer at night. This led to the occasional weirdness because if you sent a print job to the Netware server while the printer was powered off, powering on the printer would result in the laser printer spewing out several pages of PostScript source code rather than your actual print job.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
