Slashdot Mirror
Computer Voodoo?
jbeaupre asks: "A corollary to 'Any sufficiently advanced technology is indistinguishable from magic' is that sometimes users have to resort to what I call 'computer voodoo.' You don't know why it works, you barely care how it works, but you find yourself doing the strangest things because it just seems to work. I'm talking about things like: smacking a PC every 5 seconds for an hour to keep it from stalling on a hard drive reformat (with nary a problem after the reformat); or figuring out the only way to get a PC partially fried by lightning to recognize an ethernet card, after booting into Windows, is to start the computer by yanking the card out and shoving it back in (thereby starting the boot processes). What wacky stuff have you done that makes no obvious sense, but just works?"
Not sure how it works, but I've saved 2 or 3 hard drives that reported tons of bad sectors with cat /dev/urandom > /dev/hdb and then cat /dev/zero > /dev/hdb and repeating that a couple times. Seems to alleviate all the problems. The drives wouldn't format, and all the data would get corrupted, but after doing that trick, they haven't had a problem (with the longest running drive being 2 years after the fix and still going).
Anthropic principle: We see the universe the way it is because if it were different we would not be here to see it.
I once had a hard drive that wouldn't spin up if the computer had been off a few days. The only way was turning it by 90 degrees every time before booting the computer.
In my repair monkey days, my shop used to handle data recovery jobs of all kinds. The problems ranged from minor filesystem corruption or unbootable drives to physical damage - heads, and even a bullet through a hard drive (No, I wasn't able to get anything off that one).
We had a variety of methods for dealing with the physically damaged drives that had suffered a head crash, but my boss had a technique he called the 'massage'. A clicking or noisy drive would be rotated around its various axes until the BIOS would recognize it on boot. Sometimes the clicking would stop and he would sit there holding the drive in that position or prop it up to keep it there.
Another method we used was to freeze the drives for a period of 15 minutes to 6 or 8 hours. Sometimes this allowed enough contraction to let the tracks line up again, and we'd get as much data as we could with the drive cold. Once, we even froze a drive between two ziploc bags of water with IDE and power cables hanging out the edge to keep the drive colder longer. It worked!
-- Shade
Technology tips and tricks.
Whenever I boot from Windows to Linux, I have to turn the power strip off for seven seconds for the network card to work.
Please, for the good of Humanity, vote Obama.
My old TV card. No matter what cables you used, if the aerial lead wasn't bent at a 90 degree angle about 2 inches away from the computer, it wouldn't pick up a signal. In the end I just blutacked it down; I assume there was a loose connection inside, and the twist put out just enough force to make the connection.
However, after I installed the card, Windows 2000 would crash with the following BSOD:
Annoying as heck-- somewhat expected from a cheap network card.
So one day I was wat home downloading Fedora with bittorrent--- my DSL connection was maxxed out. There was too much interference on the line, so I hit the little 'channel' button to switch to a different channel.
As soon as I hit the button on the phone -- *boom*, the computer threw up the Blue Screen of Death. ANd sure enough, I reboot, hit the button on the phone-- and *boom* -- Computer crashes again.
I have since replaced all of the D-Link cards with cards from other manufacturers.
"Can of worms? The can is open... the worms are everywhere."
In general, with my main desktop machine and main laptop, if something funky is happening I will simply replace the part or parts in question to ensure a smoothly working machine, but I've had some interesting things with some old hardware I kept around for no real reason.
I used to have an old pentium (133 I think) that ran well, except that the CD drive would only actually recognize a disk if you tilted the computer at about a 20 to 30 degree angle when the disk was inserted. I never did figure out why this fixed it, luckily I didn't need to use the cd drive very often.
I also used to have a cable modem that would drop the connection if you so much as blew on the power cord. I always just figgured that was just some flaky hardware, and eventually got the cable company to replace it. Another really aggrevating hardware problem that I never figgured out was an old Sony DVD drive that I had. When you opened the tray, it would about 1 to 2 seconds later automatically close the tray, but when you opened it again it would stay open for about 10 seconds, just long enough to remove or insert a disk.
I think everyone runs into a situation where there is some voodoo involved in solving a problem, it becomes problematic when people stop carying about having any answers, and just care about getting something working.
Famous Last Words: "hmm...wikipedia says it's edible"
Actually, some people laugh at that, but I have the same thing happen. I remember a couple of years ago, I convinced a friend of mine to try switching from Mandrake to Suse on his laptop. Chatting with him on the phone, he complained that booting up the system was taking 15+ minutes. I drove over there to see if I could possibly diagnose the problem (he had been using linux for a while, but was never really much for sysadmining). I walk over, he boots up the machine, it boots up very quickly and runs flawlessly. Tried a couple of more times, same quick bootup. After I went home, he tried rebooting and ran into the same problem. We were both dumbfounded for quite a while, until I finally worked out that it was because when he had been using it, he was sitting in his living room, and it had hung waiting for eth0 to time out, but when I came over to look at it he put it in the docking station and plugged in the ethernet cable.
I've seen other situations like this. Many times, it's because the user is doing something they know is stupid/they shouldn't be doing, and with a techie looking over their shoulder they don't do it.
Famous Last Words: "hmm...wikipedia says it's edible"
Got to love old school hacking
I have mod points and I am not afraid to use them
...you gotta type 'sync' three times before it works.
Doesn't it make you feel good to know that our freedoms are protected by politicans, lawyers and journalists.
True, and throwing out equipment fried by lightning doesn't hurt, either. Seriously, what kind of questions are these? Hardware doesn't work? Isolate it and replace it. Other than freezing an old hard drive to free up bearings to get data off before you throw it out, it's not worth the aggravation.
Of course, as far as real Computer/Equipment Voodoo is concerned, there is always the Heisenbug. Just had a mechanical version of this today, the Bayer tech has spent 3 days on a machine to isolate a pump problem. To see the pumps, you have to open a panel either on the side or the front of the instrument. The past 2 days, he was working on it through the front, and the problem didn't occur. Today, after being called back because it happened again, he opened the side panel to watch it, and accidentally bumped the front panel while he was looking at it. As soon as the front panel closed, the problem occurred. It turns out that a zip-tie that holds some tubing from the pumps together was caught on the front panel, and when the panel door closed, it pulled on the zip-tie, which pulled on and pinched the tubing, causing a pressure sensor to throw a fault.
--That's the point of being root, you can do anything you want, even if it's stupid.
Just today I turned my computer on after leaving it on hibernate for a week. The "thaw" as I guess it could be called, failed (the computer hung before showing anything useful) so I rebooted. Windows starts up fine and then tells me my hardware has changed and I need to reactivate Windows. Except my hardware hadn't changed since the last boot (over the course of owning this computer, admittedly it had changed a lot). Oh wait, I can't activate over the Internet anymore, I've installed it too many times on the same machine, I have to call Microsoft, speak a 42 digit number slowly into the phone, get put on hold, be told I spoke the number wrong, put on hold again, read part of the number to a person, and then type another 42 digit number read to me over the phone. Then my computer will work again.
This reminds me of one of my favorite quotes/jokes from the Tom Baker era of Dr Who.
He stumbled across an old spacecraft on a very distant planet. As he sat down at the control console he remarked, "this looks like Earth technology". As he began to power it up it slowly came to life then started to fade back out. He kicked the bottom of the console and the rocket resumed slowly coming back to life. The Doctor remarked "Definitely Earth technology".
I just LOVE the implication that this sort of "kick it to keep it working" is a characteristic aspect of our technology that (in the world of this SF TV show at least) sets us apart from other species.
I used to have a Performa 5200 back when I started college, and if you're not familiar with the machine, it's arguably the worst Macintosh ever made. Ever. The only thing it excelled at was displaying grainy TV on the TV tuner card you could get for it.
Read that second link for all the gory details of why the follow scenario works, and you'll shudder.
I used to note in college that when doing particularly fast FTP transfers that saturated by 10-Base-T card that the machine would often lock up within a minute of starting the transfer. For months, I fiddled around and noticed that if I was actively working that this didn't happen. Eventually, I found the article I mentioned and realized that if I kept moving the mouse constantly, the machine wouldn't get in whatever weird state locked up the machine and I could finish my transfers. That's right -- to run FTP (or any other sustained, saturated transfer), I had to sit there moving the mouse in circles through the entire transfer.
Essentially, the "Left 32" bus described in the article was shared by the 16-bit Apple Desktop Bus (for mouse and keyboard) and the 16-bit networking card (as well as audio and the 8-bit SCSI controller). So long as I kept interrupting the bus with input from ADB, the networking card was unable to flood the controller that had to make sense of all the different bit-widths and clock speeds between the various busses hanging off of it, and the machine wouldn't lock up.
Now how's that for some serious computer voodoo?
If it's for-profit but free, you're not the customer -- you're the product (e.g., the Slashdot Beta's "audience").
Nostalgia - blowing on the Nintendo game cartridge.
Why in God's holy name does Windows fail to boot one time, and then boot successfully the second time?
Hah. I had something close to that one. A friend's Windows XP Home system. Boot it up, runs fine for about 30 seconds, then locks up hard. Reboot it, works fine for as long as you like. Next time you boot it up, locks up after about 30 seconds. Reboot, works fine. Lather, rinse, repeat.
I booted it up off a Knoppix disc and ran a bunch of hardware tests -- nothing. And no problems with locking up either.
Back to Windows -- same thing.
I eventually realized the pattern: after a clean Windows shutdown, it would lock up 30 seconds after the next boot. After a dirty shutdown (e.g. power cycle or reset button), it would boot up fine. Obviously the Windows shutdown was leaving something in a funky state for next time. Beats me what.
I told my friend she had the choice of doing a re-install and keeping fingers crossed, or always shutting it down with the power switch, or moving to Linux. I don't recall what she did beyond passing the box on to her kids because she'd already got a new one for herself.
-- Alastair
Well, the short answer is, because that was Apple's official service procedure.
The binary search method is theoretically the fastest way to identify the bad chip. This is a basic comp sci search algorithm, if I remember the it right, with 24 chips, you can identify the bad chip in 4 swaps. Replacing chips one at a time is, on average, the worst way to find one bad chip. Of course you could get lucky and it was the first chip you replace, row 1 socket 1. How often is that ever going to happen?
Also consider that the RAM diagnostic program could run for hours to run before a chip faulted, so swapping chips was not always the most time-consuming part of the test procedure. I sometimes used to speed up the process by gently heating up the motherboard with a hair dryer, on the theory that intermittent chips were more prone to failure when they were hot, but this was not usually effective.
Also consider how expensive RAM chips were back around 1980. We were NOT going to just replace every chip until we got the bad one, we only replaced known bad chips, if we pulled out chips, once they were ruled out as not the cause of the problem, they went back into the customer's machine. Otherwise we would have ended up with drawers full of suspect chips, and customers would have had to pay for every single chip we replaced. Labor was much cheaper than parts, back in those days.
Now I'll tell you what REALLY hurt.. when I had to service a machine that had TWO bad chips.. but fortunately, that rarely happened. They almost always went bad one at a time.
My dad works for the Arkansas Washington County Road Service, and he is something of a computer nut, so he would 'recover' the computers they were throwing out. I was pretty profoundly poor and also a computer nut, so I would take some of the stuff off his hands.
Anyway, I used to have three or four MFM hard drives in various states of disrepair. (I think they were 40 *meg* hard drives, but I only had a controller to control up to 20 meg, to give an idea how old this hardware was.) One by one they died, until finally only one was left. When it gave up the ghost, it would spin up, then immediately spin back down. I dug into it and found some connections I could short across while it was spinning up and then break the connection, and it would keep running. I was too poor to want to go spend $1 on a pushbutton, so I just had two wires hanging out of the front of the computer that I held together while booting the PC. I ran it that way for over a year...
A non-computer story, but more interesting one, is of an old Ford Escort I used to have. The starter went out on it, and, again, I was poor, so I dug into it. I finally figured out that the relay was kicking out too far and shorting out against the housing, so I duct taped a kitchen sponge to the inside of the relay housing and put it back together. I never had a problem with the starter again for the 2 years I had the car.
That same car later had the fuel pump go out. When it went out, I asked my stepdad if I should check to make sure the pump was out instead of a wiring or power problem, and he said nah, it's the pump. So I bought a replacement - it didn't help. So, I hunted around under the hood until I found some leads that were hot when the key was on, but not when it was off, and I used ties to secure an extension cord from the leads to the fuel pump. The car ran fine.
That was in the summer. When winter came along, one day I needed to defrost the front window as I was driving down the road. I flipped the vent from dash to defrost, and the engine stopped running. (I was doing 50 mph down the road at the time.) I flipped it back to vent, and the engine started right back up again.
Somehow I had found a wire that only gave power when the vent was not on defrost. I never fixed it, just kept the inside warm enough that it didn't frost over.
Now I'm a software developer and not poor. I virtually never fix (or jerry rig) anything myself, other than software and the occasional computer hardware issue.
I bought a clearance Super-Nintendo from CompUSA for $5 when I worked there. We didn't even sell them, but someone got suckered into taking it as a return. It worked find for about a week, then we ended up resorting to blowing on the cartridges for another week or two.
... :)
Finally it stopped working alltogether, but addicted to one of the games, I set to taking it apart and finding the problem. While it was apart, I found that if I held the game cartridge in with a certain amount of pressure it would work, but too much pressure or none at all and it would not operate at all.
Searching throughout the house for an appropriate weight, I ended up finding a 3 quarters empty bottle of Amaretto in the parental unit's liquor cabinet that worked perfectly. I spent the last semester of my senior year with a bottle of alcohol staring at me that I could never drink - for if I did my game console would die on me. It didn't last once summer started, though
The guy that owns my web hosting company told me a funny story along those lines once. He'd gone to the data center to upgrade the kernel on all his servers. They'd take one down, he updated the kernel, rebooted to make sure it worked, then handed it off to another guy to plug back in and the other guy would grab another one.
When they were on the third or fourth server, they realized the servers weren't booting up when they put them back in the rack. So they brought them back to the desk, booted them up and they worked fine. They'd put them back in the rack and nothing.
It took a couple of hours to figure out what the problem was. Somehow he'd managed to compile the kernel in a way that it wouldn't finish booting, if no keyboard was hooked up. He had no way to recompile the kernel there at the data center. He lives in Birmingham and the data center is in Atlanta, so going home to his own computer wasn't an option.
So he ended up buying a few dozen $5 keyboards at some computer shop and just laying them on top of the servers when they put them back in their racks. Worked like a charm.
Only on