Was he then called Vince Cable?
Tech team stood and laughed at colleague who pressed the wrong off button
Admitting to mistakes is a ticklish task, but one The Register tackles head-on each Monday in "Who, Me?" – the reader-contributed column that chronicles your worst working moments. This week, meet a reader we'll Regomize as "Bowman," who sent us a story from the early 2000s, a time when he says most servers were floor-mounted …
-
Monday 14th September 2026 06:44 GMT Giles C
I did the same….
Had to move a wheeled proliant server without the front panel, and pressed the power button, as you say those switches only toggled on release. So had to sit their whilst others said we needed to do an emergency reboot on the server before I could let go of the button, the springs on those switches are quite strong….
But I did do worse….
-
Monday 14th September 2026 07:46 GMT Dave K
Re: I did the same….
Did you get to the stage after a few minutes where you start mentally weighing up the possibility of releasing the button then quickly pressing it again to turn the machine immediately back on again - the hope being you can do it quickly enough that the server stays up?
I've done this on PCs before (never a server thankfully). Haven't managed to get it to work yet...
-
Monday 14th September 2026 08:29 GMT Sam not the Viking
Re: I did the same….
40 years on, I think I can now safely confess to inadvertently shutting down a power station by incorrectly pressing the stop/start button but immediately re-starting our essential element of the chain. Nevertheless, the major shutdown had been initiated and it took several hours to get the whole system back on line. The actual cause was never discovered because nobody witnessed the event (so it didn't happen).
The subsequent investigation into the cause did reveal a number of potential short-term events which might have invoked a major shutdown. However, my suggestion of updating the control sequence was adopted.
-
Thursday 17th September 2026 23:59 GMT FeRDNYC
Re: I did the same….
Unfortunately (for you, but fortunately for your gear), if built correctly those switches are sprung so they're only able to confidently cycle between the distinct binary states of "very off" and "very on", with no way to return to the previous state except by first fully achieving its opposite. So, the pins can't be re-shorted (which is all a computer power switch does, connect two pins on the motherboard together) until they've first fully disconnected — otherwise, you'd be able to rapidly "juggle" the two states by quickly jiggling the button slightly, which would probably make your motherboard extremely unhappy.
-
-
Wednesday 16th September 2026 07:57 GMT CrazyOldCatMan
Re: I did the same….
Likewise.
We had a room (calling it a server room would be generous - it was a repurposed meeting room that no-one liked using after about 11am because it faced south-west and you got cooked if it was a sunny day - perfect for putting servers in!).
We had a couple of servers (OS/2 LAN Server plus one MSMail Postbox server [1]) - the were not even tower units - regular desktops with monitors sat on top of them. Because the servers and monitors were both IBM, they had identical power buttons, separated by about two inches (the monitor was on top of the base unit).
On my way out of the door, I pressed the power button on the monitor to turn it off.. except it wasn't the monitor power button - it was the server. I managed to call to one of my colleagues to ask them to:
a) Tell people we had to do an 'emergency reboot' because of a problem on the server [2]
b) Log into the server to shut it down while I held in the power button.
When he finished laughing (and called over my manager - who also had a good laugh) he duly let people know and shut down the server. Which meant I could finally (after 20 minutes) release the power button.
I learnt a valuable lesson - never press a power button unless you could see which one you were pressing.
[1] Which was connected to our US side of things via an X25 line to transfer stuff between our MSMail setup and theirs. Which, given how flaky MSMail was at the best of times, even over a LAN, was a miracle that it worked more than 50% of the time. And, if it lost the far end the only way to fix it was to reboot the server. It got rebooted daily - if not twice a day. We made sure to reboot it just before we left for the evening - as the US end came into work.
[2] The problem being 'junior engineer pressed the wrong button'.
-
Thursday 17th September 2026 13:26 GMT big_D
Re: I did the same….
We had a line of six VAX 11/780s and one was due an upgrade. The DEC engineer turned up, the workload for the VAX was pushed across to the next machine in the line, all the users were pushed over to the other machine. The VAX in question was shut down and the console said, that it was safe to turn off the power.
The DEC engineer went behind the stack to throw the huge power breaker on the wall... Nothing happened. The cursor blinked, the screen still said it was safe to turrn off the power.
Then the screams started, from the operators on the next VAX, the one where the workload for the VAX being upgraded had been moved to... The engineer had thrown the wrong mains breaker!
-
-
Monday 14th September 2026 07:04 GMT Michael H.F. Wilkinson
Reminds me of the time a visiting student to our HPC centre (way back when) asked what would happen if he pressed the well-recessed power button of our Cray J932. I told him a little claw would come out and cut off the offending digit, and if that didn't work I had a pair of pliers ready for just such an event. He decided not to test the veracity of that statement.
-
-
Monday 14th September 2026 08:52 GMT Anonymous Coward
Hilarious
we would have done exactly the same back then, as in gone in with everyone we can find and have a laugh.
I do wonder what the reaction would be nowadays to a similar event, probably involve a major outage to half the nation and counselling for the hapless button presser & disciplinary action to the witnesses who laughed.
-
Monday 14th September 2026 09:32 GMT Anonymous Coward
A Dutch fable comes to mind…
namely the little boy with his finger in the dike.
You have to be rather careful of the domain of interpretation with that one.
Applying a generous dollop of Loctite® might have expedited the released of Vince.
I wouldn't imagine a decent application of superglue or loctite would do the Dutch boy or the dike much good.
-
Monday 14th September 2026 10:22 GMT ColinPa
You really pushed the big red switch?
I know a software tester who was very thorough, and as part of the recovery testing had to power off the machine (to test real-life power offs).
She powered off the box mid test and powered it on again.
This hardware was the sort that doesn't like being powered off - and it took a couple of hours to rebuild itself and do all of the updates.
The code she was testing had a few problems, and so she raised defects.
The conversation with the development team went something like
Them: What do you mean you used the big red switch?
Her: I used the big red switch
Them: Do you mean you shut the system down using a command
Her: No I used the big red switch
Them: People never use the big red switch because it causes all sorts of recovery problems
Her: I used the big red switch to test a power failure
Them: We've never tested that because of all of the problems it can cause
Her: I found lots of problems
-
Monday 14th September 2026 10:37 GMT UCAP
Re: You really pushed the big red switch?
Once, long, long ago in an office far, far away, I was representing a customer for a Factory Acceptance Test campaign for a very large and expensive (circa $1 billion) satellite communications system that they were planning to commercially deploy. Nothing novel - it was based on existing technology, tweaked a bit so that it could handle the customer's specific requirements. It should have been a nice easy piece of work - should have been.
I noticed that the tests being conducted by the supplier did not include a power-off reset simulating a sudden loss of power in the data centre (1) for one of the critical pieces of equipment, so I asked for an impromptu test to be run once all of the main stuff was out the way. "No problem" says the test engineer from the supplier. Power the equipment off, check that we are getting the appropriate alarms in the control system (we did, but I saw a bead of nervous sweat on the test engineer's forehead) then power the equipment on and wait for it to come up. And wait. And wait. After about 10 minutes, I declared a test failure and the supplier had to diagnose the problem - it turned out that the firmware was fine with a controlled shutdown, but *really* did not like an uncontrolled failure. More to the point, the development team knew about this and had been sitting quite in the hope no-one would notice.
Four weeks later with a new version of the firmware was delivered and it finally passed all of the tests including the uncontrolled power-off.
(1) I agree that it's unlikely if the DC's power system had been design & built, but I am a firm believer that million-to-one chances occur nine times in ten.
-
-
-
-
Monday 14th September 2026 17:02 GMT Paul Crawford
Re: You really pushed the big red switch?
No matter how well designed a system is, there is always *some* chance of power failing. So test for that to make sure that it is automatically recoverable (or at least gets you to the point when remote access allows a controlled restart).
Yes Igor, throw the third switch!!!
-
-
-
Monday 14th September 2026 22:10 GMT Anonymous Coward
Re: it's always the DNS - except when it's the UPS.
I was going into the grocery store today. I'm pretty sure the guy in front of me was carrying a small PC sized UPS into the store. It must have been heavy because he had to loop the cord around his hand twice. And he only held it by the cord
About as bad as the UPS delivery driver hauling my first UPS out of the truck holding one end and dragging it down the steps. Thunk, thunk, thunk....
-
-
Monday 14th September 2026 17:31 GMT BartyFartsLast
Re: You really pushed the big red switch?
(1) I agree that it's unlikely if the DC's power system had been design & built, but I am a firm believer that million-to-one chances occur nine times in ten.
Agreed, I managed to shut down a chemical plant that made the thickening agent used in the milkshakes sold by a worldwide fast food chain by tripping out their server room.
My error was assuming that the non redundant Proliant 1600 PSU had been tested for function and electrical safety before getting marked up as fully tested and being put on our spares shelf.
It hadn't and there was a short from neutral to earth (no, I know, I don't understand why it caused such problems either)
The site's error was in not testing their UPS properly, the site electrician was baffled as to how it could have tripped out the breaker before and *after* the UPS.
-
Monday 14th September 2026 19:32 GMT Anonymous Coward
Re: You really pushed the big red switch?
A colleague had a similar experience at a FAT for a piece of equipment for the military. He similarly asked for a "big red button" test - and when pressed it ... did s.f.a.
So there was a rather awkward discussion, and further investigation - whereupon it was found that many sensors/switches had been incorrectly wired up. Of course, that's a failure in the test specification in missing essential functions.
-
Monday 14th September 2026 12:27 GMT Pascal Monett
Re: We've never tested that because of all of the problems it can cause
You test it and solve the problems it can cause.
That's your fucking job.
Because you can be sure that, when the shit hits the fan, it'll be the problems you didn't solve that will happen.
She was right, and they were morons.
The real problem is that there are more of them than of her, to this day.
-
Monday 14th September 2026 12:47 GMT Yet Another Anonymous coward
Re: We've never tested that because of all of the problems it can cause
>You test it and solve the problems it can cause.
Tester> I released Velociraptors into the data centre
programmer> It was only fixing a typo on the web page, we didn't plan for Velociraptors
Tester> Well the test failed
(ob XKCD)
-
Monday 14th September 2026 17:27 GMT gosand
Re: We've never tested that because of all of the problems it can cause
This reminds me of good ol' Test Driven Development and Agile that I was first trained on back around 2009.
The theory is that you write all of your tests first, before writing any code.
You run your tests. They all fail.
You deploy code, run tests, again.
Repeat until all tests pass, and you are done.
I don't do testing anymore, but I still carry that with me and I still use my testing mindset to this day.
-
Monday 14th September 2026 19:38 GMT I could be a dog really
Re: We've never tested that because of all of the problems it can cause
Which means that unless you have tests that have 100% coverage of all program paths and data inputs, you cannot be sure that your code has the right functionality.
But the principle of writing the tests first also applies to more traditional development techniques - so you attack things from two directions, code reviews to validate that the code performs the functions required, and testing to verify that it does.
-
Friday 18th September 2026 08:47 GMT Bob Royal
Re:code Re:view
I requested a code review to look at something I had found.
The code scanned the list for somewhere to put a connection, then it did it again, find a slot or not!
Waste of processing time or what?
Waste of my time too. Think it got fixed?
It worked, so don't change it.
Good job the processor was fast, could have mucked up the real time working real good.
-
Wednesday 16th September 2026 03:20 GMT Claptrap314
Re: We've never tested that because of all of the problems it can cause
Then you mis-learned TDD. You write ONE test. You see it fail. Then you write code so that all of your tests pass. Then, except for the first test, you review your code & refactor as appropriate.
NO code gets written except to cause a specific test to pass.
100% code coverage is thus required by the process.
I cannot over-recommend this approach. I did a one-month project where the code for the second-to-the-last test cause two surprise failures. There was a deep design error & I had to rip & replace more than 20% of my code. Took less than two hours from the surprise fails and the fix.
(The code went live with 0 bugs reported.)
-
-
-
Monday 14th September 2026 17:13 GMT IGotOut
Re: You really pushed the big red switch?
When testing our fancy pants phone system, we had it in the contract that the system should stay running even in the event of a complete DC outage, due to the multiple geographic redundancy.
It was so much fun before go live going around the back and yanking the cables out of the racks UPS.
The fail over worked perfectly. It was the powering back up that caused the issues....which we sorted for the next time.
-
-
Monday 14th September 2026 17:20 GMT Boris the Cockroach
Was'nt quite
the big red switch, but another long cast back to the days of PFY for me
And I'm helping to install a big bit of kit and its in the mounting frame and lined up, but it needs someone to hold it in position while the attachment bolts are fitted and that someone is me.
So........ we get it lined up... and I'm there holding it ready when my BoFH goes 'lunch time... dont let it move Boris , we'll be back in a bit' and they all clear off for an hr down the pub leaving poor little me trying to hold 1/2 a ton of gear lined up............. what I didnt know was he'd already slipped 2 of the holding bolts in before that so I could have moved at any time perfectly safely.
Then they come back..... and laughed and laughed.... While I stood there and let anger and hate consume me.
Karma is a bitch because my BoFH was inside a steel tube fixing some test gear up and I had a large hammer to hand..................
-
Monday 14th September 2026 19:46 GMT I could be a dog really
Re: Was'nt quite
Ah, percussive revenge can be so sweet.
Maaaany years ago, I was working up a ladder on the farm when some wag drove "fairly quickly" past the bottom of the ladder, which was "a tad unnerving." A few minutes later they went past the other way, but at a more appropriate speed. As they passed, I turnd and banged as hard as I could on the cab roof with both hands - leading to very rapid braking.
-
-
Thursday 17th September 2026 09:47 GMT DwarfPants
Training Young People
We have just run a "write the instructions for making tea" with our Scouts, preparing them for later life, some may even become testers
One team, who for unknown reasons did not trust us at all, and wrote excessively defensive instructions. They were unfortunately not clear if the proportion of milk required related to the volume of the cup or the large milk container. Anyone for really milky tee?
Other teams did not specify we should be using a cup that could cope with boiling water. Don't make tea in disposable plastic cups people, its not a safe thing to do.
The leaders win again, and important life lessons were learned.
All the instructions the young people produced, would create a reasonable cup of tea if not being read by leaders being deliberately common sense free, so they all did a good job.
