back to article Turning a computer off, then on again, never goes wrong. Right?

Ah, dear reader, it's so delightful to have your company once again for Who, Me? in which fine upstanding Regizens like yourself regale us with tales of tech gone not so much right as … the other thing. This week meet a reader we'll Regomize as "Arnie" who encountered a problem that required finesse and a delicate approach, …

  1. jake Silver badge

    Reminds me of an old (early '80s) AI koan ...

    A novice was trying to fix a broken Lisp machine by turning the power off and on.

    Knight, seeing what the student was doing, spoke sternly: “You cannot fix a machine by just power-cycling it with no understanding of what is going wrong.”

    Knight turned the machine off and on.

    The machine worked.

    1. ArrZarr Silver badge
      Joke

      Re: Reminds me of an old (early '80s) AI koan ...

      Oh Jake you sweet, summer child...

      Your inexperience with Windows is showing here. Power-cycling with no understanding of what is going wrong is how 90% of the issues I have with my computer are fixed!

      Although I am reminded of all the times that stuff starts magically working as soon as the relevant expert is watching from over somebody's shoulder.

      1. MisterHappy

        Re: Reminds me of an old (early '80s) AI koan ...

        Happy memories of when I did desktop support & would shlep across site or to a different site only to hear a variation of:

        Me: "Can you show me the problem please?"

        User: "I open this & select here & then click this button and then... Oh, it's working now"

        1. imanidiot Silver badge

          Re: Reminds me of an old (early '80s) AI koan ...

          The big problem there is usually that previously they were either opening the wrong thing, selecting the wrong thing or not selecting anything, clicking the wrong button or all of the above.

          Users are idiots. (And I can say that confidently because most of the time, I'm a (L)user)

          1. Evil Auditor Silver badge

            Re: Reminds me of an old (early '80s) AI koan ...

            No. As a former hell desk staff I affirm it is the pure presence of "the expert" that makes the machine behave again.

            But yes, users are idiots. And I've considered myself a user for the past decades for I do audit and do not fix things.

            1. WonkoTheSane
              Thumb Up

              Re: Reminds me of an old (early '80s) AI koan ...

              Agreed. The mere presence of "the expert" makes the novice consider their actions more carefully.

              1. Strahd Ivarius Silver badge

                Re: Reminds me of an old (early '80s) AI koan ...

                And also make the computer stop misbehaving as long as the expert is there.

                In the '80, I had to go to a client once a month when they were starting some end-of-month task otherwise it didn't work.

                Even with a clone of the system I was never able to reproduce the issue they had when I was not present.

                1. Evil Auditor Silver badge

                  Re: Reminds me of an old (early '80s) AI koan ...

                  I think someone at the client site did like you. Even if it was the computer.

          2. Jou (Mxyzptlk) Silver badge

            Re: Reminds me of an old (early '80s) AI koan ...

            Half agree. I know cases where the time of the day changes behaviour. Doing the same action while the virus scanner is running its daily routine is a prominent example. A half-faulty network connection can cause button presses to be missed for no apparent reason. The sun shines on the case, causing temperature problems at specific times of a day, but not on rainy days of course. The list is LONG where not the (l)user is at fault.

            1. doesnothingwell

              Re: Reminds me of an old (early '80s) AI koan ...

              I had one desktop PC that would lock up but not with the cover off. The case had a small cutout on the boittom and cold air from the A/C would get pulled across the motherboard. Covering the hole with tape was a temporary fix, I didn't have a spare board with me.

          3. Alan Brown Silver badge

            Re: Reminds me of an old (early '80s) AI koan ...

            Or ignoring a perfectly valid message, saying it doesn't exist, even when it's on the screen in front of them and then - after reading it out loud - saying they don't understand it (file not found)

            Having a PhD doesn't imply actual ability to solve problems

            1. John Brown (no body) Silver badge

              Re: Reminds me of an old (early '80s) AI koan ...

              If anything, having a PhD only means we know they are good at one subject. How that translates into other subjects is highly variable. I've met Profs who I wouldn't trust to tie their own shoelaces properly, despite being very nice people and well respected in their fields :-)

        2. The Dogs Meevonks Silver badge

          Re: Reminds me of an old (early '80s) AI koan ...

          In my experience these 'issues' are normally down to 10 problems.

          P.I.C.N.I.C errors

          1 D 10 T problems.

          In about 1% of cases, I actually had to fix something instead of provide implicit instructions on how to use something properly.... for the 50th time to the same dept.

          1. The Oncoming Scorn Silver badge
            Alert

            Re: Reminds me of an old (early '80s) AI koan ...

            & usually to the same person!

        3. Outski

          Re: Reminds me of an old (early '80s) AI koan ...

          Ah, the old IT professional Proximity Field

          1. dadbot5000

            Re: Reminds me of an old (early '80s) AI koan ...

            I like to call it the Aura effect. Just my aura is enough to fix most user issues.

            1. Diogenes

              Re: Reminds me of an old (early '80s) AI koan ...

              Its amazing how many student computers start behaving themselves when I "lay hands" on the computer :-0 .

              1. Jou (Mxyzptlk) Silver badge

                Re: Reminds me of an old (early '80s) AI koan ...

                In the 80s? Literally: Because you discharged the case with your combination of clothes and shoes. Where as the student, sitting on the then-newest-shit plastic chair, and moving it on those well known plastic carpets built up enough to cause random bit flips. Tough some student flipped their bit on computers on purpose...

              2. MarthaFarqhar

                Re: Reminds me of an old (early '80s) AI koan ...

                AKA the fear effect. The computer knows we have screwdrivers, and are not afraid to use them.

                As oppose to the meddle effect, where the user had screwdrivers, is not afraid to use them, and the item never works again.

            2. Anonymous Coward
              Anonymous Coward

              Gender sensor

              My wife and I have concluded that, as most tech has been designed by men, most come with built-in gender sensors.

              She: I can't get this to work.

              Me: Did you turn/push/tap/flip that thing?

              She: I did. Nothing happened.

              I reach over and turn/push/tap/flip the thing. Instant success.

              Both at once: Must be one of those dang gender sensors.

          2. imanidiot Silver badge

            Re: Reminds me of an old (early '80s) AI koan ...

            Aka giving the "I have a hammer and I am more than willing to use it" look.

        4. Evil Auditor Silver badge
          Thumb Up

          Re: Reminds me of an old (early '80s) AI koan ...

          The most fun I had had with unjustifiably highly privileged developers: they had privileges to break things but not high enough to fix them. Typical call would be: "Hey [name redacted for I was no Auditor back then], I broke it again. Can you please come over and have a look?"

          1. Ken Hagan Gold badge

            Re: Reminds me of an old (early '80s) AI koan ...

            "privileges to break things but not high enough to fix them"

            That sounds like a configuration error, and not by the developers.

            1. Evil Auditor Silver badge
              Thumb Up

              Re: Reminds me of an old (early '80s) AI koan ...

              That pretty surely was the case, too - I wouldn't remember any specifics. It was, however, also to time of NT 3.5 and 4, Win95 and some PC plug-in cards for Mac. Such a configuration error was a lesser problem.

            2. MarthaFarqhar

              Re: Reminds me of an old (early '80s) AI koan ...

              We had one fiddler/developer that decided to change the default shell for Windows to "calc.exe". Why they did it, I never managed to find out, as they reduced a piece of very expensive lab equipment unusable until another member of staff called the IT Helpline, and sent me to have a look. The local fiddler was asked not to attend site without an appointment or a person with them at any time they were near the factory floor.

          2. swm

            Re: Reminds me of an old (early '80s) AI koan ...

            At the college I taught at we were studying permissions on files. Some student set the access pits to 000 on his user account and locked himself. Had to go to support to fix this.

            Not really the students fault - he was just experimenting (as he probably should).

            1. MarthaFarqhar

              Re: Reminds me of an old (early '80s) AI koan ...

              Absolutely. We had a sysadmin when I was studying who had the attitude, you break it, you fix it. Was an infinitely patient man, who also managed to never laugh whilst in the presence of the student or even the staff that had allowed this to happen. Down the pub was a different story.

      2. Martin-R

        Re: Reminds me of an old (early '80s) AI koan ...

        > I am reminded of all the times that stuff starts magically working as soon as the relevant expert is watching

        And all the stuff that breaks only when one particular tester goes near it...

        1. ArrZarr Silver badge

          Re: Reminds me of an old (early '80s) AI koan ...

          That tester is worth their weight in gold. If it works in their presence, it'll work in anybody's presence.

        2. Claptrap314 Silver badge

          Re: Reminds me of an old (early '80s) AI koan ...

          My grandfather could not wear wristwatches--they would stop working.

          I inherited just enough of this spooking action up close that I tended to break test boards and processors (especially at IBM) by simply getting to close. I was properly trained. I executed the steps carefully and properly. I was watched doing so. And, while being watched, I repeated burned out boards and processors. New manager insisted that I touch the boards. For a bit.

          \o/

          1. Anonymous Coward
            Anonymous Coward

            Re: Reminds me of an old (early '80s) AI koan ...

            In the late 90's, I was the go to person for testing a brand new digitising tablet that was *gasp* wireless. To save battery power, the puck was only powered on when the two sides were held, the user's skin conductivity completed the circuit.

            If it worked for me, then it would work for anyone. Fortunately they improved the design.

            And for what it's worth - I'M ARNIE!

            I was a contractor at the time, am now actually TUPE'd to the NHS! But, El Reg is correct, I never went back to that site again.

          2. Potty Professor
            Facepalm

            Re: My grandfather could not wear wristwatches

            Same here. When I was a kid, I was given a "Space Watch". It was a conventional wristwatch but the crystal was replaced with a stainless steel cover with two viewing windows, and the hands were replaced with two numbered discs. I put it on, and it stopped. Took it off, it started working again. On = stopped, Off = starts again. Perfectly repeatable. Other members of my family had no trouble, just me. Watch was given to my sister, and I inherited my grandfather's pocket watch. I was the only kid at school who regularly wore a turnip.

            Fast forward many years, and I was given a self winding watch as an engagement present (one of a pair, SO had the other). Again, would not work on me, but would on the mantelpiece or anyone else's wrist. I now just keep a cheap mobile phone in my pocket with no SIM and Night Clock running on screen.

          3. Norman Nescio

            Re: Reminds me of an old (early '80s) AI koan ...

            My grandfather could not wear wristwatches--they would stop working.

            My mother was the same - with mechanical wristwatches. For many years she wore a nurses watch (the one that hangs upside down from a safety-pin/brooch pinned on the front of a blouse or other garment), until the glorious day when suitable LCD wristwatches became available.

            I've often wondered what it is that causes some people to banjax mechanical wristwatches. They would usually stop working within a week on my mother.

            1. jake Silver badge

              Re: Reminds me of an old (early '80s) AI koan ...

              "I've often wondered what it is that causes some people to banjax mechanical wristwatches. They would usually stop working within a week on my mother."

              This is one of those things that I'd normally poo-poo without proof ... but I can't wear a mechanical wristwatch. They die on me within 10 hours. Quality doesn't matter ... Timex to Rolex, it'll die, guaranteed. Pocket watches, on the other hand, are just fine ... and I have a rather old dive watch that is still happy, but I have never worn it for more than three or four hours at a time, so whatever is causing it appears to not be cumulative. Grandad's Cartier stays in it's box until I find a proper home for it.

              I wear a standard watch on the inside of my wrist, the dive watch on the outside. Just a data point.

              It'd be fun to investigate this further, but who has the time or money to wreck perfectly good watches? And besides, I wouldn't wear one anyway ... these days, it's hard to find a place where one can not see the time.

              Note that I'm perfectly aware of the difference between testimonial and proof.

              1. Alan Brown Silver badge

                Re: Reminds me of an old (early '80s) AI koan ...

                "And besides, I wouldn't wear one anyway ... these days, it's hard to find a place where one can not see the time."

                I've seen a couple of editorials which point out that the rise of mobile phones (even the humble Nokia 3100) has resulted in near extinction of wristwatches - and if you look around you'll see that's pretty accurate (compare videos of people on the streets in the 80s/early 90s with observations today)

            2. Alan Brown Silver badge

              Re: Reminds me of an old (early '80s) AI koan ...

              I gave up on mechanical wristwatches very early. As a leftie and wearing it on my RIGHT hand, cheap watches would inevitably self destruct when the winder would snag on something and rip clean out of the case (you'd be amazed how often you brush your wrist against things)

              recessed winders solve that, but so do lcd watches and they were MUCH cheaper

        3. An_Old_Dog Silver badge

          Testing, Testing ...

          When I was at uni, I was that person (tester). Some of my upperclassmen/women had asked me to test a program they'd written for a class project, and I happily did so.

          First time I used it: I found an error; and they went off to fix it. Second time I used it: I found a different error; and they went off to fix it. Etc. Four days before their project deadline, they went off to fix the nth bug I'd found, but did not ask me to test their program again. Nor did they ever ask me to test any of their other programs.

          1. Alan Brown Silver badge

            Re: Testing, Testing ...

            "I found an error; and they went off to fix it"

            My experience was frequently that the error was repeated elsewhere but only fixed where I'd found it. The concept of looking for it across the code seemed foreign to most programmers

      3. Strahd Ivarius Silver badge
        Joke

        Re: Reminds me of an old (early '80s) AI koan ...

        Don't forget that you have to turn it off and on "again", otherwise it doesn't solve any problem

      4. Eclectic Man Silver badge
        Devil

        Re: Reminds me of an old (early '80s) AI koan ...

        "Power-cycling with no understanding of what is going wrong is how 90% of the issues I have with my computer are fixed!"

        Ah, but there is 'switching it off and on again', and finding the 'intelligent' on-off switch no longer works and you have to do the good old 'mains power cycling', because the bloody thing will simply will not do as it is told and 'power-down' gracefully.

        Now that is real SysAdmin, that is.

      5. Anonymous Coward
        Anonymous Coward

        Re: Reminds me of an old (early '80s) AI koan ...

        I have literally told people that computers love me to avoid otherwise lengthy musings on why the Computer worked for me. Either that, or we have a special way of turning it off and on. Gotta love Windows

        1. Robert Carnegie Silver badge

          Re: Reminds me of an old (early '80s) AI koan ...

          Some of us have learned, maybe unconsciously, to switch off, wait a few seconds at least, switch on. A confident firm hand in operating the switch also is best - but not too firm.

      6. Bebu Silver badge
        Windows

        Re: Reminds me of an old (early '80s) AI koan ...

        "Although I am reminded of all the times that stuff starts magically working as soon as the relevant expert is watching from over somebody's shoulder."

        Quantum version of sod's law? The superposition of the good fairy and bad fairy (gremlin) wave functions that describe the operation of the device collapse with the measurement by a competent observer into the "good fairy" state. The normal user always gets the gremlin.

        I never considered Windows in any form attracted anything but the most malign forces of the world.

    2. David Robinson 1

      Re: Reminds me of an old (early '80s) AI koan ...

      "It is not a matter of knowing which button to press but knowing how to press it."

    3. Jan 0 Silver badge

      Re: Reminds me of an old (early '80s) AI koan ...

      It's pretty sad that none of your respondents has another anecdote about a Lisp machine. Surely somebody here had a recalcitrant Linn version?

      I was more of a Poplog afficionado, but there was no dedicated hardware AFAIK.

    4. Danny 14

      Re: Reminds me of an old (early '80s) AI koan ...

      ive seenthis before when a windows service really needed a delayed start but also was dependent on another delayed start. Sort of the opposite of a race condition which i call the british queue, "you start first sir! No sir, I insist you start first!" etc.. in the end nothing starts and deriratives all grind to a halt.

      1. Jou (Mxyzptlk) Silver badge

        Re: Reminds me of an old (early '80s) AI koan ...

        I've seen that on poor written software, but have not seen such a thing from a Microsoft service yet.

        1. hplasm
          Alert

          Re: Reminds me of an old (early '80s) AI koan ...

          "I've seen that on poor written software, but have not seen such a thing from a Microsoft service yet."

          ** pleonasm analysis failed successfully- tautology detected**

  2. GlenP Silver badge

    PC Engineers...

    I've probably recounted the tale previously of a 486 based HP SCO Unix box I managed for a while. I was on holiday but had taken the company "mobile" (a car phone attached to a blooming great battery pack) just in case. Half way through I got a call, the backup tape wasn't ejecting, what should they do?

    I balanced the risks and decided the best thing was to wait until my return a few days later and I'd sort it.

    That apparently wasn't good enough, they summoned an engineer, against my instructions and without telling me, from the parent company's PC support people. He travelled the 150 miles or so from the South Coast, turned the box off without any thought of a controlled shutdown, manually ejected the tape then turned the box back on. When it wouldn't boot again he ran away after saying, "Oh shit!" or words to that effect.

    It took a few days to completely wipe the file system, reinstall and reconfigure Unix then restore the last good backup tape which by then was over a week old. Had it been my own fault I'd have pulled an overnighter and got it back within 24 hours but as it was down to senior management I wasn't going to rush (you can always find an excuse to go home* such as running a disk check that will take several hours).

    *A DEC engineer once used a thunderstorm as his reason for not continuing to fix a fault, I didn't argue as I wanted to go home too.

    1. Anonymous Coward
      Anonymous Coward

      Re: PC Engineers...

      Arnie reminds me of one or two PC engineers I've had the misfortune to encounter in the past. They can't be bargained with. They can't be reasoned with. They don't feel pity, or remorse, or fear. And they absolutely will not stop... ever, until the server is dead...

      1. MrBanana

        Re: PC Engineers...

        One coworker of mine was barely a PC user, and unfortunately got sent to a customer site that had a problem with their Univac system. This was purely for putting someone onsite to fulfil our contracted support obligations. Under no circumstances was he to do anything without explicit instructions. He managed to get logged onto the system, and diagnosed the problem as an out of space issue - this bit was actually correct. Unfortunately, as root, he took it upon himself to "tidy up" the system, and generate some free space, "/vmunix - that looks big....". After the inevitable crash, there was precious little of the system file space to enable the poor beast to boot, and we had to reinstall the OS from scratch.

        1. An_Old_Dog Silver badge

          Univac-Flavored Unix??

          When did Univac/Unisys ever run Unix on its systems?

          Burroughs (at some point absorbed into Sperry-Rand-Remington-whatever-they-called-themselves) ran Master Control Program as their operating system; Univacs ran EXEC, and some other OSes which Univac-and-its-successor companies later developed or bought, but not, AFAIK, Unix.

          1. jake Silver badge

            Re: Univac-Flavored Unix??

            "When did Univac/Unisys ever run Unix on its systems?"

            Officially? Starting in mid 2003ish, with the 2.6 Linux kernel.

            Unofficially, a trifle earlier. Probably late 2001 or so.

      2. Anonymous Coward
        Anonymous Coward

        Re: PC Engineers...

        Arnie here....

        In my defence I did a proper restart, not a forced one.

        Also medical software is some of the worst I've seen in terms of requirements to run. One obsolete version of dotNet/Java/IE that had to be installed during a full moon while chanting ancient Latin. If something goes wrong you have to WIPE the whole PC and start again. This is for PACS - a glorified picture viewer for X-rays, so didn't see out of the ordinary.

        (There was that video editing software that I had to support years before this. Was Windows 3.1 based, but you had to run it under OS/2 in emulation mode)

    2. UCAP Silver badge
      Joke

      Re: PC Engineers...

      A DEC engineer once used a thunderstorm as his reason for not continuing to fix a fault

      Well trying to fix faults when there is a thunderstorm lurking about can be pretty shocking.

      1. Stuart Castle Silver badge

        Re: PC Engineers...

        Love the description of a thunderstorm lurking, as if they hang about on street corners like Teenagers or something..:)

        1. Anonymous Coward
          Anonymous Coward

          Re: PC Engineers...

          >Love the description of a thunderstorm lurking, as if they hang about on street corners like Teenagers or something..:)

          This last weekend they certainly seemed to. We had at least 3 including one really spectacular one on Saturday night.

          1. David Hicklin Bronze badge

            Re: PC Engineers...

            > We had at least 3 including one really spectacular one on Saturday night.

            Where I am in the East Midlands we always seem to miss the worst weather, never get the spectacular thunderstorms, snow piled up deep.....really quite boring!

        2. PRR Silver badge
          Trollface

          Re: PC Engineers...

          > Love the description of a thunderstorm lurking, as if they hang about on street corners like Teenagers or something..:)

          {as AC above says:} They DO "lurk". I'm on the coast of Maine. That rainstorm that washed-out NYC over the weekend is "lurking" on the Boston to Quebec line waiting for me to leave the house tomorrow.

          Image

    3. Anonymous Coward
      Anonymous Coward

      Re: PC Engineers...

      SCO Unix box dies if power is cut without proper shutdown?

      1. Doctor Syntax Silver badge

        Re: PC Engineers...

        Remember it's a server which will be running some application serving multiple users. There'd probably be a lot of data in un-flushed buffers. It will have left user data in an inconsistent state. The OS itself would likely boot up but not the application service on top of it.

      2. Peter Gathercole Silver badge

        Re: PC Engineers...

        Back in the dim and distant days of non-journalling filesystems, such as the original UNIX Edition 7 filesystem (which was even more primitive than what is generally referred to as UFS), where free space was held as a linked list of unused blocks, just turning off a system could lead to a situation where a filesystem 'lost' all of it's free space because the updating of the list would not be completed.

        This was especially important to make sure that filesystem superblock (which contained the pointer to the start of the free list) was written out to the disk before pulling the power, as it became possible to lose the entire filesystem as a result, with almost no chance of recovery (wasn't ancient UNIX so interesting!)

        This was back in the days when fsck was a new tool, you may have used icheck, ncheck and dcheck in single user mode, and you might have thought about using fsdb to fix a filesystem (slightly later than fsck, but not by much) by hand. Sometimes, the wise decision was just to go to the backup.

        In addition, quite often updates to inodes and other filesystem control structures would be written to in-memory copies, with a write to the underlying disk storage being put on the I/O queue for the device to be carried out at a later time (anybody who is familiar with "sync;sync;sync;kill -1 1" will remember these times).

        Then remember that SCO Unix (and before it Xenix - originally a UNIX Edition 7 port) goes back a long time, so it is quite possible that the article was referring to an early SCO operating system. This could have been well before Windows was even a thing!

        1. heyrick Silver badge

          Re: PC Engineers...

          I recall one device (SJ MDFS file server, if I recall correctly) that "solved" this problem by starting up with the entire filesystem marked read only until it had had time to walk the entire filesystem to determine what was used and what was free space.

          Thankfully in those days the harddisc was only a couple of hundred megabytes, but the little Z80 did take a good few minutes. Though, to it's credit, it was fully functional as a server during, so long as you didn't want to write anything.

          1. Peter Gathercole Silver badge

            Re: PC Engineers...

            The main problem with early UNIX filesystems was that UNIX was already fully multi-tasking and multi-user, even in the '70s. This often meant that if you. for example, deleted a file so that it's space was put back on the free list, it was possible that the space was immediately used by another user creating or extending a file (I had to keep explaining this when users of CP/M and MS/DOS asked me to undelete a file they had just deleted when using UNIX). And as even relatively low-powered systems would support 8-16 users, if the system stopped unexpectedly without writing out the dirty disk blocks and inodes, it could be unclear which file an orphaned block (one that was not on the free list nor in the blocks allocated to files) belonged to. And worse, it could sometimes be that a block could appear in the block list of two files.

            The tool to fix such a problem before fsck (fsck was first provided outside of AT&T on the V7 Addendum tape which was sent with later Edition 7 distributions for PDP-11s) was icheck, which was not interactive. You would get a list of problems, and then go through fixing them with rm and clri, and then have to run icheck again and rinse and repeat until you had sorted out all of the issues. And then you had to sort out which files you'd lost or damaged. (I made sure that I did most of this on a hard-copy terminal when I had to, so I had a record of what had been done)

            The commands I referred to were icheck, which would check all the blocks listed against files, and reconstruct the free list from the blocks explicitly not in files, dcheck which would check whether all of the allocated inodes had appropriate directory entries and that the link counts were correct (useful for cleaning up pipe-files), and ncheck (not really a check) which would go through all the inodes, and list the directory entries which referenced that inode or reported an orphaned file (interestingly, ncheck still exists on late UNIXes like AIX, and is often useful to look at a filesystem that may have directories and files hidden by overmounted other filesystems or point-to-point mounts without unmounting the covering object). Icheck and dcheck were been completely replaced by fsck, but remained for several versions.

            Some wag at AT&T created an exp-tool called ipatch (arrrrr!) to help with cleanup, which would allow you to change individual values of an inode without having to dive in with a full blown fsdb session.

            These sorts of thing just didn't happen with single-user OS's, although that was obviously not the case with the system you're talking about (I knew I recognised it, but I had to admit that I had to look it up before it clicked), but that was a file-server, so the circumstances were a little different. With BBC micros with ordinary NFS or DNFS, file operations were quite often full load and saves (although Econet did support opening files as read, write and even update), but if I remember correctly, most operations were whole-block reads or writes, and files were normally contiguous blocks on a disk (at least on Level 3 econet servers which I had experience of - defragging the disk was a real pain!). I believe things were a little different with ANFS, but I had little experience of that.

            1. J.G.Harston Silver badge

              Re: PC Engineers...

              ANFS doesn't change the file system vs NFS, they are both the *CLIENT* end, absolutely everything in a networked filing system is implemented at the *SERVER* end, regardless of client. If the server doesn't implement it, there is absolutely nothing you can do to the client, the server will not implement it, because it's the server that implements it.

              1. Peter Gathercole Silver badge

                Re: PC Engineers...

                When I was talking about ANFS, it was really the file block caching that it would do that I meant. And that was effectively a change to the client that improved the entire experience. Reading a data file through the original NFS actually transferred the file one byte at a time, because NFS did not provide any memory pages for doing block reads (unlike DFS, that read whole sectors at a time). This made file read and write/update very slow. I actually wrote a small piece of 6502 assembler that trapped the OSBGET and OSBPUT vectors, and implemented a single page buffer using OSGBPB to speed up file access for several BBC languages.

                This speeded up the from-file compile time for Acornsoft Pascal about 20 times, and also reduced network contention (making our BBC micro lab a much more useful resource), as the 'net would slow to a crawl when transferring one byte at a time.

                I am aware that there was not a one-to-one mapping between the filesystem that the BBC used, and that actually running under the covers on the server. Acorn Econet Level 3 server used ADFS under the covers, which did reasonably closely match what NFS and ANFS would do. What was under the covers in SJ fileserver, I don't know, but as Acorn NFS was just so simple, it could have been anything provided it kept to the API. But I think that the file naming convention and directory structure was probably defined by the server, not the client, but was exposed to the client, so the server was not totally transparent to the NFS client.

        2. WonkoTheSane
          Coat

          Re: PC Engineers...

          "This could have been well before Windows was even a thing!"

          Knowing the UK's National Health Service, this was more likely to be last week!

        3. swm

          Re: PC Engineers...

          We had a file server at Xerox that powered down when the building lost power. There was a good file scavenger program that could put all of the pieces together etc. Running it showed bad records in a spiral pattern across the disk. Evidently, the heads retracted while still writing!

          (Backups restored everything.)

        4. Norman Nescio

          Re: PC Engineers...

          anybody who is familiar with "sync;sync;sync;kill -1 1" will remember these times,

          Oh yes. I have seen claims that modern journalling file systems don't need sync. It might be superstition, but when I need to do Alt-SysRq REISU[B|O], I do 'S' three times.

          And not all file systems in use journal. When working from the command line with FAT-formatted USB flash memory drives, sync is helpful to ensure that a large file copy is complete before removing the drive from the USB socket. Your Desktop Environment might well be nice enough not to allow you to press the 'Eject' button before writing is complete, but the buffers are not necessarily written out to the drive when the cp command returns.

          As for why you might need/have needed to do sync thrice, these answers gives some background:

          ServerFault.com: Is execution of sync(8) still required before shutting down linux?

          and

          UNIX StackExchange: Is there truth to the philosophy that you should sync; sync; sync; sync?

      3. Jou (Mxyzptlk) Silver badge

        Re: PC Engineers...

        Of course! The filesystems were far from what we take for granted now. No journaling as the best known example. No snapshots. Simple management structures prone to corruption beyond any repair. Making the writes to update files organized in a way that the changed block is first written somewhere else on the disk, and THEN update the management structures, for example the FAT, was uncommon back then.

      4. hplasm
        Linux

        Re: PC Engineers...

        "SCO Unix box dies if power is cut without proper shutdown?"

        Just like the SCO court case...

    4. John Miles

      Re: PC Engineers...

      When I started work - whenever there was a thunderstorm in area we'd shut the computers down, mind you this was in the 80s with HP computers probably 2 decades old, controlling test equipment, and no UPS so probably a good idea.

      1. heyrick Silver badge

        Re: PC Engineers...

        I do that these days. Thankfully a modernisation programme means that I no longer have four bare wires coming to the house as before (it's a twisted bundle of insulated wires), but then I'm only a half kilometre of wire from a transformer mounted to a pole fed by an overhead 11kV line.

        I've lost a router, which I think was due to induced current from where the phone line ran just below a length of medium voltage line, and in the past have had tungsten bulbs burst. Lightning here bites. I do wonder if the fancy smart meter would survive such a thing. The old mechanical one did (and quite a smell of ozone from it afterwards!).

        For that reason, everything is WiFi, no wired network, and everything electronic gets unplugged if there's a storm nearby. Taking no risks.

  3. Anonymous Coward
    Anonymous Coward

    And Arnie? He will not be back.

    Ah, too bad. Seems the "blame meeting" converged rapidly onto him ...

    1. Dabooka

      Re: And Arnie? He will not be back.

      Probably correctly from my reading of it?

      Maybe a bit harsh.

      1. Pascal Monett Silver badge

        No, not really.

        Doing anything to a box you know nothing of, especially in a server room, is a big no-no.

        1. Strahd Ivarius Silver badge
          Coat

          It is the NHS, the server was under the tea trolley...

        2. Eclectic Man Silver badge

          Indeed, there would have been no professional indemnity insurance (probably). He should have at least checked whether it was 'serving' any life-critical activities before touching it.

          1. Anonymous Coward
            Anonymous Coward

            Arnie here -

            IT doesn't touch those - they fall under medical electronics. The amount of times I have been called out to them, only to discover it's not out kit - I wouldn't even clear the paper jam if it had a printer.

        3. Anonymous Coward
          Anonymous Coward

          Arnie here.

          Agreed - however I was 30 miles away at the time!

    2. Anonymous Coward
      Anonymous Coward

      Re: And Arnie? He will not be back.

      No way of knowing if the manager came back, or brought his nephew in again.

    3. heyrick Silver badge

      Re: And Arnie? He will not be back.

      "Seems the "blame meeting" converged rapidly onto him ..."

      While Arnie isn't blameless here, when nepotism is involved they'll be actively looking for somebody else to blame.

      Reading these columns, it's amazing how many times nobody ever says "this isn't what's supposed to be here" and back away. I mean, different machine, different OS... that's as good a reason as any to walk and get a second opinion (ie cover your arse).

      1. anothercynic Silver badge

        Re: And Arnie? He will not be back.

        If everyone did that, we wouldn't have any Monday morning musing material, mate! Life would be booooooooooooring (although that's how sysadmins like life).

      2. doublelayer Silver badge

        Re: And Arnie? He will not be back.

        It probably happens more often than not, but nobody writes in about the time they didn't find what they expected, walked away, found someone else, and it got fixed. It would only end up here if the second opinion didn't help and the box went down anyway.

        However, having dealt with my fair share of boxes with uncertain support, I get why it happens. I've conducted tasks on systems that didn't seem right, but I either couldn't get a second opinion at all or the people I could find didn't know any more than I did about what the system was doing or how. They still wanted it fixed, though. For context, I'm a programmer and work mostly with programmers, so not everyone has as much administration experience as I would like them to, and not everyone on the team is a good choice for disentangling an ambiguous config. My general solution to this situation is backups and walking very slowly through the system to make sure I've thought of several things before I utter that sentence "Well, that's all I can think of".

    4. Dave559
      Coat

      Re: And Arnie? He will not be back.

      It sounds like he…

      Terminated, and did not Stay Resident…

      (Yes, yes, wrong "OS" (ahem) entirely, but I'm not going to let that get in the way of a bad pun…)

    5. Zarno
      Terminator

      Re: And Arnie? He will not be back.

      One might say, he could have been Terminated?

    6. Anonymous Coward
      Anonymous Coward

      Re: And Arnie? He will not be back.

      Nope. Still here. I pointed out that a) what I did was under a generic login, so server was not configured properly and the database was installed by the vendor.

      The culture was such that if you make a mistake and own up that would be more or less fine, not one of these make a mistake and hide it in case you lose your job.

  4. Korev Silver badge
    Terminator

    Network problem

    I was expecting a network problem for Arnie because it hadn't been terminated properly...

  5. Blofeld's Cat
    Terminator

    Hmm ...

    Let's see: unexpected hardware, unfamiliar software, not the system you installed ...

    Use the pink paint, deploy the SEP field and GTFO as rapidly as possible.

    1. b0llchit Silver badge
      Coat

      Re: Hmm ...

      Unfortunately, they didn't apply BistroMath to settle the account.

  6. wyatt

    Ah yes, the joy of software that runs as an application on a server. Frequently we'd have to support customers who had installed updates (go them!) but failed to complete the process by logging the server back on to the required account. As regulated customers the lack of records could prove expensive..

    1. Peter Gathercole Silver badge

      If a system requires you to log in on the console as a specific account, I'd argue that it was not a proper server at all, more like a workstation running a peer-to-peer application!

      1. Fred Daggy Silver badge
        FAIL

        Don't assume ... always question ...

        But remarkably common practice. Especially among single/small team developers who haven't got their head around Windows. Or running as a service, or not running as root/administrator, or assuming everyone else is.

        What is it, 20 years since MS published guidelines about not using these account. More than 30 years since Windows NT 3.5X, but especially Windows NT 4. Or at least Windows 2000 ... which at least had a usable GUI and a real Administrator account.

        The first problem solving question from a "Un-helpdesk" droid is normally about "Is the user an administrator? Have you disabled UAC? Have you tried running it as Administrator?" So, they are still out there.

        1. JulieM Silver badge

          Re: Don't assume ... always question ...

          A lot of software development for Windows was done in back bedrooms using dodgy pirated copies of the proprietary development tools, without printed manuals describing best practices.

          As a result, much "legitimate" software depended for its normal operation on the same fundamental techniques used by viruses and worms, and broke in any environment that tried to impose any sort of security.

        2. Strahd Ivarius Silver badge
          Devil

          Re: Don't assume ... always question ...

          Install LAPS on the server, with a GPO to change the local admin account name, and then enjoy the fireworks...

  7. ColinPa

    It's never been installed

    Someone told me that they worked on a project where the development was done in one country and the testing was done in a different country. Due to the executive announcing the customer date before development was quarter done, it was pretty chaotic and corners were cut.

    The installation team followed the process of getting a clean machine and trying to install the product. It always failed. They executive blew his top and sent his top expert to "SHOW THE INSTALLATION TEAM HOW TO DO IT". The guy got off the plane, went directly to the lab and tried to install it. It didn't install. He spoke to the development team - they hadn't installed it either - they just replaced the binaries every day and were careful not to turn the machine off. I think no one had actually been given the job of writing the installation process, people just used some scrappy notes.

    In the development team, because the project was running late, they were asked to resize the work. Now they knew what they were working on, instead of some chart-ware design, they resized it properly - and said it had gone up from 3 months work to 9 months work. The executive blew his top as this was not the answer he wanted to hear. He wanted them to resize it again! but someone told him - every time they resize it - it costs 2 weeks work. Do you really want to ask them?

    1. Pascal Monett Silver badge

      Re: resize it again!

      Yes, because obviously if it didn't give the proper result the first time, Shirly it will the next time, right ?

      Ah, manglement. They deserve every meme they get.

      1. Anonymous Coward
        Anonymous Coward

        Re: resize it again!

        It's not about results, it's about power.

        (And money. Money is power.)

    2. Anonymous Coward
      Anonymous Coward

      6 weeks

      I was working on an installation in Germany of a product from the US of Texas.

      We could not get it to correctly configure for our environment. I edited, tweaked, wiped, reinstalled.

      Customer was getting angry at delays, Vendor was showing how it worked on their Dev environment and blaming me.

      We had a marathon session where I reinstalled and configured from scratch with their team remotely watching.

      They agreed I had done everything right.

      3 days week later they rebuild their code, issued a patch release and it worked.

      Many drinks later with a friend in the EU branch I found out why - they'd HARDCODED the path to the config file in their dev environment.

      1. cookieMonster Silver badge
        Pint

        Re: 6 weeks

        I've attached a pint, I have had to suffer similar difficulties in the past.

        For some reason each time I requested that those responsible be beaten with mackerel and then fired, my request was always denied

        1. Geoff May (no relation)

          Re: 6 weeks

          Of course the request is denied, mackerel do not deserve that kind of treatment.

        2. heyrick Silver badge
          Happy

          Re: 6 weeks

          A soggy piece of linguine would be my preferred method. Forget firing them, this is terminal levels of rage (and ought to be quite cathartic given that murder by linguine might . . . . . take a while).

          1. Strahd Ivarius Silver badge
            Angel

            Re: 6 weeks

            Don't fire them, send them to a support center, dealing with end-users...

      2. anothercynic Silver badge

        Re: 6 weeks

        I absolutely detest hard-coding. I'm in the midst of ripping hard-coded values out of code at the moment, and at points I can only go "Whyyyyyyyyy!?!?!?!?!!?!?" when they have a perfectly good config file to shove stuff into!

        And thus, I am now shoving things into the config file and codifying it there (so at least we know what the hell we need).

        Hard-coding is the QAD way of doing it, but then when it needs moving somewhere else, it becomes a tangled mess.

        1. david 12 Silver badge

          Re: 6 weeks

          If it's not in ROM, it's not "hard" coded.

        2. David Nash

          Re: 6 weeks

          It's because it was added as a temporary implementation. Of course it's there for ever.

      3. OzBob

        Re: 6 weeks

        Re: Hard-coding. I was at polytech in NZ in the late 80s and some of my part-time classmates were coming in looking in rag order. I asked what the problem was and they said that GST (NZ version of VAT) was changing from 10% to 12.5% and everyone had hard-coded the original value into their programs, rashly assuming politicians would not change their minds. So they were working all hours to find each value and convert it to a variable (in the days of mainframes and really crap IDEs).

        1. Bebu Silver badge
          Headmaster

          Re: 6 weeks

          (NZ version of VAT) was changing from 10% to 12.5% ... assuming politicians would not change their minds.

          I doubt that politicians possess any such function Implemented in what passes for their brain viz mind.

          I quite readily believe depending on the wind direction that they could not only change the rate of VAT but also apply multiple rates to different products and services. I would allow for a table of rates v. product etc class to be loaded from a configuration file and hooks to enter or deduce the product class.

          1. Alan Brown Silver badge

            Re: 6 weeks

            New Zealand's VAT/GST implementation is unusual in that the GST is applied to EVERYTHING except banking(financial) services

            The huge mess in Australia and Britain, "where bread was exempt but raisin bread was not" was one of the major drivers of that decision. The government didn't want a system which had been demonstrated to cause expensive court cases over interpretations (Jaffa cakes, anyone?)

            The result of that decision was astounding.

            GST went in at 10%, income taxes reduced, the hideously complicated system of import duty and sales taxes (which could result in tax being 120% on some items) went away.

            The government netted MORE income and within a decade was able to reduce inland revenue staff numbers by 1/3 (nobody in IRD was complaining of overwork either), vs being on track to double staffing numbers on the existing system, within a decade just to keep up

            By not having exempt categories, business calculations were easy and it became _very_ hard to cheat the system. In most cases if a rort was going on the government only lost GST income on the MARKUP applied, rather than the entire tax amount (which was a constant feature of a sales tax system only applied at retail sale and required that wholesalers be specially licensed/only allowed to sell to retailers)

            1. david 12 Silver badge

              Re: 6 weeks

              I did an implementation in AUS. Yes, the biggest part of the implementation cost was handling the two categories of transaction. Only two, but that meant there had to be a database field to assign the categories, and calculation at the line-item level, and calculation and display branches, and display fields to indicate GST status for each line item. And I had it easy: the mainframe implementation ran more than a month late, and major suppliers stopped supply of millions of dollars of product while they waited for payment of outstanding invoices held up in the GST implementation.

              Which all could have been avoided by simply boosting the pension rates and tax cut-off rates as compensation, but the party in power had invested so much moral capital in opposing GST, painting it as an evil right-wing plot, that when they came to implement it anyway, they had to agree with every 'injustice' claim as a special case.

        2. Alan Brown Silver badge

          Re: 6 weeks

          "everyone had hard-coded the original value into their programs, rashly assuming politicians would not change their minds"

          I knew people who worked in the vehicle registration system. Management swore black and blue that personalised registration plates would never be deployed in New Zealand and ordered the staff NOT to allow for it in the new system they were developing

          About 8 months after the new system came online, personalised registration numbers were announced as a new policy from the Beehive. Thankfully said manager had been ignored and modules written "just in case"

          I crossed paths with him myself about a decade later.

          He was attempting to do offsite backups

          By emailing the database files to an office in another city

          Said files were ~90MB

          Over a 2400bps dialup modem

          Using a mail client which didn't declare the size of the message being sent

          Starting at 5:30pm each day

          That was when I discovered that SunOs 4.1.3 sendmail uses the swapspace partition (containing /tmp/) to buffer messages too large for memory

          About the same time as I discovered WHY the system was running out of swap around 2am Tuesday to Saturday was when he sent a snide mail (dead trees) complaining that our service was unreliable. Pointing out the default 10MB limitation on email size was simply unacceptable to him (the receiving end also had a 10MB limit and had no intention of raising it, so bumping my limit was pointless).

          It turned out he'd sold "email backups" to the higherups as a great way of reducing costs and had received a large bonus after demonstrating it - with 100kB files

          He left the MOT shortly afterwards and then showed up at another large customer. A friend of mine worked there and whilst technically under him, was aware of the damage he was capable of (more importantly had a good personal relationshop with the company owners) and managed to prevent him making changes which would have impacted a multi-hundred million dollar mail order outfit

          About a year after that it was mentioned that this manager had left under a cloud. The owners are conservative religious types (Not American conservative types) and one of them had walked into his office to find him broswing penthouse.com on his work machine, at lunchtime. As 98% of the employees were female it was deemed to be a hostile workplace issue and he was walked to the front door by security within minutes (Did I mention that there was a specific prohibition on accessing unauthorised resources from the workplace network as part of all employment contracts? And a firewall which whould have stopped that website even being visible?)

        3. Terry 6 Silver badge

          Re: 6 weeks

          And yet, even when I was learning coding in school in the 70s and later teaching myself other stuff, like 6502 assembler in the 80s I was given a good inkling that you don't hard code data components. Especially not ones that will be used in a calculation. I'd somehow had it explained/taught to me that this was just inelegant and poor practice, even if you thought it'd never need to change.

          1. An_Old_Dog Silver badge
            Happy

            Hard Coding Numeric, Path, and Filename Constants

            I had an instructor which taught us not to do this in a very zen-like way. We started the course writing a program, which, each successive week, we had to modify, changing, among other things, all the constants.

  8. Phil O'Sophical Silver badge

    development team - they hadn't installed it either - they just replaced the binaries every day and were careful not to turn the machine off.

    Oh yes, we've all been there.

    That's why our QA engineer always started a new test run of a new version by installing the product exactly as the documentation said, no matter how well she knew the product by then. The result was often bug reports filled against product and/or docs, but we rarely had issues from customers.

    1. gnasher729 Silver badge

      At two or three companies I managed to get a process for installing a development environment for new developers.

      It started with a printed sheet of paper. The paper said “follow the instructions on this paper. If they don’t work, then ask for help, and change the instructions so they work”. That was needed because what’s on a brand new machine would change over time.

      And one part of the instructions was where to find the instructions as an editable document so the new guy could update them.

      1. John Brown (no body) Silver badge

        "And one part of the instructions was where to find the instructions as an editable document so the new guy could update them."

        I think my brain just farted. I read that as "edible document" and the rest of the sentence made no sense :-)

        1. jake Silver badge
          Pint

          Honestly Boss, John Brown (no body) ate my homework!

          Beer, to wash down the ... well, you know.

    2. anothercynic Silver badge

      I absolutely adored our QA team for doing just this back when I was in commercial software development. Inevitably they'd find something I'd just assumed would be fine (but wasn't if a metaphorical knob was turned a specific way or a metaphorical switch flipped), or I'd forgotten to update them on changed instructions and they found out the hard way, and it's made me appreciate the QA function so much.

    3. Elongated Muskrat Silver badge

      The result was often bug reports filled against product and/or docs, but we rarely had issues from customers.

      And that, laydeeez and gennelmen, is an example of QA functioning as intended. It's to stop us devs from marking our own homework.

      A QA tester like that is worth their weight in HP inkjet ink.

    4. Terry 6 Silver badge

      And in my experience this is a place where instructions often go wrong. One of the places*. The writer knows what's meant to happen. And what every step does. So there are all sorts of ambiguities or missing details that the customer won't have a clue what to do with. As in something along the lines of; The writer, presumably, knew that for a given function of a device you had to keep a certain button held down for 5 seconds. The instructions say "Hold down the button n then press button y" but no mention of the 5 seconds or that you have to release that button before you press the other one.. The user presses n, immediately presses y and then snatches their fingers of the buttons like the device was on fire. And nothing happens. Or the device acts as it would if button n was pressed for normal use.

      *Even more egregious imho are the instructions that explain the simple and obvious steps- but carefully omit the bit that's hard to explain.

      1. Alan Brown Silver badge

        "One of the places*. The writer knows what's meant to happen. And what every step does."

        In my last job I was expected to write the instructions and inflict them on users with no further in-house testing

        I know full well that marking your own work is a deadly mistake, but the manager REFUSED POINT BLANK to assign someone else to handling this part of the job

        Unsurprisingly there were lots of user complaints about stuff and documentation issues took years to be addressed

        The same problem cropped up with in-group procedures - even worse, after going through instruction scripts, finding the holes (missing steps. assumptions, etc) I was roundly criticised for treading on people's toes (they were having to write their own docs too, but resented suggested edits being handed back and refused to add them to their official versions, believing the instructions were complete enouigh as they were - this also applied to stuff they'd written for the users).

        The result (unsurprisingly) was lots of do-overs and support requests

        These are the same people who wouldn't document all the user support they were doing ("stopped in hallways" was a common way of users filing issues instead of using the ticketing system and our people woudln't create/close a ticket when they got back to their desk), resulting in manglement coming down hard from on high about the staff supposedly sitting around idle and not doing much - wanting to reduce headcount as a result.

  9. Notrodney

    Printer Problems

    I was sent to a remote site to install some new software many years ago. The senior engineer had mentioned (with a grin) that although he had never been to the site, the younger lads in IT were always popping in there to fix 'printer problems'. The printer seemed absolutely fine, but oddly enough the young lady who worked there was absolutely stunning.

    1. Anonymous Coward
      Anonymous Coward

      Re: Printer Problems

      My younger self had a customer with weekly problems requiring that I go on site on Fridays, after my normal work.

      It was an actor agency, and the issue was cabinets overflowing with Champagne bottles...

    2. An_Old_Dog Silver badge

      Re: Printer Problems

      "No problem, I'm just doing a routine [printer|JetDirect] firmware upgrade, carry on."

  10. Howard Sway Silver badge

    the server had been configured by the son of a senior manager at the facility

    Having read the story, that's a very generous use of the word "configured". "Had the installation disk inserted when delivered" is most certainly not the same thing as configuration. But what did they expect when the manager thought he could save money because "my lad could do this he knows how to work these computer thingies"?

    So, not really a "Who, me?", definitely a "Someone else's fault".

    1. Anonymous Coward
      Anonymous Coward

      Re: the server had been configured by the son of a senior manager at the facility

      @Howard Sway

      Oh puhleeze,

      When it comes to an I.T. problem, it is ALWAYS someone, anyone but their fault.

      1. The Oncoming Scorn Silver badge
        Coat

        Re: the server had been configured by the son of a senior manager at the facility

        Another Zummerzet County Council story (That I've not repeated for a while).

        Ticket comes in for THE Head of the Council, nice plush office for "Email not working".

        Get there & discover Head's son had been using the machine for something unrelated to council business & covered his tracks by re-installing XP Home & without the usual Office suite etc etc.

        I removed the machine for investigation & have the ticket re-assigned to a permy member of the ITSC, as one of our ex-colleagues had upset a councillor by asking her politely not to park in our loading bay & had been terminated from his support contract with extreme prejudice within the hour.

    2. Strahd Ivarius Silver badge
      Facepalm

      Re: the server had been configured by the son of a senior manager at the facility

      Save money? are you kidding?

      He shared the money, with an added markup, with his son more probably...

  11. corbpm

    Sausage Factory

    I was once asked to look at a machine that had been running for a few years without apparently ever being switched off , so cautious approach, visual inspection.

    What's the strange card in the machine ?, no one had an idea.

    Where are the manuals for the machine ?, who was responsible for the machine, gone - possibly dead or they just didn't want to tell me.

    What are the ramifications of this machine never starting back up again ?, Dire, world ending (no more Scotch Eggs from this factory).

    Left it alone, backed away and declined to play guess what's wrong.

    I had no faith that it would ever start up again from cold, it was old enough that it had MFM disk drives , couldn't identify a inserted card so no faith i could build a new machine to replace it, disconcerting level of complacency from the management, and they had stressed i would be responsible if i couldn't get it working again if i switched it off.

    Apparently someone else came in after me chickening out and the machine didn't restart and the machine was down for a while.

    1. Anonymous Anti-ANC South African Coward Bronze badge

      Re: Sausage Factory

      Apparently someone else came in after me chickening out and the machine didn't restart and the machine was down for a while.

      You dodged a bullet there.

      1. Prst. V.Jeltz Silver badge

        Re: Sausage Factory

        well , what he described was pretty much Russian Roulette.

        1. Doctor Syntax Silver badge

          Re: Sausage Factory

          With all chambers loaded.

          1. A.P. Veening Silver badge

            Re: Sausage Factory

            With all chambers loaded.

            I would say with a pistol instead of a revolver.

            1. Strahd Ivarius Silver badge
              Coat

              Re: Sausage Factory

              This is the famous Soviet Roulette, that dissenters were allowed to play with a Tokarev TT30 because the Nagant M1895 was associated with czarism.

      2. DCdave
        Coat

        Re: Sausage Factory

        It could have been wurst, he might have ended up as mincemeat.

    2. Ken G Silver badge
      Facepalm

      Re: Sausage Factory

      The clue was in "they had stressed i would be responsible if i couldn't get it working again if i switched it off".

      They weren't ignorant they just were willing to pay for someone external to take the blame.

      1. Alan Brown Silver badge

        Re: Sausage Factory

        I've encountered clients like this a few times. My immediate response has always been to start packing up and head for the door

        If they're taking that attitude, then you will be blamed for ANYTHING that goes wrong, even if you weren't there

    3. NXM Silver badge

      Re: Sausage Factory

      "No more scotch eggs"

      On the surface this might seem to be a disaster, but my experience of Scotch Eggs from factories is that of a spheroid with something resembling a gobstopper inside with a coating made of lard and sawdust, coated in yet more sawdust.

      Make em yourself, they'll be far far better.

      1. J.G.Harston Silver badge

        Re: Sausage Factory

        A substantial meal.

    4. Alan Brown Silver badge

      Re: Sausage Factory

      Such a machine is a cleaner with a floor polisher(*) away from disaster

      I've had machines on sites with uptimes of YEARS suddenly stop working - in virtually every single case it's becase the cleaners unplugged it to run a floor polisher - despite the DO NOT UNPLUG/DO NOT SWITCH OFF label on the socket and plug itself

      At least one of them has stated that unplugging is not switching off, so they're alright jack.

      In a lot of other cases it's turned out that they cannot read (these are people handling dangerous chemicals FFS) and often blag their way through life trying to cover it up

      Using an _obviously_ incompatible plug/socket system for critical kit doesn't help. They'll unplug it anyway to check

      It's worth having penalty clauses in contracts for external organisations making them liable for damage caused by their staff for any reason and then insisting that the managers read/sign off on site rules/safety instructions. That way when the inevitable "nobody told me..." crops up, you can pull up the evidence that they WERE told and failed to pass it on or send their new staff to the safety briefings as required in the contract

      1. jake Silver badge

        Re: Sausage Factory

        "it's becase the cleaners unplugged it"

        For future reference, they make locking outlet covers that fit over an inserted plug or plugs, preventing the removal of same. The locks are trash, easier to pick than a file cabinet, but they work for this kind of thing. Under twenty bucks, and usually in stock at your favorite purveyor of sparky stuff. They make more expensive and harder to defeat versions, too. They come in both indoor and outdoor versions. Recommended.

        And of course sometimes nothing beats hardwiring the machine(s) in question into either a J-box or the breaker panel.

        1. Alan Brown Silver badge

          Re: Sausage Factory

          Kinda hard to force a client to install these at their own site.

          If you can get them to do that you can usually convince them to put the equipment in a server room off limits to cleaners

          (Unfortunately then you still have the problem of "security guards" who make it their life's mission to switch off aircon in unoccupied rooms - including rooms full of noisy equipment and racks)

          1. jake Silver badge

            Re: Sausage Factory

            "Kinda hard to force a client to install these at their own site."

            Nah. It's all in the presentation. "It would cost you upwards of $BIGNUM per hour if that server were accidentally unplugged. If I install one of these simple devices, at a cost of $smallnum (including labo(u)r), that possibility is removed forever". In fact, they are so cheap and easy to install that I've been known to install them gratis and not even bother mentioning it.

            "If you can get them to do that you can usually convince them to put the equipment in a server room off limits to cleaners"

            THAT is easy. The hard part is convincing them that the former closet now needs its own AC ...

            "(Unfortunately then you still have the problem of "security guards" who make it their life's mission to switch off aircon in unoccupied rooms - including rooms full of noisy equipment and racks)"

            So remove the switch from places accessible to "security guards". It's not rocket surgery.

        2. An_Old_Dog Silver badge

          Re: Sausage Factory

          Ah, so that's why so many factory machines I've seen were (previously, to me) inexplicably directly wired-in to the power, rather than equipped with plugs when it seemed to me to be more useful to be able to unplug them and move them around as needed.

  12. technos

    Service != Server

    I once offhandedly mentioned to a user that his label printer problem was because the service it relied on had frozen so I restarted it.

    What he apparently heard was "I restarted the server", because the next time he had trouble with his label printer, he walked into the closet they called a server room and pulled a bunch of power cables.

    Two machines and a switch didn't come back up. Two of them hadn't been plugged back in correctly, one dropped to single-user when it found a dirty filesystem.

    Worst part of the story is that, after dropping my date-night plans and bailing him out of an emergency of his own making, he told everyone the issue was a 'power blip' and stalled on approving my invoice. He wanted a significant discount because, and I quote, "The problem only happened because of your poor communication skills.".

    I emailed back: "How's this for clear communication? You can either authorize the invoice and forward it to accounting, or I will be informing the company leadership what caused the 'power blip' and having them authorize it.". There was some further whining about how it was going to come out of his budget and that I should 'have a heart' but I did get paid.

    1. Doctor Syntax Silver badge

      Re: Service != Server

      "There was some further whining "

      I think at that point I might have implemented your "or" branch without further ado.

      1. technos

        Re: Service != Server

        Naw.. I wasn't gonna blow perfectly good blackmail material when there were plenty of other ways to put pressure on him.

        Like roll it into my next one, with a late fee, send it to Accounting, and let them yell at him for not doing his paperwork.

        Ended up sitting on it for over a year and then used it to impress upon him the importance of dealing with the three employees running Gnutella and hosing the network, telling him that if I got one more baloney complaint that 'the network is slow' I'd bring pictures of the unplugged equipment and share them with his CEO as a funny story.

        1. Alan Brown Silver badge

          Re: Service != Server

          Sounds like the idiot in question needed removing from the position and what you achieved was job security by letting him keep trashing his employer's equipment (or letting him allow his underlings to do so)

          There's a risk with such users that they tell porkies to their management which results in your external contract going away

    2. J.G.Harston Silver badge

      Re: Service != Server

      I've had that: Our IT cock-up should come out of IT's budget, not our budget.

      1. Alan Brown Silver badge

        Re: Service != Server

        That's EXACTLY what's happened in a number of cases in one place I worked (A presitgious department in a Russell Group university)

        One serious IT cockup happened because the user in question (now deputy director) had done an end-run around the IT group because she was of the opinion her PhD trumped any of our experience

        1. jake Silver badge

          Re: Service != Server

          "she was of the opinion her PhD trumped any of our experience"

          Over the years I've discovered that the more traditionally educated a person is in any one field, the more likely they are to think they know more than anybody else in all other fields. By the time they receive their doctorate, they know so much about one, single narrow topic that they are absolutely convinced that they must know everything about everything (by osmosis, presumably). Doctors and Lawyers are the worst, at least when it comes to IT.

          Throw in politicians and you get the triumvirate that I refuse to do work for ...

          1. An_Old_Dog Silver badge

            Ph.D Knowledge Limits

            I've worked with many Ph.D doctors (medical researchers, mostly). Most were averagely-IT-knowledgable, and were reasonably-cognizant of the limits of their IT knowledge. A small minority were troublemakers, and a small minority had advanced IT knowledge, and used it in helpful ways, working with the IT staff.

            As to the troublemakers ... there was one we called "Doctor Gadget", who had nearly every "cool", modern gizmo hung off his Win 3.1 desktop, which had three parallel ports, two built-in serial ports, and an expansion card with two more serial ports. Driver conflicts? Port mis-configurations? Oh, yeah; he had 'em, repeatedly. Another we called "Doctor Franken-net." He had, in the days predating DHCP, plugged an Ethernet hub into a network wall jack on one subnet, and also into a network wall jack on a different subnet ...

  13. GrumpenKraut
    Facepalm

    The web server that wasn't

    Mid 1990s: customer had his own bow with Linux and a web server running on it. Few month later, after a power outage, customer calls us "The web server is gone!", me: "So it did not restart, right?", "No, the *server* is gone!", "OK, I'll have a look". Upon inspection there was nothing at all on the disk. Customer had installed everything in the installation system in RAM, nothing had every been written to disk. Took me quite a while to figure that one out. After new install I made the customer reboot the box to make sure he did not repeat the error.

    1. NXM Silver badge

      Re: The web server that wasn't

      That was Jurassic Park, Shirley?

      1. The Oncoming Scorn Silver badge
        Joke

        Re: The web server that wasn't

        You can't teach a old dog new T-Rex.

        1. Korev Silver badge
          Coat

          Re: The web server that wasn't

          No doubt one of those Dinobabies....

          1. Alan Brown Silver badge

            Re: The web server that wasn't

            When will the hurting stop?

      2. Korev Silver badge
        Linux

        Re: The web server that wasn't

        > That was Jurassic Park, Shirley?

        It's a UNIX system! I know this!

  14. t245t Silver badge
    Devil

    Server not configured to restart when the machine was rebooted.

    Sounds like someone was double-dipping ;)

  15. Boris the Cockroach Silver badge
    Pint

    The relative who

    'fixes things'

    Whenever I hear that phrase it takes me back 20+ years to a friend who worked much in my line of things.... however , his company had bought 2 spiffy new machine tools and as a whizz with the RS232, I got roped in to figure out why they couldnt talk to the PC.......(why do I never listen to that inner voice saying 'noooooooooooOOOOOoooooo'? )

    On arrival... its standard 25 pin d plug at the machine so the owner's relative... lets call him Andy, had got 2 25 pin d plugs and because the PC didnt have a 25 pin D plug on the back.. bought an expansion card that did(lets ignore the 9 pin D socket included as standard back then)

    So I plug everything in.... nowt........ lets have a look at the config... hmm using the right comm port.... ok machine setup... yeah its standard ok lets check the cable........ oh

    Lets take the D plugs apart.... Pin 1....... to Pin 1....... Pin 2 to Pin 2... errrr Pin 3 to pin 3... yeah we know where this is going...... its not a cross over cable as needed.. oh and hes not bridged out DTR to +5v(its an old machine that needs to see DTR in order for it to think theres a tape drive attatched)

    So after a spot of re-soldering and testing the machine can talk to the PC....... then I get... "can you show Andy howto program the machine... hes not used it before........." and thus went 2 more weekends.

    I did get beer AND money though

  16. John 110
    Coat

    aaah, NHS IT....

    Perennially underfunded and understaffed. Arnie will have been under pressure from the end users to fix the problem, whatever it took, 'cause we need it now! (now usually means occasionally (to print out holiday details...))

    I worked IT support for a diagnostic lab. My brief was to fiddle about with some half-arsed resourse management software strung together in a DOS spreadsheet by someone on a fixed term grant (who got a real job before they were finished) and to help our less-than-PC-savvy staff come to grips with a new lab reporting system (this was mid 90's -- any staff who had used a computer had had a ZX Spectrum or C64 as a kid.)

    As is usually the case, the job expanded to cover looking after all the PCs in the department (there was a central IT services, but you could wait weeks for them to fix anything) and as lab automation spread into microbiology, I was expected to liaise with the companies supply the analyzers and the lab system to make sure they talked to one another. -- Oh, and I was on my own with holiday cover provided by the world famous "it'll keep until I get back, or you could ask Ian in biochem..."

    So I walked through the door after a relaxing week of looking after the kids (school holidays) to be asked to look at the ELISA machine that had stopped talking to the lab system. I wandered down to the lab. No analyser.

    "Where's the analyser" "We moved it to the lab 2 doors down". Nobody had mentioned this before I went off...

    I located the missing machine and sat down to diagnose the problem. The PC and the analyser (RS232 connection) were talking, but no results were making it to the lab system. Checked that the PC was plugged into a node point. Check (surprise, I was). But there was no network connection.

    "You did have this node made live, didn't you?" "The whatnow?"

    Sorted that day by a quick grovelly phone call to Network support.

    "But only 'cause I'm going to be in that area anyway!" "Thanks Denise"

    1. Alan Brown Silver badge

      Re: aaah, NHS IT....

      "As is usually the case, the job expanded to cover looking after all the PCs in the department"

      without any change in job description, responsibilities or pay

  17. MachDiamond Silver badge

    Big giant CF

    I could see a narrative where the location needed a server, put in a request and the leaves grew and fell a couple of times with no server forthcoming. It came to a breaking point where a director has a family member that works on server installs and 'hires' them to come in and get something implemented. The HP box is spotted in its delivery company worn box in the corner and a glance at the label makes it clear it won't work in this situation. The family member writes up a BOM for all of the hardware, gets the software software configured and implemented working late nights and weekends. The 'temp' worker's post is folded up and all of the expenses to 'fix damage due to a broken water pipe' are tallied and properly assigned in the accounting. Providing documentation for the system might raise some flags so none is generated and submitted/stored anywhere. The HP machine is giving to the family member as a premium for getting the work done and not mentioning it to anybody.

    I've see this sort of thing happen a lot. There won't be a budget for tools, but there are always funds to make facility repairs for a borked water pipe so it's a great carpet to shovel miscellaneous expenses under. If you need a bunch of money, claim a backed up waste line. Nobody will squawk that you should have submitted a request up the chain, got 3 quotes and waited for permission.

    1. Alan Brown Silver badge

      Re: Big giant CF

      "The HP machine is giving to the family member as a premium for getting the work done and not mentioning it to anybody."

      In the case of the NHS, that's an ICO investigation waiting to happen

      Medical computers are in a class of their own WRT privacy issues and that applies if all they're doing is running the building climate control. Treat EVERYTHING as if it contains sensitive data and ensure everything is erased (not just signed off as erased) before it goes out the door

  18. dave 76

    rebooting is not a 'fix'

    I always tell people that rebooting doesn't fix a problem, it just makes it go away for an unspecified period of time.

    1. Bebu Silver badge
      Coat

      Re: rebooting is not a 'fix'

      I always tell people that rebooting doesn't fix a problem, it just makes it go away for an unspecified period of time.

      The wager is that the unspecified time will be on the far side of the "my problem" and the "not my problem" boundary.

      Pretty much the playbook of government and associated entities.

  19. Anonymous Coward
    Anonymous Coward

    Why isn't Windows Server Windows?

    Sure, Windows Server isn't Windows Client, but neither have a product name solely Windows, and both are generically Windows??

    1. Anonymous Anti-ANC South African Coward Bronze badge

      Re: Why isn't Windows Server Windows?

      Windows Server usually have a couple of extra goodies and tweaks in order to improve serving, and Windows Wanksta... errr... Workstation have a couple of goodies and tweaks in order to give a better user experience.

      For instance, there was a Registry tweak/hack for Windows NT4 Workstation in order for it to be a Windows NT4 Server IIRC, but it was a long time ago.

    2. Jou (Mxyzptlk) Silver badge

      Re: Why isn't Windows Server Windows?

      Windows Server is, upon base install, a cleaner version of Windows without bloat, which needs less RAM, and nearly no telemetric. It currently still allows uninstalling features like the AV, if you are that low on RAM, something which gets more and more difficult on the client OS. On the other Hand: Bitlocker and Wifi have to be installed as feature and the Windows Audio service is normally set to disabled, just to name the most prominent things you have to adjust.

      When it comes to a shootout in gaming performance the client version often wins since it's kernel, since Windows 10, is updated more often. So if you make the comparison mid 2018 it is like you'd compare Windows 10 1607 versus Windows 10 1803 with the newest CPUs and see one or two percent difference in speed. Same goes for testing right before Server 2022 was released.

      On the "I want that feature on a client" side is deduplication, which gave be extra 300 GB on my 1 TB "only games installed here" SSD without noticeable slowdown. Back then when I used it with Windows 8.0 and 8.1 (with stolen roles from Server 2012 (r2)). For using hyper-v the saving rate was above 50%, simply 'cause all those test-OSes have so much common data.

      Edit: There may be other subtle differences, like a higher security for process and VM separation, which do have their performance impact as well. But the is more guessing and "I read it somewhere" than knowing.

      1. jake Silver badge

        Re: Why isn't Windows Server Windows?

        "a cleaner version of Windows without bloat"

        That word "without", I don't think it means what you think it means.

        Try instead "with a little less bloat".

        And "cleaner" is very, very subjective ...

  20. Alan Brown Silver badge

    I'm less surprised about Arnie not going back to the site, than I am about the senior manager not getting his marching orders

    My experience with NHS (and other state health systems) is that they take unauthorised changes to critical equipment extremely seriously - particularly the installation of unauthorised servers and alterations of sitewide configurations

    Even back in the 1990s the privacy and liability ramifications of this kind of issue would have accountants and lawyers blanching

  21. Rtbcomp

    The Son of ...

    This phrase sends shivers down my spine.

    I once fixed a system where the MD's son had tried to get it working and presented the company for a large bill. The MD tried to fob me off with 30 quid in fivers instead of the several hundred he got billed for.

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Other stories you might like