Shock horror – and there goes the network neighborhood Oh for heavens' sakes is it Monday already? Far out. Well, if you're here anyway, you may as well read another instalment of Who, Me? – The Register's weekly attempt to look on the bright side of the working week by revelling in the misfortune of others. This week, meet "Vincent" who once, many many moons ago, worked for some …
...the UPS battery had failed and sent a voltage spike to the stuff it was supposed to be protecting. Luckily most of it was salvageable and my boss invested in a line conditioner to sit between the UPS and the "important" hardware. I have this issue with most of my bosses: They cut corners to increase their bonus and send the employees that trigger the traps they set to the bread lines. "Vincent" is a lucky man.
When my UPS played up, it killed the power supply on a cheap and nasty HP printer (bought as a distress purchase during COVID). It was the only item affected. HP sent a (refurbished?) replacement a couple of weeks later. It failed 3 months later, just after I had bought new ink cartridges. I replaced it with a basic Brother inkjet which is fine - Unlike the HP, it works well without requiring an internet connection.
HP really preys on people with the inability to think ahead...
Their printers are usually a lot cheaper than a full set of ink cartridges, for that money you could've bought a way better printer. (This goes for many cheap laser printers too, but I think its mainly HP doing it)
And the "but I don't print a lot!" Excuse doesn't matter here either, because inkjets are terrible if rarely used. Print heads dry out and ruin the printer.
Always go with laser printers if you can't help it!
I've also had pretty good experiences with Epson EcoTank printers. Bonus: it's literally impossible to put some kind of ink DRM on the printers because you pour it in from a bottle.
But don't plug your laser printer into your UPS!
Unless your UPS is really beefy as the power consumption as the fusor heats up is normally quite chunky and can be upsetting to a UPS of sensitive disposition...
(I know this from RTFM'ing not because I've witnessed anything in particular - honest!)
... and by "really beefy", meaning a three phase model of UPS that can run a data center.
I have a UPS unit plugged into the same circuit as my laser printer and when the laser printer spools up to print, the UPS sees a short brownout, and it's a standard US 15 amp duplex outlet.)
I once installed a small laser printer in a backroom of a very old building (pre-war im pretty sure). Every time you started a print the fluorescent lights in the room went out and then kept flickering for the duration of the print.
The worst part is, HP used to make great printers. The early Laserjets, right up until maybe ten years ago were built like tanks, and would run for years. Even better, if they did require maintenance, HP provided proper service documentation and spares. Simple stuff, like fixing paper jams, was well with in the capability of the average office worker, so we got less call outs.
Oh HP, I'm sure you're making more money now that customers have to replace their printers every few years, but you've tanked your reputation.
I've avoided HP printers for along time. Ever since a software update stopped me from using mine. The update uninstalled a previous version. Then refused to install the new one. Because there was one .DLL still installed. And that .DLL couldn't be shifted- or overwritten.So the install just fell over when it tried. It wouldn't skip or ignore it. I went through all the multiple levels of their uninstall programme multiple times with multiple reboots. Got nowhere. And to cap it all the version number of the "new" .DLL was the same as the existing one. It had not changed .
Replaced the printer with (I think) an Epson.
Never bought HP anything ever since.
hmm I guess YMMV.
I had a brother MFC which could use non-branded ink but it used to clean the heads periodically (using up a lot of ink) even if you hadn't turned it on, so it needed a new set of ink cartridges every year even though I hardly used it.
Moved to a Samsung B+W laser 15 years ago which cost £50 to buy + about £1 a year in toner for the amount I print
Funny you should say that. My litttle Brother HL-1110 mono laser printer has outlasted three multi-function inkjets, been used by various members of the family, schlepped around to an assortment of student lodgings and studio flats and still works beautifully. Now USBd to our main computer and does all the workhorse jobs with no problems using cheap third party toner.
Laser mono printers are the best!
Usually after 10-30k pages the drum starts to wear out and it can cause some black streaks, but it's usually easy to replace and a third party one shouldn't be much more expensive than a toner cartridge.
(Plus a lot of home printers won't print that many pages)
I ended up replacing my Brother MFC-8860DN unit not because it wasn't working, but because it was so out of support that drivers were getting harder and harder to support. I wish Brother would just relase the source, it's not that magical pixie dust that it needs to be hidden away, it's just a BRScript (postscript like language) they need to support. So anyone, my old unit only did BRScript v2, and they're utpo v4 and v2 is pretty much only older 32bit compat distros. I didn't want to keep running older Debian relases just for printing.
So of course I replaced it with another Brother Laser Printer. I hope I get another 14 years out of it.
> MFC-8860DN
You shouldn't have replaced it. Why? When installing the printer you can do the "select manually the driver", then click "Windows update" on the driver select page. Wait a few minutes, it really takes long. And then you will have Brother MFC-8860DN on that list. If you look around many VERY old printers, many over 30 years old, appear on that list. And this is Windows 11... How I know? I tested that before hitting "Submit" of course! "Brother MFC-8860DN USB" is the exact name listed there.
This post has been deleted by its author
Well when I had my Atari 800 (proper one with aluminium chassis) I got a printer bur the atari couldn't feed it - had to buy an RS232 interface as well. That 9 pin japanese printer cost about £400 (1982) whilst the box and the extra 16K (yep Kb) of memory cost just over £1600, as I recollect. I gave it to my business partner eventually, after about 5 years, along with the seperate interface module which was a litle beige box about 12 x 6 x 2 cm which had 4 RS232 ports and a centronics port (for the printer). It did their family printing with true descenders (9 pin) with the tails of letters beneath the printing line, and went through 2 three year degree courses. It was only retired to the bin when drivers became an issue.
Gosh, I wish I had it now. I've no idea what ribbons would cost but I came across some printouts and there more than good enough. Faded, but legible, it used fanfold paper, and you could even rip the holes off the side with the little perforations. (Shoutout to Tetly tea bags).
RIP, the best £400 I ever spent. These days it wouldn't even but a couple of weeks drinking. Espescially not at the rate I did then. Sadly beer fugit even more rapidly than tempus fugit.
@TDog - £1600? are you sure? Im pretty sure that an Atari 800 came with 48Kb. and cost no where near £1600 in 1982 (it had been on the market for 3 yeasr by then)
What was the printer, even a NEC Spinwriter was sub £400 in '82
At least 10 years ago I was gifted a secondhand HP LaserJet 1015, a dinky little mono printer which I use at home/office. I changed the drum once and buy 3 party toners and it's never missed a beat. Probably the only HP printer I've never had trouble with.
However, around the same time I was already recommending or supplying Brother Laser printers, both colour and mono, as a solution for reliable and economic printing, even to my domestic clients. As far as I know they're all still working.
If you buy a printer and it works well, buy a spare. When your first one eventually dies, you can port across the cartridge with unused toner.
Why would a printer need an internet connection to work? Or perhaps, why would anyone buy a printer that needed an internet connection to work?
FWIW, I also use Brother lasers.
I replaced it with a basic Brother inkjet which is fine - Unlike the HP, it works well without requiring an internet connection.
I've said it before, but I'll say it again. Replace with a colour laser, not an inkjet. The up-front cost might be higher, but consumables are much more cost-efficient, they don't have print heads that clog and dry up, toner is typically much more colour-fast than water-based ink, it doesn't run if you accidentally spray droplets of water on it, and doesn't tend to expire or evaporate if not used. Brother is a good choice though. Our previous Samsung colour laser finally died enough to not be worth the percussive maintenance during the pandemic, and the Brother replacement has excelled, and has a much higher build quality than the Samsung one, without the cartridge DRM and HP drivers (!). I'd be very surprised if it lasts less than ten years.
Oof, that's rough. I once had an electrician mess up the phases and he sent 380V through the customer's entire building...
Had a pretty beefy line interactive Eaton PSU in between but it wasn't ready for a fuckup that colossal... probably the most intense electrical fire smell I've encountered, the whole rack was fried!
Good for my companies' turnover, bad for his insurance premium!
I once had an electrician mess up the phases and he sent 380V through the customer's entire building...
At one point we had the aircon in my computer room [1] at home serviced. We were about 100 miles away in Snowdonia and my parents were house/pet sitting.
The power strips that came out the UPS were all clearly marked "UPS, computer use only" and mum quite clearly told the man that, if he needed to use a vacuum cleaner, under no circumstances was he to use *any* of the sockets in the computer room but to use the one outside the door.
We can all guess what he did.
Yup, he plugged his vac into one of the UPS strips and blew the UPS.
I got a very distressed call from my mum that everything was beeping loudly and it was driving her up the wall. Fortunately, the UPS had failed into passthrough mode so the servers were still being powered. I talked her through how to silence the alarm and then drove back home to power everything down and reset the UPS. Fortunately, the UPS, after a power-cycle, came back to life happily. Otherwise the aircon service man would have got an £800 bill for the replacement.
He claimed to have not been told. Mum claimed he had. I trusted her a hell of alot more than I trusted him - and pointed out that the UPS power strips were all labelled and he would have had to have stripped off one of the labels in order to plug his vac in.
I never used him again.
You basically can't run a company with employees unless you insure yourself for employee accidents/mistakes. One slip-up can bankrupt you, especially electricians, plumbers, etc. who can cause a lot of damage by making pretty minor mistakes.
If you're self-employed it's a good idea too, but it can be quite expensive.
The secret to successful lying to to make yourself look at complete idiot.
If you're late in to work you could try saying the bus was late/that car broke down etc. No one lends it any credence.
On the other hand, if you say "I crapped myself! I had to go home to clean myself up and get changed!" no one will ever question it.
to make yourself look at complete idiot.
Then you can be just telling the truth.
If you're late in to work you could try saying the bus was late/that car broke down etc.
If your car broke down, actually take photos, same with the bus. Get the receipts. Always.
Tell the boss you can't afford a better car and maybe if they could give you a raise your attendance could improve.
Or if you got late because of the bus, just say you couldn't get out of bed, because all you see that you are just a "bus w*nker" and then it came late.
and talk about that raise.
anyway if you keep coming late, just use different time zone.
Once had a discussion with a supervisor about the difficulty in sacking someone who was very unreliable. He'd put stuff on his self certificated sick form like "had to go shopping", "couldn't get be bothered to get up" - for a while it worked as Personnel said we can't sack him as is obviously isn't firing on all cylinders, he was eventually dismissed.
From observation, (of an ex-friend), the secret is to repeat the lie over and over and over again until the victim has no choice other than to accept it. Add in manipulation, psychological control over many years and whenever caught lying claim "I was only joking" and distract further. For brownie points steal money from a shared bank account for 10 years too, and lie about this and then lie about being happy that it's in the open now.
The other option is to just tell a more outrageous lie and repeat this process until the victim, or country/voter pool, is distracted by this and has forgotten about the first lie.
Yep, been there, didn't done that (i.e. blaming the students for I was one myself). And before you know it, police is wandering the campus, interviewing* students about an alleged hack attack. What allegedly happened was a large-scale password crack. (In our alleged defence, those passwords were, allegedly, really weak. Honestly.)
*No interrogation took place as the cops themselves didn't seem to have a clue what they were supposed to investigate. Presumably, having all had "123" as their own password.
This post has been deleted by its author
《I have learnt over many years to leave things well enough alone. Merely the act of checking on something can cause a system that had been working perfectly to break.》
Not peeking inside the box lets Schroedinger's cat have a happy or at least longer half life ;) ie don't go about recklessly collapsing wave functions.
I always had a problem with that. Why not place the cat in a humane evironment (if you must) or just ensure that life support is available for a week. Run the test with your 50% chance with your radioactive source of choice. Wait a week. open the box; if it is smelly and horrible then the cat died without the event being observed (unless microbes, viruses and whatever count as observers). If there is a strong smell of cat shit and pee, and the cat tries to claw your eyes out then the cat is still alive.
This in itself is not a proof nor a valid test. But the cat has a 50% chance of dying. Should either event occur then repeat the same until you get six sigma results; this is the gold test.
OR
you could simply note the amount of decay in dead cats. If it were bugger all then they have obviously only just died supporting the basic hypothesis of observer effect. And if they smelt like buggery then they died a week ago - so stuff anthropic principles (unless cat's are observers, so repeat the same test with bacteria and penicillin cultures doing the turtles all the way down thing) or accept actually it is a non proven load of bollocks. And then wait for the theoretical physicists to produce several hundreds of papers criticising the methodology, the ethics of the methodoloy, the new string theories that explain this, cosmic (insert your own word here), But the true winners will be the experimental physicist who will explain that you didn't specify the colour, spin and charm of the cat.
Oh dear. Next somebody will tell me that I really should be testing small spaces which are described as "not enough space to spin a cat around in" using actual cats. I admit it's not easy, as cats are not the most cooperative creatures in the first place, and one sometimes to retire worn out cats, and cats come in different sizes, but one has to do one's best.
To be fair, you're working from a broken premise anyway and your grant issuing authority ought not to be providing the funds for your research. The "not enough room to swing a cat" refers to the cat'o'nine tails used to lash recalcitrant sailors in the Royal Navy, not the soft furry variety.
Yeah no... when I first got to my current job aeons ago, we had a problem with power to the entire building, of which were just one small part. But we were lucky and had a beefy UPS and a generator out back. Anyway, I kept getting called in on all the three day holiday weekends, because the power would go out due to squirrels getting fried on sub-station bus-bars. The UPS would pick up the load, then generator would start, and then the batteries would die. All very mysterious and all the testing we could do didn't find any problems in the system. We even took a weekend outage where we put a breaker into each of the three lines feeding our system and would flip it to see if there was a problem with only losing one phase causing problems.
Nope, the UPS, Transfer Switch (to change from street to generator power and back) all worked just fine and dandy. Queue much head scratching.
So I took it upon myself to check the Transfer Switch each month with a manual failover. Low and behold, then damn thing would only switch over two of the three phases when it had been sitting idle for a month, but if you did another switch over test right away, all three phases flipped over just like that.
Ended up with *another* all weekend outage to rip out and replace that Transfer Switch. Took about 18 months to figure out just because it was so intermittent.
I agree and disagree. Looking can be dangerous. But not looking can be more dangerous, regarded as neglect. If the site you are on is that bad (humanly bad, not the machines) get at least an email with the OK to check. If they deny that OK, THEN you can officially run the blissful ignorance. Depending on how bad the site is you make it in writing, with delivery stamps for confirmation.
《it always helps if you make sure that the meter is set to measure voltage, not current.》
With very few exceptions I wouldn't let anyone "detained" within a University near a screwdriver let alone a multimeter.
There are those that don't know the difference between EMF(Voltage) and current, those that think they do and a very few that do. I am not sure that it ultimately makes much difference :)
There are individuals in those institutions whose very proximity to any piece of equipment more sophisticated than a slide projector can rapidly cause said device's mysterious failure.
Elsewhere never let a developer anywhere near critical equipment - lock up your gear if you cannot lock them up.
My long deceased but probably sectionable granny didn't often make much sense but "the devil finds work for idle hands" was one the few.
At one company, new programmers had to build their PC for work on their first day.
Company had a utility room with motherboards, CPUs, memory, graphics cards and so on and if something was missing or they liked something else, they could get a company credit card and order anything they wanted online (within reason).
Wouldn't work at larger business though.
Not really IT related but I was on a site abroad where one of the contractors was dealing with the 're-shaping' of a cliff in order to provide a sturdy, flat base for a power station. He was in charge of laying the blasting charges and only he was authorised to set the fuse.
One day, on his return from lunch and walking up the hill to the next area for 'adjustment' he was surprised by his team of local labourers running down the hill. They ran past him without comment but at speed. He had a brief pause before turning on his heel and running after his squad, faster than his years suggested; a huge explosion behind caused him to dive into a ditch and make his outline as small as possible as rocks and debris plunged and scattered all around.
The contractor's compound was set around a nice grassy area fitted out with a few tables and chairs. It was now enhanced by a huge rock which had landed exactly in the centre of the sward.
If you visited the pretty Northamptonshire village of Islip (across the river Nene, from Thrapston) and sat in the pub garden looking towards the river, you could see the strange outline of a tree stump in the branches about 30' up one of the Poplars. It was there Summer 1982 or 3 until the great storm of 1987 (Michael Fish for those in the UK.)
'Lifting' stumps with a little bit of plastic explosives was a quick and simple technique. Unfortunately the roots of this particular tree were much weaker than expected, so rather than lifting a couple of feet the stump went about 40' high at the top of its arc. It was private land and the owner never asked us back to remove the stump, but I was still worried about it until it fell in the river.
Anon for hopefully obvious reasons.
From The Seventy Maxims of Maximally Effective Mercenaries:
2. A Sergeant in motion outranks a Lieutenant who doesn't know what's going on.3. An ordnance technician at a dead run outranks everybody.
This post has been deleted by its author
There are those that don't know the difference between EMF(Voltage) and current, those that think they do and a very few that do. I am not sure that it ultimately makes much difference :)
Like the time I used a Quicksilver power supply in my brother-in-law’s G3 and was roundly castigated by the Online Applescenti because it wouldn’t work (despite the fact I was telling them in as many words that it did) due to the “trickle voltage” being too low.
This has informed my opinion of Apple users ever since.
Working for a submarine telecoms cable manufacturer in university holidays when I was required to get a bit of "shop floor" experience. The company was pretty good about letting me have a week in several different departments which ended up in the "final testing" area of the shore-end equipment - think lots of full height 19" telecoms racks.
I was doing the final tests on the "spare set" for something like TAT-9 (which dates this whole story if you look that up) prior to delivery to BT. Along with various messing around with optical fibres and high speed data testing equipment, etc. the final tests were "how much current does each shelf in the rack draw?". The technique was to grab one of these wheeled stair/ladder things, go up to the top with your multi-meter to where the fuses all were, set it to current mode, put the probes into the terminals on each side of the fuse, hold them with one hand whilst you unscrew the fuse with the other. You can read the current, then screw the fuse back in and write the current down on the test sheet - all of which means you get the normal operating current without anything being turned on/off. You then rinse-and-repeat down the line of fuses at the top of the rack.
The final measurement was supply voltage to the rack. This came from a standard telecoms 48V DC supply (think big pile of lead-acid batteries).
Notably the instructions in the test did _not _ include the helpful advice to set the multi-meter to voltage mode first.
As a multi-meter in current mode is a basically just a thickish piece of wire the resultant spark and clunk from the circuit breaker on the battery pack were quite impressive. Followed by the extensive "rattle" as every relay in the various PSUs clicked off and the descending whines of the many, many fans in _all_ the racks that were currently under test ...
Apparently I went white-as-a-sheet which I suspect did something to dampen the laughter.
Thankfully the foreman scuttled over pretty quickly to put me out of my misery and said something like "Don't worry - it's just the circuit breaker, the racks will be fine. Everyone does this - once. You forgot to set the multi-meter to voltage didn't you?"
Using a clamp requires you to have access to the just one of the cables in the cable since if you measure both the "supply and return" the net current will be zero unless you suffer from an earth fault in the equipment. And I guess that having the outer insulation on the cables stripped to reveal the internal cablers would make the customer ask annoying questions...
Yes, you could do that in theory if you measured the voltage across the fuse and knew its resistance.
But in practice that would likely be pretty inaccurate because the fuse resistance is probably quite variable from fuse to fuse.
That is pretty much how a multi-meter works on the 10A setting though: it is a thick piece of wire with two sense wires soldered on to it at each end. The multi-meter will then actually be reading a very small voltage drop between those two wires but will (hopefully) have been calibrated so that variations in the thick wire and exact positioning of the sense wires between devices will be compensated for.
By connecting the meter across the fuse terminals then removing the fuse, the current then goes via the thick wire in the meter until you put the fuse back in.
Fuses are actually pretty inaccurate devices anyway. The perception is that a 10A fuse will run fine at 9.9A but blow at 10.1A but that isn't really true...
>Fuses are actually pretty inaccurate devices anyway. The perception is that a 10A fuse will run fine at 9.9A but blow at 10.1A but that isn't really true...
To add to this, they also have (equally inaccurate) curves for low long a particular current will take to blow a fuse. For example, taking 10 seconds to blow at 2% over rated capacity.
Indeed - in another aspect of that same job we used shaped fuses that looked a bit like keys. The idea was that the different thickness bits of the fuse would take different amounts of time to blow as the thin bits would have higher resistance and less heat capacity so would melt at a different time to the thicker bits. So they could have different ratings depending on the time that rated current was actually flowing.
Given that they were bits of metal physically about the same size as a key I recall that the ratings were in the sort of 50A to 200A range - so pretty chunky. I never witnessed one blowing though. Presumably that would have been quite loud.
We also had sand-quenched fuses where the glass or ceramic tube is filled with sand so that when the fuse itself explodes it doesn't take out the containing tube too.
Why did we need these things? Well, submarine telecom cables have a very simple power supply. You stick one wire into the ground at your side of the Atlantic and connect the other end to the cable core. You do the opposite at the other end. Then you put enough volts on to get around a 1.5A current flow. "Enough volts" for the Atlantic is around 18,000
So one of the toys we had for testing/developing with was a 12,000V 2A DC power supply. A fairly sizeable beast that was quite scary when you consider what it was capable of. We had lots of safety interlocks around that lab!
I remember killing a multimeter, just one of the cheap no-name brand jobbies. Put it in voltage mode (500V) to test a live 240V IEC socket and the internal glass fuse blew so spectacularly that it deposited a fine metal "plating" over the orange plastic case inside :-)
Oddly, it was something I'd done many times before without issue, a quick'n'dirty, continuity, fuse and supply test before checking if the PSU was the faulty device on the dead PC <shrug>
I'm just a computer engineer, but that degree was about 50% EE courses at the time.
So why you can't calculate the current measuring the voltage across a fuse. Fuses are not precision resistors plus the fuses socket isn't a precision resistor either. So the reading would only be a relative thing that couldn't be compared to even the next fuse on the rack. Might work for saying that yes, it's drawing current, but thats about it.
With that costly of kit. I would think built in amp meters would not be unreasonable.
> should be able to calculate the current if you know the fuse resistance right?
Fuse resistance is rarely specified, or maybe as a maximum. You can't measure resistance while power is flowing. (Unless you know the current, but that's what you wanted to find.)
If you have an *identical* fuse you can measure that... except it will typically be two few Ohms to measure without specialized 4-point techniques. Don't bet on two fuses same part number different lots being identical.
Current transformers pre-date 1900, no Hall Effect needed for usual AC. H-P used to have a DC clamp-on which balanced a probe current against core nonlinearity, but these are now collector-items for very odd collectors.
Standard multi-meter probes are DANGEROUS. I will not loan a meter to anybody who is not deeply experienced -OR- it is not my stuff being probed. Even if they seem to know what they are doing, I have better things to do than call ambulance or fire services. If they don't carry their own meter, they do not have enough experience to use mine.
I don't loan mine out mine either. Not after I lent it to the "experienced" printer engineer* complained that there was a lot of sparks when he was measuring voltages on a PSU. Yes, he had it on Amps.... PSU and more importantly, my meter was fine.
* I claim that he's old enough to started his career clearing paper jams on the Guttenberg printing press....
Your foreman is bad. Honestly, if you want to measure current, especially high current, you use a current clamp, a hall sensor or something similar. You never ever bridge a fuse, unplug the fuse to let the current flow through a multimeter, and then do it backwards. While in operation. Doing it that way is beyond stupid. If you measure VERY HIGH current you can use the voltage drop two points after measuring/calculation the resistance of the big fat wire.
But since your actual mistake was forgetting to set to voltage mode AND you stand by your mistake you deserve thumbs up :D. Your foreman deserves something else.
TAT-9 "old" ? OK, now YOU make me feel old.
EDIT: There are current clamps made for high frequency above 100 kHz if you want to measure spikes. For my stuff the cheap "max 20 kHz" ones are good enough by far.
"Notably the instructions in the test did _not _ include the helpful advice to set the multi-meter to voltage mode first."
Yes, instructions like that, for jobs like that, are usually written for people with experience to use as a crib sheet. Not a wet-behind-the-ears "work experience" student.
Although it was certainly an experience for you :-)
So when I was younger and a lot more stupid, I was on my own racking an old A-Class server. It was an awkward height, and I was doing the work at the end of the day after racking half a dozen others (so I was getting a bit tired).
Queue me trying to get the damned rails in and my arms just gave out. Luckily the thing missed me, but it didn't miss the floor. Bent the front panel a little, but nothing I couldn't fix up. The system still booted up and ran fine.
My one and only secret screw up.....
Still should not have sparked if it slipped from one test point to another, just displayed the different voltage.
Much more likely that test point was receded, in a hole surrounded by the case which was of course conductive and grounded, and he held the probe at an angle.
never would happen to me. (cough)
reminds me of
ACHTUNG!
ALLES TURISTEN UND NONTEKNISCHEN LOOKENSPEEPERS!
DAS KOMPUTERMASCHINE IST NICHT FÜR DER GEFINGERPOKEN UND MITTENGRABEN! ODERWISE IST EASY TO SCHNAPPEN DER SPRINGENWERK, BLOWENFUSEN UND POPPENCORKEN MIT SPITZENSPARKEN.
IST NICHT FÜR GEWERKEN BEI DUMMKOPFEN. DER RUBBERNECKEN SIGHTSEEREN KEEPEN DAS COTTONPICKEN HÄNDER IN DAS POCKETS MUSS.
ZO RELAXEN UND WATSCHEN DER BLINKENLICHTEN.
Doing tech support for a Uni, I was doing some work in the Vice Chancellor's office. I didn't normally support them directly, but I think my boss was trying to curry favour, so, when they needed a new computer set up, which required moving stuff around, I was sent to do it.
I'd got the new PC set up. It was a state of the art (for then) P4 machine with bags of ram and a large HDD. Not sure why they needed it, they only seemed to run Excel, but I supposed that's a benefit of power..
However, to make a nearby socket available, I had to unplug a fan and move it to the next socket. I did this, and everything went off. I'd tripped the circuit breaker. The trouble is, that circuit breaker powered all the sockets in the VC's office, and the offices of their staff and most of the University senior management. The advantage of so many higher ups being affected was that when I reported it, and electrician was on the the case in less than 10 minutes.
Reminds me of the engineer who decided to test some circuit resilience during working hours.
We had a process to test resilience on customer sites at least once a year. Resilience testing followed our change process. We needed user approval with a scheduled time window and field engineers on call just in case we somehow managed to completely disconnect the site. But somebody decided that this didn't matter on the grounds that it was going to work.
On this occasion there has been a configuration change and out of the now legendary "abundance of caution" the customer decided they would like a resilience test carrying out in the site in question outside the normal testing schedule. The correct process was for the engineer to raise an RFC for failover testing out of hours. The engineer who had carried out the change decided that they would simply test during working hours.
Now even if the test was successful it wouldn't have gone down well with the customer as the failover was implemented using HSRP. With HSRP there's always going to be a brief interruption to service as everything fails over to the standby connection. But our overconfident change engineer didn't consider that for a moment.
The test was not successful. Or normal test procedure for testing is to drop the primary connection where it hits our core by shutting down the interface. For some reason on this occasion the change engineer decided to log onto the primary router and shut down the WAN interface. Because of some config error there service didn't fail over to the standby, but now our hero couldn't get into the primary to bring the WAN interface back up. No problem, just jump onto the standby and hop across to the primary. Except that the config error that meant that the standby wasn't accessible. The connection was up. You could ping the router. But that was all. It wasn't passing any traffic.
Our hero now went into panic mode. Their first instinct was to try to get a field engineer to site. While they were trying to do this however there customer called in and service desk went thorough the usual checks for a down site one of which is to reboot the router. Luckily as there config hadn't been saved with the WAN interface shut down this restored service.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
