... turn off a working PC.

Alas, I broke the rule of a lifetime when we went off to Norway, and I powered-down the IT infrastructure in the house. My PC came online OK yesterday (phew), but it developed funky signs of hardware failure after a soft reboot this morning (following regular security updates to Ubuntu). Curses, and thrice curses. Now I can order a full set of new PC-gubbins AND try to recover some data from a suspect hard-disk when I get back from Switzerland.

OF COURSE, I had put making a full backup of the disk on the long finger until after the holidays... CURSES.

The moral of the story - make backups, and if it's working, never, ever turn it off!

ETA: I'll add one more moral;- never, ever buy a mainboard not manufactured by Intel. The one which just went wonky was a non-Intel item, and I now regret the small saving I made upon its purchase. I've not yet managed to kill an Intel-made motherboard before its time...

CURSES.

It's those bloody leprechauns mate, they've been re-enacting the fight scene from 'The Quiet Man' all over your motherboard

I had the same thing happen but I was lucky that I could transfer my two 160GB hard drives onto the 500GB that I installed...it seems that the 3.1 gHz dual core processor and mother board w/3GB RAM that I bought won't take two disc drives, one floppy drive and two hard drives and the cost of the connector for the two drives was almost the cost of the bigger hard drive (which I partitioned) but they didn't tell me that they would charge me the same amount as the connector again to transfer the data, bastards!

ive never been able to kill a gigabyte or a MSI or an ASUS board either...

frankly i think the key with hardware is to never ever buy cheap, and never buy from a no-name brand...

My suspect board is an Asus... I've just ordered a true Intel one to replace it$ plus CPU, Kingston memory, Zalman cooler, Antec TruePower power supply, a new 1.5TB hard-disk and a Canon scanner (I've been scannerless for years, but never got that particular tuit, till now). Right on the heels of our holiday expenses - I can hear my wallet screaming....

So does Intel make AMD boards?

On a serious note, in IT the most common cause of unplanned downtime is planned downtime.

Is this high-end comps or regular comps? I turn off my comp almost every night to save on the electric bill, and also restart the programs(a lot ask for a restart when adding on).

I've not seen any errors on my end with a regular one and I've been through 3 comps.

he is most likely talking about turning it off without using the shutdown command.

Sorry for resurrecting an old thread, but... No. He was talking about shutting down his PC which he apparently had previously kept running for many, many moons.

Often equipment which runs fine for eons will fail when powered down, and refuse to come back up.

I'm sure Mike and Chris (among others) will chime in if they see this, but take it from a guy with 2 decades in IT: One of the most traumatic events in a geek's life is having to shut down a big data-center, because it's inevitable that some antique server(s) - without which the enterprise cannot run - will refuse to come back up after an outage.


Naturally this will also be the one which - in spite of years of logs showing successful backups - has no restorable backup available, or perhaps it will be some other equally-maddening fiasco, but... I digress...

Usually - as in this case - it will be disk-drives which fail, usually because they sieze up when they stop spinning though they may have run for another decade if kept running...

We geeks REALLY don't like powering off servers if they've been running for a long time.

DD

meh, ive never been one to subscribe to the notion that you shouldnt ever shut off your PC... servers is a whole different discussion... but i shut my machine off all the time...

Doesn't have to be antiques, but it helps. Power-cycling a machine is one of the more traumatic things you can do to it. You can think of it as culling the herd- anything 'iffy' will let go and need to be replaced when you power-cycle. Only the strong come back up without a problem. Our quarterly patching tends to work that way.

With the exception of a fan that went dead, every computer, printer, or monitor I've had to replace over the past 5-10 years died when I powered them back up after shutting them down over a vacation period when we weren't at home.

Indeed, powering off was what I meant. All my hardware failures that I can recall (other than equipment that was DOA) have as a result of restarting a machine that was turned off. As a result, I even set my hard-disks to never power-down, I just leave everything running happily, and well enough alone...

Seems to strike more often when you power down, and leave equipment powered down for protracted amounts of time. Shutting down for a night or two is no biggy, leaving things off for a month or more seems to invite trouble.

During the downtime, dust and debris has a chance to settle down into all the little nooks and crannies that it doesn't usually settle into due to the cooling airflow. Same could be true for moisture and condensation.

There is also the online updates, the longer a computer is down, the more of these stack up. when you power back up, you have to install them all. The more updates that are waiting, the greater chance that one of those updates will have issues.

The worst sin, is to physically disconnect your computer from power, (unplug it). When you do that, you're also disconnecting it from its grounding. This invites Electro static discharge issues. Turn off your computer, but leave it physically connected to power. There may also be 'passive draws' in place. the computer itself is shutdown but some items may still draw power to stayed warmed up. Your TV does this, (The little red 'power off' light) These features are typically used to reduce warm up time during power on sequence. Those items may also be particularly sensitive to initial power spikes, and thermal differences.

I'm not very good with modern day systems, but I am an Electronics Maintenance 7 level on a rack based ATE system that is physically older then I am. My Test equipment was designed in the late seventies, fielded in the early eighties, and is still going strong when we're not scavenging the dead for parts. Keep it grounded, keep the power physically connected. (In the modern era of electronics, "Turned off" is a relative term.)

A little "happy" update...

I had to stay at home today, as the heating repairman was overhauling our boiler's control electronics. That gave me some time to do a bit of "electronics" of my own - I took the components of the dead PC out of the chassis, and installed the new ones - Intel DG41TY motherboard, Intel Core2Duo processor, 8Gb of Kingston ram, an Antec TruePower power supply, and a 1.5Tb Seagate HD. Now it's happily running Ubuntu Linux 9.04, and I'm typing this on it. The only little "gotcha" is that the system won't connect to my network printserver, although its predecessor running Ubuntu 8 printed to it just fine...

Aargh!

I should take my own advice sometimes...

The other day, I powered off my computer and accessories, because someone had to sleep on the sofa-bed in my study, and I wanted to spare them the flashing lights and the fan noise overnight.

When I powered on again, the router was dead. Apparently its 240V AC to 5V/3A DC transformer was running hot, and when it cooled down on being powered off, some internal circuit failed, never to work again. Fortunately, I keep a spare (cheap and nasty) router in a box on the shelf, but I still had to replace the good one... Now I could buy a replacement transformer so I'll have a spare good router too, but the part costs more than half as much as a whole new router, so I probably won't bother.

Aargh. The next house-guests will just have to put up with the noise and flashy lights

Mine get turned off every night just before I turn off, and get turned on shortly after I wake up. It's for saving energy, I swear it.

Now which one will fail first, the computer or me.

I just had this happen to me.

Had some guests come over, who stayed in the guest room (natch). However the guest room is also the office where the computers live, so to be nice I turned them off. After they left, the desktop came back up but the fileserver had... issues. Intermittent crashes then failure to boot at all (lilo would run fine, kernel would start to load then lock up, completely) then nothing at all, not even fans spinning. Power-supply replacement surgery is never fun, the cabling ohsweetmercy the cabling.

Once the machine would power on again I needed some time with a rescue disk- I had a 'spare' kernel on the machine but no entry for it in lilo. Assemble the arrays, activate the volume groups, mount /, edit lilo.conf, run lilo and we're IN BUSINESS. Booted off the old kernel, replaced the newer messed up one and we're golden. BTW raid can protect you very nicely against drive failures, but what it doesn't do is save you from data corruption.

Anyway, next time we have guests I just may force them to deal with the noise.

My condolences. Glad you were able to recover from the failure.

I like your comment about RAID - I lost a lot of data recently on a RAID array which was being fed gobbledegook by a faulty memory module in the PC. Thanks to RAID, I had two perfect copies of the garbage. Aargh.

At least I was able to junk the bad module and turn the remainder of the machine into a simple Linux/Ubuntu PC for a colleague whose Windows box was dying a slow death from malware and the accumulated cruft of ages.

Thanks. I got lucky- I had backups but they were old and incomplete, so I would have lost some recent stuff (digicam pictures mainly) if everything- not just my kernel - had gotten corrupted.



Oh. I'm sorry to hear that, that could be very bad.

In my case, would you believe I had 4 copies of an unbootable kernel? I've lost data before due to drive failure (early in the Deathstar days) so RAID was a natural step. And I'd always heard it only protected you against drive failure, not data corruption. I was OK with that- drive failure had bitten me, but never data corruption. Until now. It's true. Bit of a wake up call.



Now that's cool. Funny thing though, computers have gotten so much better and so much cheaper (and I hold on to mine so long) that these days I have trouble giving away my old surplus machines. I have a ~1.4GHz Athlon/GF4/512MB/40GB (at least I think that's what's in there) machine sitting around that no one will take. It's weird.

Turned off my Toshiba laptop on Friday night, as I do every night. Saturday morning, it wouldn't boot up . Seems to have some corrupted system files. Currently in the process of transferring all the information I need to save to my old computer (huzzah for external hard drive enclosures, boo for USB 1.0), then I'll reformat at see if I can reinstall windows properly.

First time I've ever had this problem in 6 years of ownership, though. Reformatted a few times, but it never truly NEEDED it to work (even in the most rudimentary fashion) until now. Here's hoping.

Since one of the few things that I have which can't be replaced are pictures (even old fashioned photographs can be rescanned back into digital format), 'cause once digital photos are gone, that's it. When my last HDD crashed, wouldn't you know it, one of the few things I wasn't able to recover were the photos I had taken at the service for a parent from my childhood neighborhood. I just hadn't gotten around to copying them to another disk or tape drive yet when my drive decided to take a dump.

Now, whenever I copy my photos from my CF card, they go to THREE physically separate drives, one of which I can quickly disconnect and put in the gun safe in case of fire or evacuation (and I didn't have time or room to shove it in with the other stuff being afraid it might get damaged).

With the prices of today's external hard drives, using them instead of tape only makes sense - besides, they're MUCH faster than tape drives.