Have Fujitsu Harddrives Been Failing in Record Numbers?

Michael_Angel asks: "If your hard drive has started to show garbled characters in the BIOS at boot, or just does not pick up. You may be victim to what could be the biggest hard drive manufacturer failure rate yet! Our company is small OEM system builder and we have been hit by a failure rate of %90 of the hard drives we purchased a year ago. We might be lucky because we stopped buying after rumors of hard drive issues 3 months after Fujitsu Limited made some major changes. IBM had a pretty crazy rate of failure and was telling people to turn off smart mode. I've called Fujitsu and they said that there is no problem! However, a simple search for bad fujitsu hard drives on any search engine will point to some angry folks. One notable link is this Register story." Has this problem followed Fujitsu drives into other countries, or might they be limited to the UK markets? Have you noticed an unusual failure rate in Fujitsu drives compared to hard drives from other manufacturers?

Comments

Trends

[swordboy]

I've noticed one thing -

As drives have gotten smaller/increased data density, they've become increasingly unreliable. I'm pretty sure this coincides with the new 1 year warranties (versus the older 3 year standard warranties).

Laptop drives especially...

Re:Trends

[runenfool]

Its probably related to the increasing drive density - but it may also come from companies cutting costs to move more units to stay alive in the soft economy.

Just looking down the list of comments, it does seem that everyone has noticed the increasing number (or at least it seems that way) of massive drive failures from certain manufacturers.

Re:Trends

[Anonymous Coward]

I can only comment form personnel experience on this one, but I don't think that this is the case.
Drive companies for the most part produce very reliable hardware as long as that is their primary concern. Seagate, WD, Maxtor all produce drives as their main product. IBM has been doing it since the beginning. These companies don't try to touch all of the market, IBM excluded; they only try to be the best in a niche. That is where the good get separated form the mediocre. Fujitsu has never been one to shine in the HDD market, only just make par.

Re:Trends

[Anonymous Coward]

I work for a major OEM, and we have had nightmarish issues with Fujitsu (also Maxtor and Western Digital but mostly Fujitsu).

You need to watch out for clicking noises, "Drive Not Ready" messages, and 172x errors during POST.

From the official Fujitsu response regarding their hard drives, it is all how you categorize the failure. If we send back 2000 drives that have all genuinely FAILED, they will look at all the drives and will count those with scratched labels as "damaged during shipping" not a product failure. This helps them skew their numbers so they meet the 3-5% acceptable failure rate. It's a creative accounting method right out of Enron.

Not coincidentally, Fujitsu has stopped manufacturing hard drives.

Re:Trends

[jhunsake]

correlation doesn't imply causation

If only that could be pounded into most people's heads!

Re:Trends

[squarefish]

absolutely, I've replaced four drives in the last few months for myslef and others, two 7200 rpm maxtors, a laptop Fujitsu and a standard Fujitsu. All have still been under warranty and barely more than a year old. It's interesting that a lot of these companies have now gone to a one year warranty as opposed to the old 3 year which was the industry standard for years. My next motherboard will have intergrated raid.

Re:Trends

[buckeyeguy]

More info on the Zip drives can be found here. There are apparently known problems with the Zips, hope this helps (check bottom of linked page for the info).

Re:Trends

[Fweeky]

Although increased density perhaps makes drives a little more prone to failure, I think a big factor is how people treat their drives and where they buy them from - something that's becoming increasingly varied.

If you buy cheapo "OEM" drives from some box stacker, chances are it'll be poorly packed, and/or handled badly before it reaches you. Manufacturers can't do much if the box shifters keep throwing boxes of drives about. Just because they're rated up to 300G+ doesn't mean you don't want to handle them like eggs.

Heat's another factor; modern drives run damn hot - you really want a fair bit of airflow around them, either from your normal case intake fans and convection, or dedicated active cooling. Just because it runs fine doesn't mean you're not cutting it's lifetime in half, or worse.

The warranty situation I think is more down to the price war that's occuring with low end drives rather than any real change in quality. You can still get higher end drives with full warranties, and in some cases purchase extended warranties for another $20 or so. The 97% of users who don't experience a drive failure are probably happy to keep their $20, while 90% of the remaining 3% will likely get a replacement from their retailer anyway. The rest of us can spend the extra on a quality drive :)

Re:Trends

[rw2]

then why go from a 3-year to a 1-year warranty?

Because two extra years costs the companies money (even if the drives are quite reliable it still costs something) and the market has pressured price down more strongly than warranty up.

Re:Trends

[Fjord]

A long time ago, a friend of mine said to get a CD-RW instead of a zip drive and use it to do archiving. I'm glad I followed that advice since I currently have 3 working CD-RWs (including the original one from 5 years ago, a blazing 6/4/4, the two others came with each computer I bought since then). While CD-RWs aren't perfect, the medium is a lot cheaper (about a dollar a disc, millidallars per meg), and the software has been getting better at using the discs efficiently. They archive well, with tons of 3rd party products like jewel cases and racks for storage (they are CDs after all). You always have the option to use a format that all computers can read (CD-R), and with external USB burners they are as convinent (and cheap, both are $130 at best buy) as Zip Drives.

Re:Trends

[Killall -9 Bash]

and you thought HDD fans were just for those crazy OCing kids, didn't you? I think a big part of why HDDs are failing is improper case ventalation (due to cheap-ass manufacturers not wanting to spend the extra 3 cents on a case fan, AND smaller and smaller cases beccoming more popular) combined with drives running hotter than previous ones. A few days after installing a Maxtor 60gig, i was inside the box cleaning up the spaghetti and happened to touch the top of the Maxtor.... and was scared shitless to find that it was *HOT*. 10 minutes later $7 bought me a HDD fan/drive bay mount combo. nice and cool now, and i'd bet dollars to doughnuts my Maxtor 60gig will outlast any Maxtor 60gig without a fan.

Re:Trends

[Steve Franklin]

"...give it room to breathe!"

4 fans. Aluminum case. 6 GB WD IDE 4+ years old. 18 GB Seagate SCSI 2+ years old. So far, no problems. I would imagine cooling is one of the big factors, as you suggest.

Re:Trends

[tomhudson]

I had one of my CDs explode while doing a linux install (I thought the dog had knocked over the box, it made so much noise), but I would tend to agree that my writers are more dependable. It's just that the IDE zip drives are FAST!.

Time to buy a few 220-meg CD/RWs (I like to carry my backups in my shirt pocket).

Re:Trends

[WowTIP]

Ouch! I only have 3 fans in my box (CPU, GFX & Power), but still I think the sound will drive me nuts sometimes. How do you cope? Earplugs? :)

Someone should really invent a better method of cooling computer devices and make a fortune in the process.

Re:Trends

[rw2]

If the drives are so reliable (MTBF of a million years my ass!), then the extra cost would be $0.00 per unit.

1) I hope your joking, but to be pedantic the MTBF is only 500K hours, not years, of course.

2) The extra cost, even if the drive itself is fine, still exists as you vet problem reports from your customers. In fact, the cost of vetting the problems is probably about the same as the drive itself, so even if they make perfect drives that never fail their warranty support costs would still be half of what they are if the make drives where not a single one lasts for the entire three years.

Now, that said, re-read my note. Nowhere do I say that drive quality is as good as it used to be. I only point out that there are valid reasons, market based reasons, to reduce the warranty that have nothing to do with the quality of the drive.

Re:Trends

[rw2]

so even if they make perfect drives that never fail their warranty support costs would still be half of what they are if the make drives where not a single one lasts for the entire three years.

Heh.

This is incorrect. Teaches me to type and run.

The point is, there are significant costs to a longer warranty even if the drives are high quality. Those costs are passed on to the consumer and the consumer has decided that he would rather have a cheaper drive than a longer warranty. I'll not belabor that point further.

Re:Trends

[tomhudson]

Drive manufacturers require you put up the (supposed) retail cost of the drive before they issue an RMA. Why do I say supposed? Because it's more than twice what I would pay for the same drive from the shop down the road. And way over their current MSRP.

Then they ship you the replacement.

You return the crapped-out drive.

Then, if they determine that the drive is buggered outside the warranty, you've bought yourself a "new" (because they reserve the right to ship you a used drive as a replacement) drive at twice the cost that you paid for the damn thing. And you're paying the shipping.

It's because of tactics like this that warranty repairs become a revenue generator, as well as discouraging people from making claims in the first place.

In my experience, real-life MTBF is only in the thousands of hours.

Re:Trends

[meldroc]

It's definitely related to cost cutting. The manufacturers are perfectly capable of keeping the drives reliable at higher densities - that's just a matter of using the amount of error correction, sector sparing, interleaving, scratch detection & such to bring the error rate down to a statistically insignificant number.

The hard drive manufacturers are under intense pressure to cut costs. If they can reduce the price per unit by five cents, when that is spread out over hundreds of millions of drives, that adds up to a lot of money. Especially in this economy, this means you'll see drives made with cheaper components, with less testing done, in clean rooms that may not be as clean as they used to be, by workers that don't have the training their predecessors had, using firmware that has been hacked and rehacked until "spaghetti" doesn't even begin to describe it. (Don't ask me how I know this...)

But this is no excuse for a 90% failure rate. Making drives cheaply is one thing, but we as customers still expect them to work for at least two or three years without problems. I still expect vendors to own up to their screwups and make them right.

Re:Trends

[meldroc]

Just because they're rated up to 300G+ doesn't mean you don't want to handle them like eggs.

Let me give you a bit of perspective on what a "300G shock rating" really means. If you drop a can of pop on your counter from three inches, that will induce a shock on the can greater than 1000Gs. When I worked in the HD industry, I learned that simply tapping a drive with a pencil induced a momentary shock of 40-50Gs. I could fire up some diagnositic firmware on the drive, and watch the drive detect and fix errors as I tapped it with a pencil.

Moral of the story, hard drives are fragile The only reason why they seem so tough is because the firmware detects and fixes thousands of errors that you don't even see.

Re:Trends

[0x0d0a]

Fortunately, there's a fix. You can buy server-class drives SCSI drives if you're uncomfortable with cheapo ones. Sure, they're going to cost you more, but that may simply be the cost of reliability.

It really comes down to how much you're willing to pay for peace of mind about your data. A monitor failing is no big deal. A hard drive failing can cost you years of work, source code, everything.

Unfortunately, "backing up" is no longer a really good option for most people. Perhaps buying a second drive and mirroring, but tapes (except for the very most expensive) and CD-Rs are simply too small compared to drive size to be very useful for backup. It's actually cheaper to use a second drive to back things up these days (compared to tape). Writeable DVDs still are expensive, still aren't popular or standardized, and even when they get so, are very fragile and likely to require at least ten discs to back up a complete hard drive.

If anyone knows of a less expensive, large-amount-of-data-per-unit backup system, I'd be interested to hear about it.

Hard drives got too big too fast. They outstripped CPUs and Moore's law. They outstripped all competing storage devices. They actually outstripped consumer demand in the last two years or so. The people doing the research on them are *too* good. I remember buying an 80MB drive not-so-many years ago. Slow. Physically huge. Cost something like $3 per megabyte. Now drives have a price/performance ratio 6000 times better. No other product in any field I know of has come anywhere close to this.

Re:Trends

[0x0d0a]

few days after installing a Maxtor 60gig, i was inside the box cleaning up the spaghetti and happened to touch the top of the Maxtor.... and was scared shitless to find that it was *HOT*. 10 minutes later $7 bought me a HDD fan/drive bay mount combo. nice and cool now, and i'd bet dollars to doughnuts my Maxtor 60gig will outlast any Maxtor 60gig without a fan.

Unfortunately, this is an artifact of people trying to jack speed on cheap drives. It's a really, really stupid idea. If you want a reliable, quiet drive, buy a *5400 RPM* drive. I can't stress this enough. They're getting unpleasantly difficult to find, but a few people still make 'em last I looked. 7200 RPM drives run *much* hotter, though FDB may eliminate that. Have to take another look...

Re:Trends

[twilightzero]

I don't know what HD company you're referring to in this, but it's obviously not the one I work for. And I just happen to work for WD, in tech support/RMA's, so I think I'm qualified to write a rebuttal ;)

Drive manufacturers require you put up the (supposed) retail cost of the drive before they issue an RMA. Why do I say supposed? Because it's more than twice what I would pay for the same drive from the shop down the road. And way over their current MSRP.

When setting up an advance replacement (where we send the HD out to you first), we do require a credit card # but we do NOT charge it. We do an authorize operation, which checks that the card is valid and that it has at least enough available on it at that exact moment to cover the drive should you not return it. And we do not charge double fair market value for the drive, in fact we get a lot of comments that our authorize amounts are VERY low - about $25 for a 20 gb drive and $31 for a 40 gb (both 7200 rpm) to give you a quote off the top of my head.

Then they ship you the replacement.
You return the crapped-out drive.

Standard practice.

Then, if they determine that the drive is buggered outside the warranty, you've bought yourself a "new" (because they reserve the right to ship you a used drive as a replacement) drive at twice the cost that you paid for the damn thing. And you're paying the shipping.

We do indeed test the bad drives when we get in, however we don't wait for the testing to get done before sending out the replacement, mostly in the interest of time. And if your old drive tests out with NDF (no defect found), big deal, you won't hear about it or get charged for it. And if you've managed to damage your drive outside the warranty, I'm sorry I have no pity for you. Almost the only way it's damaged outside of warranty is if it's physically damaged, and if that's the case you're the stupid one for even thinking you'd get it replaced. On the off chance this would be the case, the drive would be rejected back to you and you would have the option of returning the replacement or purchasing it, see my note above about fair market value. Yes you would pay shipping back of the replacement drive back, as I said physical damage is one of the only ways to get it rejected back and if you're dumb enough to send it in like that, you deserve what you get.

As far as sending remanufactured drives as replacements, yup we do, as well as I believe every other IDE manufacturer (scsi I'm not as familiar with so I don't want to conjecture). And repeat replacements from failed refurb drives are actually more rare than you'd think. Sorry I don't have any numbers, just personal experience. And sorry for the long-windedness but either you're REALLY dealing with the wrong company or you're stretching the truth like Gumby on a taffy machine.

Re:Trends

[kbielefe]

You are assuming that people need to back up their entire drive. I use only 2.6 gigs of my 20 GB disk at work and consider my installation disks to backup 2.5 of that. You are also assuming that the average user actually fills up their hard drive with crucial stuff. Most people that really need backups don't need a backup of their entire hard drive, but they only need a backup of a specific project. And projects fit pretty well onto a CD or even a floppy disk, especially personal projects. I'd rather backup my budget separately from my ogg collection separately from my personal programming projects anyway.

I agree with you about the ease and cost of using another hard drive for backup, though, as long as they are physically separated (at least not in the same machine, and the farther the better). I frequently copy files from my laptop to my desktop and vice versa. I also consider the local copy of my website to be a backup of the remote copy, and vice versa. I think most businesses nowadays use hard disks for most backups and only use tapes for old archived stuff.

Re:Trends

[drinkypoo]

On the other hand, an old school seagate barracuda (the first generation, which was 3.5" half-height, rather than the 1" or even lower drives which are commonly sold now) drive which is loud as hell is too hot to pick up after it's been running for a while, even with a 5" fan blowing across it, and most (but not all) of those ended up more reliable than the average first-generation hard disk.

Meanwhile, let us not speak of the slow and sloppy Quantum Bigfoot. Yecch. I am only going to use it as an example.

Now remember, the original barracudas, unlike the ones sold now (or is that just recently?) were performance hard drives, supporting such cutting-edge features as Ultra SCSI. :) I still have a 4gb, I trot it out whenever I need a narrow scsi disk. I also have a nice 9gb IBM LVD disk but I haven't used it enough to really comment on it. Anyway the Bigfoot (The only kind of Quantum drive I've ever had problems with, which further illustrates my point) is cheaply made in ways that matter. They were really high capacity but very slow (slower is cheaper) and very sloppy, IE prone to failure. You know, like old (ST-506 interface) MFM disks from Seizegate. Boy, was that a dark time for those old boys.

Keeping your disk cool surely helps, but older drives would last and last without any special cooling considerations, and many of them ran quite hot. Sure they were slow as hell, but manufacturing technology has come a long way since then. You would think they would become more reliable, not less. Unfortunately it costs more to do business the right way, which is to always build everything as well as you can, than to do it the prevalent way at this moment, which is to offer just the right length warranty to where you can only have to replace so many. This indicates that they know a whole bunch of them will fail, and they know that a certain significant percentage of them are going to fail. This is no different from the way business has ever been done, but the percentage they find acceptable seems to be slipping in the face of the economy.

In short, use an IDE RAID, always buy twice as much disk as you need, and mirror it. If you need more performance ignore what I said about IDE but don't forget the RAID.

Re:Trends

[Proc6]

Might it also be related to drive "usage"? Think about it. When there was Windows 3.1, an operating system that fit in 8 megs of RAM, and the average person probably opened 10, 50k Word docs a day. What about now? Games that take 2 gigs of drive space, and probably load half that during play? Playback of 600 meg DiVX movies, web browsing that writes, deletes files absolutely constantly. Streaming hundreds of megs of MP3's off the drive while you work on your 128 meg OS, editing 3gigs of digital video?

Seems to me they also work harder. Not saying the manufacturers aren't at fault but we may be overlooking just how vastly different today's harddrive usage is.

Re:Trends

[Tycho]

There is another fun thing about first generation half height Seagate Barracuda drives: the voice coil magnets. The magnets in those half height 'cudas are easily the strongest permanent magnets I have ever seen in a hard drive or not originally from a hard drive. Don't worry the 'cuda was already dead when I took it apart. I killed the 'cuda by inserting the power plug in upside down. The drive was already experiencing bad blocks, the idiot former owner put it in a computer with no air circulation.

Re:Trends

[uberdave]

Water cooling is a lot different than what your freezer does. With a freezer the working fluid (freon) goes into the freezer and is allowed to expand. As it expands, the heat of the surrounding environment (the food) causes it to boil. The gas is then forced out of the freezer compartment and passed into the condenser coils. A compressor squeezes the gas until it condenses back into a liquid. In the process it sheds heat to the environment (the air surrounding the freezer). The liquid then circulates.

A water cooler just moves water around. There is no phase change (liquid/gas - gas/liquid) to absorb heat. Actually the working of a water cooler is a lot closer to that of a fan than that of a freezer.

Re:Trends

[0x0d0a]

Oh, and on another note, I get a few people talking about performance issues in using 5400 RPM drives instead of 7200 RPM drives. That's silly. If you're using either of the two, you're probably using IDE drives, and thus probably building a workstation. The performance hit is just not important. A 5400 RPM drive is 25% slower than a 7200 RPM drive. If you're hitting the disk at *all*, you're going to be slow compared to working entirely in RAM. Increasing rotational speed by even a factor of two or so is not going to be an issue -- the amount of RAM you have to cache stuff on the disk is, in almost any real-world situation, a far more influential factor. The filesystem you're using, the fragmentation level, what OS you're using and how you have the caches set up...all these can have far more impact than a 25% change.

And when it really comes down to it, what do you care about more -- your data not being lost, or large file copies finishing 25% faster (under ideal circumstances: assuming the bottleneck is entirely from the hard drive and that head movement time is not an issue). Background the copy and do other work while it's going on.

Re:Trends

[tomhudson]

You can drop a drive off the desk and it will still work. The heads are parked when the drive's not powered up, and most drives have non-powered G-ratings in the hundreds.

As I pointed out in another post, RMAs of good (not defective) drives has become a profit and sales generator.

Besides, let's get real here. User incompetence in the installation shows up in the first few minutes, 'cause the box doesn't boot!. This doesn't make a jot of difference between 1 year and 3 years warranties.

Re:Trends

[Reziac]

The problem with a well-established system, one used for real work day in and day out, is that reinstalling and reconfinguring everything can take far longer than even the slowest backup. I once figured out how long it would take me in realtime to redo my old Win16 system from a blank HD, and it totaled at TWO WEEKS of work at 8 hours a day. Gods know how long it would take to get my Win95 box back 100% how I want it, if I had to start over from scratch. It's not just reinstalling the software, it's remembering and finding all the little tweaks and updates you did to make everything play nice together and behave just how you want it.

Yeah, for Aunt Minnie's email, on a box where the only stuff that ever has to work at all are Windows and Outlook, reinstalling the software is not such a big deal, and you really only need archive the user data. But not so at all for a complex system that does a variety of tasks (especially if those tasks are interconnected).

I don't have any good solutions either, given the size of current HDs and how much data a person can accumulate (both programs and documents -- BTW my sister routinely works with documents in excess of 4gigs). A RAID server with the sole task of keeping the network backed up seems the least impractical for large setups.

Re:Trends

[tomhudson]

Re: charges on credit card: up here in Canuckistan, Maxtor Canada did a charge of more than double what I would have paid for a new drive, and shipped us a replacement. about several hundred $$$ for a 40 gig, two months ago.

Then, when we shipped back the bad drive, they did a refund of the charge.

Of course, we ate the shipping.

Mind you, if you want to get rid of a whole bunch of remanufactured 40 gig drives at $31. a pop, I'm in!

Re:Trends

[twilightzero]

If you buy cheapo "OEM" drives from some box stacker, chances are it'll be poorly packed, and/or handled badly before it reaches you. Manufacturers can't do much if the box shifters keep throwing boxes of drives about. Just because they're rated up to 300G+ doesn't mean you don't want to handle them like eggs.

You mean the places like Tiger Direct (there are many others but that one's fresh in my mind) who ship you the HD in a bubble envelope? Then they say they don't take returns even though it states it in clear print right on their website.

You can still get higher end drives with full warranties, and in some cases purchase extended warranties for another $20 or so.

AFAIK, Western Digital is the only company still allowing you to purchase the extended warranty on 1 year warranty drives (Disclaimer: I work for them). Someone please correct me if I'm wrong.

Re:Trends

[Reziac]

I had the same thought last time I bought a large HD. Do I want a cooler 5400rpm drive that will likely last 50% longer, or do I want a 7200rpm drive that will run a good deal hotter and probably die sooner from heat stress? Easy choice, I'll take the slower, cooler HD. It's not like the speed difference is all that significant in a workstation (unless you're working with multigig files all day every day, then the time saved with the faster HD might add up to far more than the cost of replacing the HD sooner).

I use roomy, well-ventilated cases (or make them so as needed) and if a given HD runs hotter than merely comfortably warm to the touch, I give it its own fan. Found a flat fan unit that screws to the bottom of the HD; quiet and does the job very well, for a mere $8 investment.

Maybe I'm just paranoid, but could be this is why (other than a couple that were DOA) I've only had one HD fail in everyday use, and that because it likely got head-crashed when I moved. (Was fine before, had the creeping crud after. Still got 5 years use out of it.)

Re:Trends

[Reziac]

I've been buying W.D. HDs exclusively for a long time, for myself and my clients. At last count I had .. um .. [adds in latest salvage gleanings] 18 working W.D. HDs in the house. (The oldest is a 20mb model dated 1991, still 100% error-free.) Also have half a dozen more with age-related problems. Of course, I tend to treat 'em like raw eggs and am careful to avoid overheating too :)

I've had to RMA three new W.D. HDs for DOA, plus one that died on the LAST day of the then-3-year warranty. I will say W.D. made it plenty easy. Tech support dude determined that I had already tested the HD and knew what I was talking about, and two minutes later we were done.

But the warranty cut to one year was disturbing. I'd noticed that my most recent purchase (first I'd seen made in Malaysia rather than Singapore) is visibly not as well-made as previous W.D. HDs. And that was my most recent RMA as well, having developed data errors at only a couple months old. Next time I buy some HDs, I'll be looking for last-year's models.

As to "but you can buy the 3 year warranty for $20" -- sorry, that does me no good at all if I first had to pay *double* the OEM price for a retail-boxed HD (which my clients sure as hell are not going to go for either). And I can't use the enterprise class drives (which still have the 3 yr warranty) on most of my own and clients' machines -- that big a HD isn't supported by the system BIOS.

And yes, I've ranted about this to W.D. sales, public relations, and tech support. As a SOHO integrator, it puts me in a bad spot (you think YOU have trouble making money off HDs!) and makes me wonder if maybe I should just settle for the cheapest HD if they're only going to be backed for a year anyway. :(

Re:Trends

[twilightzero]

Which just underlines the point that MTBF isn't worth a load of jack anyway ;) Part of the problem is that it's IDEAL conditions, not everday people taking their computers to LAN parties and dropping coffee in them. See the discussion above about shock rating of the drives and how incredibly easy it is to exceed this. Also don't forget you're dealing now with MUCH smaller sector/head sizes than previously. I definitely wouldn't say this makes the drive less reliable, but it absolutely makes it MUCH more delicate.

Re:Trends

[kbielefe]

I agree with you that it would be most convenient to just backup the whole computer. However, the whole point is that the size of today's hard drives makes this no longer convenient.

This is exactly why my parents have Mandrake on their computer. They don't know how to set things up anyway (on Windows or Linux), so I do it for them. What to backup is easy: their home directories. It's just as easy for them from a user's perspective (if not easier) than windows and they don't have to spend money for software upgrades. Since all the software they use is on the distro disks, a reinstall is easy. It's a far cry from the 10 cd switches and 15 reboots to completely reinstall a windows machine. Just a little slower than a full restore operation and you get the latest distro version as a bonus. And they don't have to worry about their grandchildren accidentally erasing their work files.

Re:Trends

[twilightzero]

Yeah I've been hearing rumors that Maxtor has changed their "no hassle" RMA policy quite a bit recently. Haven't had the change to experience it yet, but what I've heard agrees with what you've said...*comfort*

That being said, did you use a debit/check card? I don't remember if I mentioned before (and I'm too damn lazy to look back) but when we do an authorize on a credit card that's a debit/check card, the actual bank that issued the card usually puts a hold on the funds for several days. Basically they expect a charge to come through after the authorize and since it's working directly with a checking account balance, they reserve the funds aside in anticipation of the charge. When the charge doesn't come like they expect, they drop the hold. Just something to think about :)

As far as reman 40 gb drives, if you wanted to directly purchase re-mfg drives they would end up being somewhat more than that - I don't have direct quotes available. Basically if we charge you for a non-returned RMA drive, you get the warranty voided and no tech support and a nice note in your case that you don't return drives ;) When you straight out buy a re-mfg drive, you get tech support, warranty of varying periods (you can specify this at time of purchase), etc. - basically full support for the drive.

Re:Trends

[rw2]

Quibble away.

He said a million years, not 57. Read the parent then retract your goofy quibble if you like...

Re:Trends

[tomhudson]

no, was a regular Mastercard account. Thanks for the info.

Re:Trends

[achurch]

I could fire up some diagnositic firmware on the drive, and watch the drive detect and fix errors as I tapped it with a pencil.

Out of curiosity, is there any way for end users to get at the error information (like what sectors have gone bad) on IDE drives? SCSI drives don't automatically remap bad sectors, so I can catch a failing drive before it dies hard, but when my 40GB IDE (coincidentally, a Fujitsu) kicked the bucket this past summer, it was a pretty major mess--it would have been nice to know ahead of time that it was starting to go bad.

Re:Trends

[peter]

> i'd bet dollars to doughnuts my Maxtor 60gig will outlast any Maxtor 60gig without a fan.

I'd take that bet, but donuts cost a dollar now, so whatever. (You did say any, instead of every, so it's not so unreasonable for you to propose such a bet. :)

Hard to imagine

[ekrout]

Every single one of my friends here at school who purchased an IBM DeskStar-line hard disk drive had the drive fail on them less than a year after purchase.

I never thought that dependability could be much worse than for that particular line of IBM HDDs. But, this Fujitsu story sounds like it's a dire situation as well.

As a side note, I'd highly recommend (and do so to family, friends, etc.) purchasing only Western Digital or Seagate drives.

Re:Hard to imagine

[Skyshadow]

I dunno, I've had trouble with Western Digital drives in the past, too. In fact, I remember one of their drive lines (although I can't remember which) being notorious for failing quickly.

Personally, I think this sort of discussion is useless just because there are people out there who have had trouble with any given manufacturer's drives.

I think a collection of real stats which were somehow reliably collected would be really useful in terms of all this commodity hardware ("Gee, those ShitCo drives fail twice as often as most others" or "Gee, there's no difference in drive reliability, so if I got IBM I'd be paying for a brand name"). I just don't see how you'd go about collecting that data.

Re:Hard to imagine

[escher]

And I have never had trouble with the IBM Deskstar drives in my file server (been running for 3 years now). Maybe I got lucky and was shipped drives from a good batch...

Re:Hard to imagine

[ekrout]

Here's just one example of an unhappy owner of a large IBM DeskStar drive:

I have bought 4 75GXP drives all of the 60 GB variety. Initially I bought two to connect to a RAID system but one failed after only 5 months. Just as well I had my RAID set to mirror otherwise all would have been lost. I have in excess of 25 GB of MP3 files which have taken years to collect hence my need for reliable storage. I then contacted the suppliers of my drives and asked about replacement. I was told it would take at least 6 weeks as they had to go back to IBM.

Given the importance of backup I bought a third drive whilst the 1st was being checked by IBM. Guess what 6 weeks later a second drive failed. BY this time I received back a drive from IBM. This was a second hand drive that had been returned by another customer under warranty. I know this because I was able to unerase the data on the drive and the former user was from Germany. This drive failed after only 4 weeks.

The second drive to fail was also replaced by a second hand drive. This also is making ominous noises.

In fairness when they work they are fast and very quiet but the uncertainty about when they will fail has left me very unimpressed. Of the 4 purchased 2 are new and working fine 1 is broken and I can't be bothered to send it back as I know they will send me another dodgy 2nd hand drive and the final one is noisy and I am sure would fail if it were used as a RAID drive.

My advice therefore is to look elsewhere. When I upgrade my system shortly I shall buy 2 120GB drives from another supplier but I shall research carefully first.

-- From http://www.dooyoo.co.uk/computers/hard_disk_drives /ibm_deskstar_75gxp/_review/393167/

Re:Hard to imagine

[cheezedawg]

Well, even the Lawyer that is representing the class action lawsuit against IBM (streaming video) is only claiming a 1% failure rate, and the only model in question is the 75 GXP.

The fact is IBM had some quality control problems with the 75 GXPs, but they have been fixed. And it wasn't too many years ago when people were saying the same things about Western Digital drives.

I think you always have to take this anecdotal evidence (i.e. "My friend had this drive and it sucked") with a grain of salt.

Re:Hard to imagine

[Zathrus]

I think a collection of real stats which were somehow reliably collected would be really useful in terms of all this commodity hardware

Storage Review had such a database at once upon a time. It was widely hailed as the most comprehensive database of the kind, and pretty accurate (given that "reliable collection" is an oxymoron when it comes to the net).

Then their hard drive crashed and they lost everything.

Yes, it's horribly ironic. It also struck me as really freaking idiotic that a website dedicated to storage wouldn't back up their own data. I'm not an SR regular, so I didn't follow the story that closely at the time.

As it happens, SR is now restarting the reliability database. It'll take time to get accurate data, of course, but it's better than nothing. Here's hoping they succeed.

And that this time, they have backups.

Re:Hard to imagine

[Anonymous Coward]

Western Digital has been one of the crappiest of disk manufacturers -
Maybe you don't remember their massive recall in 1999 - 2000 ?
I have a couple WDs that refuse to be slave drives on an ide channel EXCEPT with another Western Digital drive as master.

Maybe they've had to clean up their act since the big recall but I haven't been willing to be the guinea pig to settle that question. I've also seen some Maxtor drives crapping out within 12 months of purchase.

In general the reliability of product in the IDE drive market has sucked as margins have declined further and further with the tech slowdown. IBM leaving the market was very demoralizing to see; if anyone could have turned the trend around - started making drives with high QC and charging more and getting it, it would probably have been them. There aren't too many makes that have not experienced a quality crisis like theirs with the 75gxp deskstar product and stayed in. But they left I am sure, because they concluded that margins would never bounce back.
And we're all gonna keep suffeing for the shortsighted cheapness of the consumer.

Re:Hard to imagine

[DeltaSigma]

With such consistent failures in a single setup, have you thought of checking other components?

I'm not trying to patronize you, I'm merely saying, before you go spend the money on a couple 120 gig drives, you might want to research your RAID controller, motherboard, etc.

And if it's reliability and speed you're looking for, as apposed to size, perhaps abandon the future double 120 Gig drives and instead get a couple SCSI drives. I've been running a Western Digital (favored brand) and a Seagate (second favored brand) for three years, and neither show any signs of wanting to write themselves into oblivian any time soon. Compartively, the Maxtor I had quit on me in just outside of a year.

In any event, the speed of a few good SCSI drives has kept my old PII 450 capable of competing with much higher clock speed setups with IDE drives, even the 7200 RPM ones until recently.

While my next comp won't be SCSI (money) I wish it could be. I highly reccomend SCSI drives and with the continuing extended warranty and fast access, it might be right up your alley.

Re:Hard to imagine

[tchuladdiass]

What was the airflow like in your server case with those drives? Were the drives sitting in adjacent drive bays? Many brands of 7200 rpm drives run extremely hot, and need proper airflow over the drives, which some cases don't provide. One of my drives was running hot enough to almost cause blisters, but when I got a new case, it ran much cooler. Even if the case provides proper ventalation, you've gotta watch the drive cables, they tend to block the air and cause warm spots.

Re:Hard to imagine

[xenofalcon]

On the bright side, at least they had an entry to make immediately after they started rebuilding the database...

Re:Hard to imagine

[BdosError]

There is a class action lawsuit going on currently with regard to exactly those hard drives: see more info here.

Re:Hard to imagine

[nelsonal]

Mostly because they never made any money at it (The PDF is much easier to read). It just got worse last year when all the news about their 75GXP line having quality issues. They lost $423 million on the hard drive business in 2001, and almost $100 million in the first three months of this year. The hard drive business is in the middle, adjustment, column and the division's total loss can be found in the line net income from continuing operations. Because it is an adjustment column, all the numbers are of the opposite sign. Hitachi hopes that by focusing on harddrives they can return the division to profitability, or it can gain an advantage of EMC in storage arrays, I'd assume. IBM will probably continue to do some R&D to get licensing income, although I don't believe that the final transaction has taken place so the drive business might still be a part of the company.
They have also been focusing on services, their high end server business, storage, and semiconductors, and getting out of most of their other businesses. They fully removed themselves from PCs and the like, licensing their name and selling the products only to those who ask for it. If you wanted to be an IBM only shop or something similar.

Re:Hard to imagine

[slaker]

I *am* a Storagereview regular (I post there as Mercutio, the second non-admin user whose account was re-created after the crash).

The issue was *not* a disk crash, but the fact that SR's colocation facility wanted to charge $x more to run proper backups, and SR couldn't afford it. During a regular upgrade to either MySQL or phpBB (don't remember which), their DB got dumped on accident. Eugene, SR's admin, posted very early after the site came back up that he has a small stack of DDS and DAT drives sitting around his home that he would've loved to install, if only their ISP would've let them.

Incidently, Storage Review's self-reporting reliability database is back up and running now, if you'd like to participate, feel free, but I'm convinced that self-reported statistics are of fairly little value.

Also, a lot of SR's regulars, including myself, chose to create our own community, distinct from SR, in case Storage Review either shuts down or loses its database again. We can be found at Storage Forum. SR's general membership is not aware of our site - we don't advertise it there out of courtesy to SR's admins, but if you spent time on SR's forums and wonder where Tannin, Clocker, P5_133XL, JamesW, time and some of the other mainstays went, well, now you know.

Re:Hard to imagine

[Zathrus]

Ah, cool... I was actually hoping someone would pipe up and say what actually occurred. I did some (very minimal) searching prior to posting but couldn't quickly find the story.

The self-reported statistics are of fairly little value, but it's more than what we have now - stories like this one, vague rememberances of failures past (yes, I remember the Seagate Stickition problem, but I really doubt it has much to do with their drives nearly 20 years later), and utterly useless MTBF numbers. Sure, you can go to various forums and read people who say they are a dealer, or work for a wholesaler, or whatever and have huge returns on xyzzy's drives, but that's about even less reliable than the self-reported statistics.

I'll check out your website sometime... thanks for the post.

Re:Hard to imagine

[cat_jesus]

I have a deskstar 30 Gig as well. In fact I just made the mistake yesterday of turning S.M.A.R.T. back on after I upgraded my bios. All the sudden I started getting all sorts of strange errors and sudden reboots. Luckily I realized my error and turned it off again. Wierd.

Re:Hard to imagine

[ckedge]

> I have in excess of 25 GB of MP3 files which have taken years to collect hence my need for reliable storage.

Backing up such data *is* problematic.

An EXCELLENT cheap (near zero cost) fun off-site redundant storage solution is to get together with some friends and pass around a high-capacity hard-drive! You get all your friends mp3's, your friends get all of yours, and each one of you is a completely redundant off-site backup of all your wonderful mp3s! Sneakernet rules!

Ok ok, you're incremental backup isn't "nightly" but rather becomes "monthly", but hey, you can always do the weekly incrementals to a CDR/CDRW and use that after the initial big burst of "traffic".

Re:Hard to imagine

[slaker]

We're funded by our own contributions - no ads, and our forums are backed up nightly in a distributed fashion to multiple, global locations (i.e. three of the founders with broadband ftp the database backup tarball "just in case" every night).

Re:Hard to imagine

[slaker]

SR depends on advertising to survive. If our community takes traffic away from SR, Eugene and Davin starve. We'd rather see SR flourish, and have our site continue with an identity distinct from SR's.

If our site takes traffic away from Slashdot, um, well, maybe Taco has to wait a few seconds before buying another Anime DVD.

Re:Hard to imagine

[Reziac]

At least W.D. had the balls to issue a recall. Most HD mfgrs have let consumers find out about bad batches the hard way.

As to your master/slave problem, what model is this? W.D., all the way back to their earliest IDE HDs, is the ONLY brand I've *always* been able to rely on accepting slave status to ANY other HD. I've never had a W.D. argue about it, and I have one of just about every size from 20mb (yes, meg, not gig) to 60gb.

Try that with a Conner. As a rule, they won't play nice with another HD unless it's a Conner. (May also apply to some Seagates after they ate Conner, since word was they rebadged some Conner HDs.)

we need...

[2MuchC0ffeeMan]

we need something like the automobile industry's recall system, but it's too bad nobody tunrs in those registration cards...

hard drives are so important, they should be the most quality product of a computer... you can replace a cpu, motherboard, etc... but without backing up, you can't get everything on a hard drive back.

Re:we need...

[Camel Pilot]

It seems these days every manufacture wants you to "register". And we all know the registeration process is just a marketing ploy to get your personal information so they can profile you and stuff your mailbox/e-mail/phone with advertising.

Re:we need...

[Usquebaugh]

I'm sorry to say that people in IT are becoming lax in running backups. When I started, back in the mid 80s, the only rule was thou shalt not endanger the data. Well that and don't shag the CEOs wife.

Most companies these days are not performing backups of live servers and have no idea how they will recreate the data when a problem occurs.

In my current role I'm using a lot of different machines FTP servers, SEQUEL DBs, Oracale DBs, DB2 etc etc I asked the ops for a copy of the Sequel DB no copy made since installation! Same for FTP. The really scary thing is they quote 95% of systems backed up. So it's CYA time, send emails keep proof of reading and wait for the fireworks :-)

Speaking to other old farts it's clear that a lot of companies are in the same boat, where the hell did all the good ops managers go?

Re:we need...

[swv3752]

Would those be the same email servers that are not backed up?

This might be the digital age and all but think paper, man. Print up a memo saying you need backups and have your boss initial it. Photo copy enough copies to cya.

Re:we need...

[gorilla]

It helps that us old farts have all seen a disk pack explode. Reliability might be down from it's peak a few years ago, but it's still a heck of a lot better than it was in say 1975.

Re:we need...

[tmark]

Given the absolute profit margins on hard drives, and how cheap hard drives are now, I can barely imagine that it makes much financial sense for hard drive companies to implement recall systems. The automobile industry is different because the makers may be liable to huge judgements if someone is injured from a defect they are aware of. But has a hard drive manufacturer ever been liable for large judgements because of a defect in a hard drive ? I doubt it.

Re:we need...

[McFly69]

What registeration cards? Since when you get those on OEM drives? But I may be wrong.

Re:Fujitsu makes hard drives?

[ekrout]

what the hell. stick the with the big names.

Right on, man. Bigger is always better (not!)

Didn't I read this a month ago?

[ackthpt]

I wish I could remember where, but I've read about this already somewhere some time ago.

I've been wondering if the recently revealed electrolytic (ha, spelt it right that time) capacitor problem (bad taiwanese electrolytics) was related.

On a different note, Seagate's ST380023AS and ST3120023AS (Serial ATA) drives which were expected in Mid-October, then late-November, are now, according to a Cnet article a Seagate employee who shall remain nameless, pointed me to, is indicating shipping dates in Mid-December.. hopefully the two are unrelated.

Re:Didn't I read this a month ago?

[Jucius Maximus]

"On a different note, Seagate's ST380023AS and ST3120023AS (Serial ATA) drives which were expected in Mid-October, then late-November, are now, according to a Cnet article a Seagate employee who shall remain nameless, pointed me to, is indicating shipping dates in Mid-December.. hopefully the two are unrelated."

That sentence deserves to be taken out and shot.

Thats why I like Maxtor......

[vertical_98]

But seriously, I have a 5400rpm 1.2g Maxtor that has been in use for over 4 years. I had a 7200rpm 20g Seagate that crashed after 14 months in a machine. I think the combination of high rpms with super dense platters is what is causing the most problems.

Of course, My father thinks that people just don't give a shit about quality any more.....

Vertical

Re:Thats why I like Maxtor......

[bogie]

I hear people swearing allegence to a particular brand of consumer Hard Drive all the time, but I don't buy it.

It sucks that your Seagate died and I'm not trying to convince you to buy another one, but in general the reliability for Seagate, Maxtor, and WD's consumer drives are all about the same. If you had bought a defective Maxtor you'd be saying the same about Maxtor and praising the new Seagate you just bought.

Re:Thats why I like Maxtor......

[Zathrus]

I think the combination of high rpms with super dense platters is what is causing the most problems

Well... kinda...

A vast number of problems are being caused by the side effect of high rpms and dense platters -- heat. Modern drives get really, really hot, and most people don't adequately cool them. Heck, they don't even adequately cool their CPUs.

Look at the operating temperature of your drive. Get a probe thermometer and read the ambient temperature of your case. Then realize that the air around the drive is probably 5-10 degrees C hotter than the ambient temperature, and unless you've specifically addressed it there's little or no ventilation of the drive cage.

So most people end up operating the drives in excess of their rated operating temps... and they fail.

There are some easy things you can do for drive ventillation - the easiest is to put the drive as far down as you can get in the case. Most cases vent from bottom front to top back. Take advantage of that. More extreme measures involve mounting a heat sink on the drive or even fans (either on the drive bay or to the sides).

Re:Thats why I like Maxtor......

[1WingedAngel]

Darn kids.

I've still got a Maxtor 7080AT (80 meg) drive in my toolbox with DOS 6.22 and a variety of DOS-based systems diagnostics on it.

Bigger != Better.

Re:Thats why I like Maxtor......

[Citizen of Earth]

There are some easy things you can do for drive ventillation - the easiest is to put the drive as far down as you can get in the case. Most cases vent from bottom front to top back.

I have found that an effective way to ventillate an entire computer, including hard drives, is to remove a side of the case an position a desk fanto blow directly into the case at full power. In all seriousness, it's *very* effective. It may be a little noisy, though.

Re:Thats why I like Maxtor......

[Zathrus]

It's noisy, it's power inefficient, and no good if you have pets or children. And it winds up putting even more dust into your case -- and dust is a wonderful insulator for heat.

Re:Thats why I like Maxtor......

[Jucius Maximus]

"So most people end up operating the drives in excess of their rated operating temps... and they fail. There are some easy things you can do for drive ventillation - the easiest is to put the drive as far down as you can get in the case. Most cases vent from bottom front to top back. Take advantage of that. More extreme measures involve mounting a heat sink on the drive or even fans (either on the drive bay or to the sides). The world is run by idiots because they're more efficient than hamsters."

I think that power is also an issue. Some power supplies have very weak +5V channels that often drop more than 10%. (*cough*Enermax*cough*) This can also kill a nice HDD.

About heat: One other good strategy for keeping your drives cool is to use a cooling bay. Instead of having 2 x 40 GB maxtors right on top of each other due to the small amount of room in my case, I put one in a 5.25" cooling bay with an integrated fan to get good airflow. This can also prolong the life of your drives.

The cooling bad was pretty cheap (only CAD$10 refurb) but the suction is definitely present through the unit and since it's front loading, I can easily swap drives without opening my machine.

Some modern cases now adays have a cooling fan right next to the HDD mounting area, which is also good for keeping things frosty.

Re:Thats why I like Maxtor......

[Jonboy X]

I have found that an effective way to ventillate an entire computer, including hard drives, is to remove a side of the case an position a desk fanto blow directly into the case at full power. In all seriousness, it's *very* effective. It may be a little noisy, though.

Personally, I have found that an even more effective way to cool down that red hot 'puter is to simply submerge the sucker in icewater. I tried it with mine a few weeks ago and haven't had a single heat problem since!

Re:Thats why I like Maxtor......

[legoboy]

I just had my oldest WD drive fail on me a month or two ago. It lasted about 11 years. Capacity was something like 16 MB, which was enough to be a DOS-running POS (as in point of sale) terminal.

I can only conclude that I'm lucky, because I have never had a drive fail on me in less than three years of normal use, and we're talking hundreds of drives here. Admittedly, they're all 40 GB or less, which may nicely illustrate the 'large drives suck' principle.

Re:Thats why I like Maxtor......

[Citizen of Earth]

It's noisy, it's power inefficient, and no good if you have pets or children.

We use it on a server that has a worn out CPU fan. The CPU heat sink is actually cool to the touch.

And it winds up putting even more dust into your case -- and dust is a wonderful insulator for heat.

I find exactly the opposite: it blows all of the dust out of the case and keeps it away. Be sure to cover your mouth, nose and eyes when first turning on the desk fan.

Yeah

[Tsar Ivan IV]

If you do a search on the net for _any_ manufacturer or _any_ line of products you are likely to find a number of unhappy customers. Every hard disk manufacturer has sent out a bad batch on occasion -- I've had various people recommend to me at different times "Never buy Maxtor" or "never buy Seagate" or "never buy Western Digital" and so on .. because that particular person had a bad experience with a drive.

Re:Yeah

[Animixer]

If you do a search on the net for _any_ manufacturer or _any_ line of products you are likely to find a number of unhappy customers. Every hard disk manufacturer has sent out a bad batch on occasion -- I've had various people recommend to me at different times "Never buy Maxtor" or "never buy Seagate" or "never buy Western Digital" and so on .. because that particular person had a bad experience with a drive.

I've worked in a small lab of various vendor's systems and storage arrays, at any given time, there's probably somewhere around 1000 disks or so running there. They have arrays from Compaq, Hitachi, etc, and many PCs/Unix servers as well. My overall experience is this: Everyone's drives fail occasionally. It's a fact of life, you can't expect that something spinning at high rpm and has heads floating at miniscule distances from platters will last indefinitely.

As far as FC-AL or SCSI disks go, reliability-wise, they're all pretty much the same. I have seen an IBM, Fuji, Hitachi, Seagate, Quantum, etc pop once in awhile, but nothing major. IDE disks are a little different, the IBM 'deathstar' drives caused problems, and the older (sub 8 gig) Western Digitals seem to have a high mortality rate after a few years of use. Same goes for the defunct Seagate 'Hawk' line. Granted, this is beyond the warranty period, but many of the other seagates/ibms/etc are still going strong after several years. I normally swear by Seagate, but I just had to send back an 80gb drive from my home machine for replacement...not sure what the exact failure was, though. I have hit the random-ascii-garbage-during-boot from a 10gig Fuji, but that was an isolated incident.

Advice: Assume that your critical drive WILL fail, and plan accordingly. Now I mirror everything, AND back it up. Keeping redundant online storage is different than maintaining backups of earlier versions of data! Even in a home PC, it is really worth it to get two drives and mirror them, with software or otherwise.

You don't have to go hog-wild though....I'm a storage redundancy zealot, and I tend to go overboard. (Clustered fileservers, two fibre HBAs in each [on different buses], multipath through redundant switches to RAID arrays, power from different circuits to each array, etc. Fun stuff to play with, if you have the equipment around. :-)

wha?

[meekjt]

How is 90% a failure rate? 100% of all hard drives are going to fail sometime.

Re:wha?

[kermyt]

On a long enough timeline everyones survival rate drops to 0%. The first rule of Fujistu is... you do not talk about Fujitsu.

Re:wha?

[IvyMike]

How is 90% a failure rate? 100% of all hard drives are going to fail sometime.

To paraphrase Twain, the difference between "90% failure rate in a year" and "90% failure rate sometime" is the difference between lightning and lightning bug.

Re:wha?

[McFly69]

In theory yes.. but not all the drives are used for the longest time they can be. Drives get replaced and tosed, they get moded to create ash trays, and otherse are just "lose" either by theift or misplacement. Like the other day.. I found my 80 meg Maxtor drive in a sock drive. And I know the last time I used it was around 96.

I've probably already shortened my one's life

[P-Nuts]

I have a Fujitsu drive and I've cut a hole in it and added a window. It still works fine, but if it stops working I'll never know whether it was due to my customization or not.

The Register and Fujitsu

[HealYourChurchWebSit]

This isn't the first time The Register has fried Fujitsu' sushi. Check out an article from this past September entitled PCA attacks 'shabby' handling of Great Fujitsu HDD fiasco.

It makes me wonder if The Register, or at least one of the writers there, didn't get stuck with a few sand grinders doubling as hard drives.

Re:The Register and Fujitsu

[beanerspace]

You may have a point there. Add to your list the following Register articles reagrding Fujitsu:

• Crash! Dud Fujitsu HDDs all over UK
• Ouch! Fujitsu to replace 300,000 faulty HDDs

One word...

[dannycim]

Maxtor. We have over 5000 PC Workstations at my previous job. We've had problems with just about every manufacturer (Quantum, Seagate, Fuji, WD, etc...) except one: Maxtor. Personally, I've got around 8 of them at home, 3 up and spinning 24/7 and one actually trashing all around the place continously (compiles, builds, rendering, etc...) and never had a bad block.

Does your mileage vary?

Re:One word...

[Fugly]

.....We've had problems with just about every manufacturer (Quantum, Seagate, Fuji, WD, etc...) except one: Maxtor....

Does your mileage vary?


My ex-roomate had two 8GB Maxtors fail on him when we built his PII a few years back. The first one failed within a day of use. He called Maxtor who were very helpful on the phone and sent an advance replacement. The replacement drive lasted a little over a year.

Regardless of brand, there are only two types of hard drives out there:
1) A hard drive that has crashed
2) A hard drive that is about to

I think we're slowly reaching the end of magnetic media's life as our primary secondary storage mechanism. There are just too many delicate moving parts requiring extreme precision to even function due to the density of data we're storing. I think we'll see more and more solid state storage solutions replace hard drives and more optical solutions used for backup.

Re:One word...

[0xA]

Well I too have seen it with every manufacturer including Maxtor.

Maxtor 1.2 GB to 2.0 GB models were horrible, I was a tech in a retail store at the time and we sold a bunch of NEC desktops with Maxtor drives. Got a lot of them back with dead drives.

I've seen bad drives or batches of drives from every manufacturer, there is no best brand IME.

Sounds like you've been pretty lucky.

Re:One word...

[Lumpy]

very wrong.

I have several 9 gig U160 IBM and MAxtor drives that are rock solid and I am betting will run for another 5 years.

The only thing i worry about on my drives (Scsi Enterprise class drives) is that they have been spinning so long that the bearings are worn out and will not spin back up after one of the rare shutdowns (2 shoudown in 5 years... you get very worried about powering something down that has an average uptime of 24 months)

Moral of the story? dont buy consumer grade crap. Yes U160 Scsi is more expensive, but I do not worry about it dying on me for silly reasons.

Re:One word...

[Fugly]

I have several 9 gig U160 IBM and MAxtor drives that are rock solid and I am betting will run for another 5 years.

It's stupid to bet anything on the life of a hard drive.

The heads are so close to the platters that a smoke particle can't fit between them. If you were to put a fingerprint on the disk, it would hit the head as it spun by. On top of that, they're spinning extremely fast, (7400rpm pretty much being the norm these days with 10krpm drives out there). Assuming the disk is 3.5" in diameter and travelling at 7400rpm, the outside edge of the disk is travelling at around 75mph. What is gonna happen when that head collides with the disc due to a mechanical failure? Everything is gonna be hosed - stick your foot out onto the road while you're doing 75 in your car sometime.

I once heard an analogy regarding this whole mechanism. I don't remember the exact numbers or wording but gist was that a hard drive head reading data was analagous to a 747 flying at 500mph 1 centimeter off the ground and counting every blade of grass that passes underneath. Something like that...

These are delicate machines with many moving parts. The only parts of a PC that I've found to be less reliable than hard drives are CD-ROM drives and CPU fans....

Backup everything. Backup often. (Well, that or get cocky and lose all your data, no skin off my back)

Re:One word...

[Lumpy]

Backup everything. Backup often.

DUH! any sysadmin or netadmin that doesnt have a solid backup plan and disaster recovery plan in place needs to be fired,publically humiliated and completely blackballed in the industry (I.E. call all your tech industry friends and tell them not to hire XXXXX because he/she is a idiot, and spread the word.)

even with that said.... I still bet these drives will run another 5 years. so all my backups will have been useless.... I love it when my backups and disaster plans are never used.

Lotsa Fujitsu Drives

[truffle pig]

I don't know I have a bunch of Dell Servers that are using Fujitsu Hard Drives in RAID Arrays. In the past year and a half of using the dells with Fujitsu drives, we have only had one drive our of about 40 go bad. I can't speak to their IDE drives but the hot plug SCSI's are working pretty well.

Our experiences

[Superfreak]

I work for a small company as the solo-IT guy. We have had a total of 18 Fujitsu drives, all 10GB, from one batch purchase in October, 2000. I've had one failure out of them, and we're at the two year mark, so I certainly haven't seen a fail rate anywhere near whats described. Just another anecdote for the pile...

First thing I found in a google search...

[Evro]

http://www.classactioncounsel.com/fujitsu-litigati on.htm

These cases are being brought on behalf of purchasers of the MPG3xx series hard disk drives, irrespective of the entity from whom it was purchased. Additionally, Hewlett-Packard is sued in connection with its sale of the hard drives as components in certain HP computers and its processing of warranty claims. Please note that the MPG3xx hard drives were also distributed to retailers and to other computer manufacturers, although none of them have been made a party to the litigation at this time.

http://www.theinquirer.net/?article=5666

This took me 5 seconds. http://www.google.com/search?hl=en&ie=UTF-8&oe=UTF -8&q=fujitsu+hard+drive+failure&btnG=Google+Search . I'm not sure what the point of this "Ask Slashdot" is, is the person just trying to inform everybody that there is a problem with Fujitsu drives? I didn't see an actual question in that "Ask Slashdot" except for the ones Cliff tacked on.

Re:First thing I found in a google search...

[mickwd]

"...is the person just trying to inform everybody that there is a problem with Fujitsu drives?"

Well, judging by his user details:

User #624901

Michael_Angel has posted 0 comments.

has submitted 1 stories.

Have Fujitsu Harddrives Been Failing in Record Numbers? on 12/11/02 18:06

that's exactly what he's trying to do.

Sounds a bit fishy to me. He wouldn't happen to work for a competitor, do you think?

I've only lost one...

[BenTheDewpendent]

I have about 60 fujitsu drives at work. One failed a couple weeks ago. However my friend with same job at another location has sent back around 20 drives if not more, in the last few years. While hearing from yet another friend who worked for the state that he had fujitsus failing all over.

Problem
I think so.

Failed Drives

[Tha_Big_Guy23]

The company I work for built and installed 15 systems with identical configurations, all having a 20Gb Fijitsu hard drives. Each system was installed within the same week. Approximately 10 weeks later, each of the hard drives failed, in almost the same order they were installed.. I'd say this is definately a problem they need to look into.

IBM 300PLs

[Mechamse]

I work for a large Oil Company in the North West US. We have roughly 1500 IBM 300PL systems in our inventory. Of those 1500 we have had to replace 700 or so Fujitsu HDDs due to various problems. Fortunately for our sake, IBM was using a mixed hardware pool when our systems were built because out of 1500 systems, all of the Fujitsu drives have now been replaced. Now we are suffering through Maxtor drives, but that is a tale for another day. This to me seemed to be a huge problem. We filed a complaint with IBM on this issue for not having a recall of the effected drives. IBM and all of the service centers in our area know of the problem, but that doesn't seem to be of importance.
Not only is this the largest mass failure of a product, but also probably the largest cover-up to protect all of the parties involved.
What really takes the cake on this whole issue is the pure audacity of Fujitsu in making this appear to be within the bounds of standard failure. That will keep me from ever using their equipment.

Re:IBM 300PLs

[Halo-]

I suspect you got a small/weird batch of 300PL's. I work for IBM in software development. (so I know nothing about the hardware) Nevertheless, I can tell you that the 300PL used to be the standard issue PC to all developers. We've got zillions of them around, and most are still in service. These things have the crap beat out of them all the time disk-wise, but I've never seen (or heard) of a disk failing. I certainly don't debate that you got a bad batch of HD's, and it's possible IBM hasn't responded as well as they could, but they might have internal reasons. (Legal restrictions/haggling over who is allowed to claim a part is defective, possible libel... etc...)

Re:IBM 300PLs

[Chanc_Gorkon]

Want to say I have a 300PL that's on it's 4th year. It's been her almost as long as I have. The drive is just going on like gangbusters. We have not had alot of problems with these machines. Now the newer IBM's, well, I can't say I care for how the power button works (you have to hold it in to power off), but other then that they are fine. I will say this...in the many years I have been around computers I have had people tell me never to buy hardrives from EVERY manufacturer. Even Seagate and IBM. Sure, you can get a bad batch every once in a while, but, in general, I have only had 2-3 hard disks fail on me and in most cases hard disks have been upgraded long before they fail.

Re:IBM 300PLs

[Col. Panic]

We also mainly use 300PL's and IBM has a preventative firmware update for the model 6565 here. Note that if the h/d has already failed the update does no good.

(Hope I found the right one - I've been using the same floppy forever :/)

Re:IBM 300PLs

[Chanc_Gorkon]

Not sure....it's a Pentium 3 500. It's quite old. I will be using it a bit longer I think. I share it with the wholke department now, but it will be mine when the new system comes in to replace it. It's not to obad for what I do. They even run XP at a decent clip too once you up the memory although it could DEFINITELY use a better graphics card (Trio 3D sucks!)

I for one am happy with my Fujitsu harddrive...

[suman28]

One of the many harddrives I have is a Fujitsu and they are prone to go bad, but as much as you say. I replaced 2 out of 20 that I have. I even got a better replacement because they were out of the drives I bought 2 1/2 yrs ago. So, I am happy.

Re:I for one am happy with my Fujitsu harddrive...

[Hadlock]

I even got a better replacement because they were out of the drives I bought 2 1/2 yrs ago.

i think you just hit the nail on the head as to why drive makers no longer give 3 year warranties on drives. they go through inventory/technology waaaay too quickly and it's just uneconomical to give out new drives for ones that failed after 3 years.

It affects HPs too....

[WhiteKnight07]

My friend has an HP Pavillion xt926 with a Fujitsu hard drive in it. That thing gave her no end of problems. We had to pull it out of the machine on a daily basis and put it in my machine to correct file system errors caused by bad sectors that kept her machine from booting. When she finally called HP about it they said it was a common problem and replaced her whole system for free. They were very eager to keep it quite and make her happy, aparently they use the same hard drives in many of there low end server platforms. The machine was running Win2k with NTFS and worked perfectly after they replaced it.

Drive Service Company seems to agree

[Brett Glass]

Before I buy a hard drive, I always ask data recovery companies what they think of the most recent models. This page, created by Drive Service Company says the following about Fujitsu:

Fujitsu (Desktop drives only) Their 10, 15, 20 and 30gb desktop models have been failing left and right with either servo loss or electronic failure. Notebook drives are only so-so but are no longer manufactured. They have had so many returned drives, that they have stopped making drives all together.

It then goes on to say:

Fujitsu Notebook drives of any kind are prone to head crash, desktop drives are bad now too, sorry. Again, they have stopped making drives and now barely support what is left out there.

Believe it or not, their most recommended brand is now Seagate (the high end models). And they strongly recommend anything with a SCSI interface over IDE -- not for performance reasons (there's really not that much difference if you cache) but for reliability.

Re:Drive Service Company seems to agree

[Magila]

I'm sorry but thoes rankings are totaly off. They put WD at dead last yet Dell, which goes through many times as many drives as these guys see, rank them as thier #2 suplier (they were #1 for a while but a single slightly bad batch droped them to 2nd, that's how demanding Dell and other major OEMs are). Clearly they're nowhere near as bad as IBM, which is ranked second (WTF?).

Re:Drive Service Company seems to agree

[Fjord]

You're right, it is in the mechanics, but the fact is that SCSI drives are used in high performance applications and so are expected to be more reliable. This reliability comes with a price tag, since SCSI drives tend to cost significantly more per meg. IDE drive could be as reliable, but they would cost more too.

To use a crappy metaphor, it's why a ferrari performs better than a honda civic, even though they have basically the same mechanics (gas tank, gas line and pump, engine, exhaust system).

Re:Drive Service Company seems to agree

[Fjord]

One thing to note here is that there is going to be sampling error wit this analysis. Only people with enough money and important enough data are going to take their drives to a data recovery service.

Re:Drive Service Company seems to agree

[GigsVT]

IDE drive could be as reliable, but they would cost more too.

Maxtor is closing this percieved gap at the beginning of next year, their Maxline II are ATA drives with the reliability ratings equal to SCSI, aimed at bulk storage archives.

http://www.maxtor.com/en/products/ata/enterprise _a pplications/maxline_ii/

You may ask why a MaxLine II, when there was no MaxLine I... Well, the 5400 RPM 120GB and 160GB 540DX were basically the MaxLine I. You will notice that these drives have been discontinued in favor of the 7200 RPM forms with 1 year warantee. I suspect the covert MaxLine I drives were actually aiming for SCSI reliability, because we have over 50 of them in service here in large archive RAIDs with no failures for over a year.

Re:Drive Service Company seems to agree

[John_Booty]

To use a crappy metaphor, it's why a ferrari performs better than a honda civic

Oh, I don't know about that. You can expect to get 150,000 miles from the Civic even if you treat it like shit, but you'll be lucky to get 30,000 from the Ferrari! :P

Re:Drive Service Company seems to agree

[Reziac]

On reading your cited reference -- this guy doesn't like ANY hard drive, at least not for long, which tends to make me wonder -- after all a data recovery person isn't going to see the mass of live drives, only the dead ones. I notice he didn't make any attempt to correlate marketshare vs number of dead HDs that come into his shop. Also, recommending SCSI over IDE for reliability doesn't necessarily wash either.

Frex, Micropolis -- at one time rock-reliable SCSI HDs. But their final year's worth of SCSI HDs (mostly sold by surplus dealers after Micropolis went tits-up) have had, in my observation, a near-100% failure rate. Bad handling in transit or bad HDs? We'll never know.

As to what I've heard locally about Fujitsu -- the general comment is "really unreliable", especially their SCSI HDs. I've never bought any Fujitsu HDs but at one time had been looking into 'em for consumer SCSI. No one had a good word to say about 'em, including HD dealers.

Too bad their HDs aren't as durable and reliable as their floppy drives. I've got several Fujitsu 1.2mb 5" floppies dated 1986 that still work just fine, and some were the everyday data drive for systems in the pre-HD era. Don't think I've ever seen one fail yet.

I think with the corner-cutting that all the HD mfgrs have been doing of late (cf. the recent cut in consumer-HD warranties to one year) we're going to see a steep increase in crap HDs from everyone, at every level. :(

Re:Drive Service Company seems to agree

[Reziac]

And look at the high failure rate in Yamaha SCSI CDRW units. I don't know of a single one over two years old and still working. (So far I know of ten dead ones, about half and half SCSI and IDE. Dismangled one of each and found the innards are *identical*.)

I can tell you from personal experience, YES!

[19Buck]

I recently ended a Contract position i was working as a Service Technician for Gateway. My job was basically to perform diagnostics and replace parts on the systems (duh!).

I can tell you from experience, that Fujitsu drives were easily, by far and the way the most failed brand of drive that we replaced. It used to be Maxtor's that died in record numbers some time back, but the difference there is that Maxtor's were much more widely installed.

A majority of the time that we had a system in with a bad HDD failure, we'd say "I bet it's a Fujitsu".. 90% of the time, that's exactly what we'd find inside the computer. After a while, we just stopped doing diagnostics troubleshooting on Fujitsu drives..we'd just close the system up and order a new drive.

And if we got a Fujitsu drive back as a replacement, we wouldn't even install it, we'd close it up and send it back requesting another replacement HDD.

They stopped us from doing that, said we couldn't send back drives that were working fine just because we didn't like the brand. So.. we said "ok", and resigned ourselves to the fact that the unlucky customer who got a Fujistu replacement drive would be back within a month.

And guess what? A majority of the time.. they were.

After some research...

[twoslice]

I found the actual cause of the Fujitsu HD malfunction...PrOn!

Apparently they are very allergic to PrOn and the more PrOn you have the more likely they are going to die.

And from all of the horror stories posted here on /. about bad Fujistu drives it is probably true!

Re:After some research...

[twoslice]

and I thought girls were always after the big "O"

Silly me...

Dell Inspirons had bad Fujitsu drives

[Kheldar99]

I had a Dell Inspiron with a Fujitsu drive that used to make clicking noises all the time. When I asked Dell about it they sent me an IBM drive to replace it... they didn't say if they were having massive problems with them but they did indicate that there were a lot of these swaps going on.

Definitely a problem in Ireland...

[1by1]

I know that there in the west of Ireland that we are having severe problems with Fujitsu drives, especially ones beginning with the serial MPG3. The drives seem to have been giving up the ghost in high numbers for the past 6 months. (ie: in September, one site that had 15 PCs suddenly had 3 hard drives go in a period of 5 days...all three were Fujitsu with that serial number.) I seem to be receiving a call about once every two weeks now about a failed drive, and the majority of them have been Fujitsu ones...

Known Problem with some drives

[ivan256]

Some Fujitsu drives manufactured just over a year ago have faulty chips from Cirrus Logic on them that cause the controllers to fail. Check the article for details. I believe there's a class action lawsuit in progress that you can join.

Lost 2 hard drives last year

[vorwerk]

I lost one of my SCSI drives last year (a 4GB Quantum Atlas-2). I was not amused. It was still under warranty, so Quantum (now Maxtor) replaced it with another Atlas-2. The replacement (which came with a 90-day warranty) failed shortly after its 90-day warranty expired. Bummer.

I can't speak for the rest of the industry, but I can say this: none of my older (~300 MB) hard drives (which I've been using in my 486s) have ever failed. They rattle a little, they're rather slow, but never once have they let me down. Can the same be said of more recent media? I suspect not.

Fujitsu Drives - Bad for some time

[CrackHappy]

I owned a small computer shop for three years. We used Fujitsu drives for about one of those years. The main reason was to drive down our costs. However, it turned out that it cost us more in the end. We had a failure rate around 60%. Most of the failures were not spectacular, which made it worse! Strange things would happen. This was about 6 years ago, so I'm not surprised to see that they're having even worse trouble now. I also recommend Western Digital. They have been quite reliable for a long period of time for me and my users.

Re:Fujitsu Drives - Bad for some time

[runderwo]

I also recommend Western Digital. They have been quite reliable for a long period of time for me and my users.

Click, click, click, click, click, click, click....

This is easy for Fujitsu to fix.

[Kenja]

They just need to fire the guy with the rubber mallet at the end of the production line.

had 2 fail in 4 months

[shomon2]

Yes, I had this problem on my puter: I have a compaq that comes with "same day support" - which in spain can mean a lot of things. One day my drive broke down. It was a fujitsu. The tech guy came the next day with a new one and even let me keep the old one for a few days while I submerged myself in hardware trying to mount it and copy my stuff out. Yes I do keep backups, but it's nice to just copy stuff back exactly how I had it.

Second time, same problem: hard disk just stopped. Same exact one as before (although I don't remember what it is just now exactly). The same day technician this time was a few days later than last time, because they'd "had to order the part from madrid". The guy didn't even check the drive. He just changed it. He said: All these fujitsu's just crash on us. I don't even check them anymore to find out why. We ordered in a seagate. This time everything was lost. The computer couldn't even read the broken drive.

Ale

Re:had 2 fail in 4 months

[rlangis]

Next time you have a drive that 'just stops', try this:

Stick it in a freezer bag with a dessicant pack. Put it in the freezer for a few hours. Take it out, hook it up. If it doesn't spin up, tap it lightly on the side with a hammer or somesuch UNDER POWER. If it spins up, furiously copy data after boot, and THEN RMA it.

I did this with a neighbor's drive. Still kinda bummed I didn't get to fsck her, too. ;)

Not bad, /.

[Loki_1929]

Not bad slashdot, only missed the boat by 2 months this time. The Register has been following this for a good long while now.

It's Our Own Fault, To Some Degree

[occamboy]

A very large problem here is that almost everyone buys on price alone. Over the years I've seen a number of manufacturers of really superior stuff get beaten up because manufacturers and consumers are so price-oriented.

IBM and Fujitsu hard drives used to be the best -- really really solid and reliable. But they cost more. I remember when, several years ago, Fujitsu dropped their drive prices to bring them in line with seagatemaxtorquantumwesterndigital... -- I was surprised that Fujitsu could build a much better drive than their competition, at the same price. Turns out that they actually could not -- Fujitsu drives quickly started getting ungood.

Sigh. I'll gladly pay a little more for quality, but since few others will -- I'm hosed.

Re:It's Our Own Fault, To Some Degree

[multipartmixed]

> Fujitsu drives quickly started getting ungood

I took one look at the Picobird series when the first came out -- 1996? -- and say, "uh-uh. my Fujitsu days are over, these look *cheap*". I'm glad I did, they certainly weren't the most reliable drive from day one.

My 9-year old Fujitsu drive actually looks like it came from a different design team/plant entirely. Big surprise, it still works.

IBM DeskStar

[T-Kir]

When doing my internship a friend at work recommended IBM drives, mainly on the principle that they had the best record for reliability. I have been buying IBM drives for years now (apart from a nice quiet Maxtor) with now problems whatsoever.

But about two years ago, my uni housemate got an IBM DeskStar drive which died on him after 3 weeks from getting it. Turns out he got the drive where they had the glass platters, and the heads on the drive literally crashed and cracked the platters. He had all his Uni work on there, although we kept yelling the work 'backup' to him. I don't know how many of these drives had this problem, but IBM pulled the drives as soon as they found out about the problem.

oops

[tps12]

I have no idea whether or not Fujitsu's hard drives have been "failing in record numbers." But if they haven't, then I imagine that Slashdot will be looking at some sort of lawsuit in the near future. Well done.

Is it silly not to do RAID/0?

[wytcld]

The ex-girlfriend (still friend) had her hard drive failing. "Okay, order a WD, I'll come by and replace it." So she did, and the WD was DOA, and we end up out at Staples paying too much for a smaller Maxtor. But even too much is so cheap these days. Given that drive manufacturers are barely holding on in this market, and are all scrimping on quality control, does it even make sense not to install drives in pairs with RAID/0 mirroring? The cost of the second drive is far less than the time involved in even doing regular backups (although these are still a good idea for when to tornado strikes), let alone restoring a system.

Re:Is it silly not to do RAID/0?

[Loki_1929]

"does it even make sense not to install drives in pairs with RAID/0 mirroring?"

Well, in all honesty, that statement doesn't make much sense as is. ;)

RAID-0 is striping, meaning there's no redundancy. RAID-1 is what you're looking for; that's mirroring. As for your question, it makes sense if you have valuable data and need maximum uptime to run a RAID-1 array. Extra costs are somewhere in the neighborhood of $150 for the card and the extra drive, unless you go el-ultracheapo, in which case you probably don't care anyway.

The short answer is for the vast majority of home users, it doesn't make sense. For anyone running a home office, it should be one of a couple different backup methods, as it only guards against physical failure.

Re:Is it silly not to do RAID/0?

[Loki_1929]

"As opposed to what it'll cost me to get my data back should the system be hosed in a non-RAID configuration..."

Actually, if you use 2k or XP on NTFS, there are lovely tools for recovering data should the drive take a dump. In all honesty, if it's important enough to require a second drive, it's important enough to be backed up regularly on tapes or (better yet) CD-RW's. A second drive won't help you if you encounter a viral infection or accidentally delete needed things; whereas a regular schedule of backups to a remote medium will protect you from almost anything. Since most people already have a CD-RW (commonly preinstalled on newer systems), the costs is only that of the CD-RW discs themselves, which are quite cheap.

One other thing...

"I'm thinking of doing a RAID-1"

RAID-1 would be a mirror, not a stripe. A stripped array wouldn't protect your data at all, only make I/O operations faster. On that note, RAID-1 (mirroring) will also somewhat improve read operation speed, as it'll read from both, but won't affect writes.

Bad News for Sun

[niola]

This is not a good turn of events for anyone who buys hard drives from Sun. At Princeton University, every time I order a hard drive for my Sun servers, it is actually a re-badged Fujitsu as of this past summer. Prior to Fujitsu all Sun drives were actually Seagate, and they were very reliable.

Though I find this news disturbing, I have to say I have personally not had a failure of any of my Sun/Fujitsu drives yet. Knock on wood...

Perhaps this problem is not in the higher-end 10k RPM SCSI drives?

Re:Bad News for Sun

[pmz]

This is not a good turn of events for anyone who buys hard drives from Sun.

Actually, this article is largely irrelevant regarding Sun computers. Sun sells tested stable configurations...I'd bet that their drives are just fine. Also, 10K RPM SCSI drives are a different breed from the IDE trash discussed in the article.

Basically, Sun could purchase drives from a different statistical pool than the ones known to fail. Whether they actually do, I don't know, but their reputation is at stake in matters like these.

Re:Bad News for Sun

[autocracy]

Sun buys from a different pool of hardware than the rest of the world. A Fujitsu drive bought from Sun will be better quality than one straight from Fujitsu... it's part of the deal they have. Fujitsu can't afford to have Sun come crashing down on their heads either. Same things goes for pretty much any part from Sun. If you've ever wondered why a 128 MB stick of RAM costs so much, it's because Sun will guarantee it...

I have no connection with Sun other than I want to own some of their equipment...

Re:Bad News for Sun

[pmz]

Anyway, if you browse the full component lists at SunSolve, you can see exactly what model numbers Sun uses in their systems. For example, the IDE-based Blade workstations list various Seagate drives, and the FibreChannel Blades list a mix of Seagate and Fujitsu drives. Again, I'd bet these drives have been selected for specific MTBF criteria (e.g., 24x7xMany Years).

Please don't classify Sun in the same boat as your local Best Buy, Circuit City, or other "white box" pushers.

Had 100% failure rate on Fujitsu drives

[msoftsucks]

I had bought 15 such drives 1 1/2 years ago for custom clones. All of these drives have since died. Never again will I buy Fujitsu. Its not that they died (other vendors are no better). Its the fact that they lied, and gave me no support in resolving the issue. They didn't even care that their products were failing with such a high rate. In that same period of time, I had bought some Western Digital drives, that have since died also. But when I called WD, and gave them the S/N they sent me a replacement drive, no questions asked. Compare the two and tell me who you would like doing business with. Vendors who don't stand by their products should be run out of business. Would I buy a Western Digital drive today? You betcha! Would I buy another Fujitsu? No way! Not even if they paid me!

Fujitsu MPG3307AT

[yoink!]

I've had a 30GB 7200RPM Fujitsu drive for some time now... it's been in use at least a year and have not seen any trouble with it yet. The article did scare me into making yet another backup, but seriously, there are a lot of good posts here talking about how we tend to blow everything out of proportion. Yes, the register seems to have quite a bit of proof of faulty drives, and yes, this drive isn't exactly the cream of the crop. I can say that it has been doing it's job, and so far it hasn't made any strange noises or emitted any foul odours. On the other side of the consumer spectrum, it's not unusual for an automobile manufacturer to recalls tens of thousands of automobiles for "issues" that can actually result in death, yet people continue to drive those cars (many of whom didn't even get the recall notice and are driving potential execution chambers.) The fact is, at least with disk drives and data, if you don't have a backup then it's your own damn fault. It's like preventive health care for your information.

What I learned working at a disk drive company...

[Anonymous Coward]

I worked on an IT project at Western Digital. The drives that pass quality tests with flying colors go to customers like Dell, Compaq, etc. The lower quality ones go to Frys, Comp USA, etc.

FUJITSU SUCKS! (really witty eh?)

[Anonymous Coward]

I work in a tech support dept. for a mid sized university. We recieved 18 Gateway desktops with Fujitsu drives in them. Within 6 months, 12 have died. All but one of those were replaced with non-Fujitsu drives and work perfectly. The single Fujitsu replacement was a different model and series that its predecessor and it failed within 3 days!

RAID

[mrroot]

As the quality of drives is getting worse, more people will be turning to RAID to protect their desktop storage. It's no coincidence IDE RAID is becoming more common on motherboards, and the hard drive manufacturers aren't going to shed a tear about selling twice as many drives.

Here we've had TONS of them fail

[docstrange]

One of the desktop models that we ordered for widespread deployment in our enterprise was the compaq ipaq desktop. The 10GB Fujtsu drives that came in the 866mhz ipaq desktop. "Hard Drive Model MPG3102A" are failing left and right. I would estimate that I have had to replace on average 2 of these drives a week for the last few months. The drives started to magically fail after about a year of use. Fujutsu says that the drives should be covered under compaqs warranty. (which is only 1 year, and since gone), and refuses to help us replace their defective drives. The funny thing is that these drives have a known hardware flaw, and there is a firmware out there that tried to fix it. All of our drives have the alleged "fixed" firmware, yet they still are failing. If anyone wants a box of the 50 or so fujitsu paperweights that I've got over here please let me know. I really wish we didn't have to eat the cost of all these drives.

Re:Here we've had TONS of them fail

[Kintanon]

I'll pay shipping if you send me all of those dead drives.
I love to take them apart, pull out all of the super strong magnets and build stuff out of them.
How many do you have?

Kintanon

I've seen this...

[EvilNight]

Had a Fujitsu 40GB report it was failing (via SMART), so I popped a brand new Fujitsu 40GB drive in off our parts shelf to clone the old one to, and it reported the same error. After spending an hour making sure it really was the drives failing and not the machine I cloned it over onto an old maxtor and sent both in for warranty service. Got new ones back promptly in 11 days. I'll have to keep an eye on them now as well, guess I'll be putting them on the shelf with the spare 75GB IBM Deskstar...

I have a Fujitsu with bad boot sectors...

[Newer Guy]

It's an 8 gig...it can not be used as a boot drive but works just fine as the second drive in my computer. With them taking over IBM drives..... Ohhh BOY!~shudder~

75% for me

[koancomputers]

I would say that the two year failure rate for Fujitsu harddrives sold in my shop was as high as 75% up till 2001 - when I stopped selling them because the RMA's were driving me nuts. I'd also say under %10 for Maxtor and Western Digital, the other drives I sold...

Theres is no such thing as "the best HD manf".

[MFHFozzy]

There is however, good bad and best LINES of HD products out there. Ive been running a test lab of over 100 computers, and supporting another 100 or so for 7 years. In all that time, i have dealt with just about every line of HDs from every manufacturer. Every single one of them has a line of HDs that suck. Had one shipment from Seagate once where 17 out of 20 HDs went bad within 6 months. In all the yeras, the only IDR drives i trust now are the Seagate U series drives. They arent the fastest, but they are built to last. And they are quiet. Ive bought hundreds of them for work by now, and have never regretted it. Out of all of them, maybe 5 have gone bad.

Fujitsu HDD Failures are quite common

[SkankhodBeeblebrox]

Until recently I worked as a technician in a retail computer shop, and we had terrible luck with Fujitsu hard drives, the MPG3204AT in particular. Some drives wouldn't detect at all on POST, most just had the "click of death", and as a result were subjected to the "freezer of doom" so we could try and rescue some of the customer's data (not that it usually helped)

Maxtor/Seagate/WD drives seem to be quite a bit more reliable, but one of the OEM's we were buying premade systems from was using "Fush*tsu" drives, so we encountered quite a few of them (I'd say at least 50% failure rate)

We also had problems with MSI K7T Pro mainboards we recieved from the same OEM, so it could just be we were getting shafted w/ known-bad product.

In any event, in the past few months I've seen the same articles on The Reg and other spots, and I'm not at all surprised to have seen it.

It may be news to some of you that Fujitsu has subsequently pulled out of the desktop HDD business (they still manufacture laptop and enterprise drives) Fujitsu's Hard Drive Lineup

Mine are OK

[dusanv]

I know there will be about 10000 people here saying Fujitsu sucks but I have to say that my experience has been different. I have a bunch of ATA Fujitsus (MPDxxxx & MPExxxx). They have been all on 24/7, some for four years straight. Excellent drives, running very cool, unlike the stupid IBM's 34 & 75 GXP series.

My two cents...

Gotcha Beat

[multipartmixed]

> My five year old Fujitsu 4.5GB SCSI-2
> HD is still going strong.

I bought a 1.054 GB SCSI-2 Fujitsu in 1993. It cost about a grand US. It still works. :)

It actually performs decently, too -- 5400 RPM, 10.5 ms access time, 512 KB cache. Not bad for a piece of 9-year old hardware to still perform about as well as entry-level current stuff.

The freaky thing about that drive, is that you can use one corner of it (where the arm pivots, presumably) to pick up quarters. It will hold four if you're patient.

Re:Gotcha Beat

[Reziac]

Pikers.. I've got a 1.3g Seagate SCSI dated 1987. Gods know what it cost new (I got it as salvage when previous owner retired that system at the ripe old age of 11 years). Still works, with good performance.. and at 20W, you can fry eggs on it even when it's sitting out in the open air.

One does have to wonder what possessed you to use a HD to pick up quarters ;)

MPG3 - MP3

[xenofalcon]

we are having severe problems with Fujitsu drives, especially ones beginning with the serial MPG3

Must be one of those new RIAA-compliant hard drive models.

Check those part numbers..

[Dynamoo]

We too have had a failure rate of over 50% on 10Gb Fujitsu HDs about 18 months old fitted to our early Compaq DeskPro EXDs. Compaq have a BS firmware fix that doesn't work. We insisted that Compaq give us replacements for ALL the HDs, which they did, but they won't admit to the problem.

If you've got a installation of more than a couple of these HDs you'll *know* about the failure rate. If not, then the 10Gb unit is part MPG3102AT dated early 2001 - if you have one of these replace it NOW. I guess that MPG3204AT, MPG3307AT and MPG3409AT are faulty too.

There's an interesting thread here. But trust me, if you have a home PC with one of these units in, replace it right now.

Unfortunately I've Suffered

[Psychotext]

I work for a large soft drink making company that through a third party contractor ended up with Fujitsu hard drives in all of the equipment that we use to control the blending and dosing (Putting in bottles) of our drinks. About 6 months ago we started noticing failures of these machines in large numbers but could not work out what was causing them.

We initially put it down to heat (Surely these drives can't all be naturally broken) and fitted expensive cooling gear. They kept failing.

We then thought that it was the contractor messing with the machines that caused the failures so we put in better access control (Simple key to allow dial in). This didn't fix it either.

It was only when I ordered 80 western digital hard drives and started replacing the Fuji's once they broke that we started noticing that the WD drives were not breaking. We are currently scheduling downtime of the plant to replace the rest (Not easy given it all runs 24x7 and we are always behind schedule).

Needless to say we are not happy at all. I would hate to think how much money all that downtime has cost the company, and how much lost sleep the IT team has had to endure from the endless call-outs.

Which models?

[Wakko Warner]

It's pretty stupid to say "Fujitsu hard drives are bad" without giving model numbers. IBM drives were fine, except the horrid GXP line. The MAN-series 10K RPM SCSI drives I have at home are all running beautifully. As far as I know, it's just a certain line of cheap IDE Fujitsus that are displaying these problems.

Please be more specific.

- A.P.

Re:Which models?

[FBCrack]

In my case it's and MPG3409AT and it's rev. no. is 4. I have another disk with a higher rev. no (6) and it hasn't failed (yet). I have also heared of some other people having problems with this model too. The problem with this disk is not with the "moving" parts but with the cirucit board as I have tried to change it by the other disk and it worked perfect. I have noticed the tho boards differs in some of the SMD components.

Bizarre.... It happened to me just today!

[Dman33]

I just worked on a server here at work today.. inside was a Fujitsu 34GB U160 hard drive. Anyway, I got done doing what I had to do, plugged it back in and "No Boot Device Found"...

Great. Then it dawned on me.. this happened to be before on this same server! I opened it up, adjusted the cables, power on, and nothing. Hrm. Did it again, this time wiggling the power molex connector (which seems quite loose) and now it works! I am calling the vendor for a replacement ASAP because I think there is just something wrong with the drive.

Serious IBM notebook drive problem

[Anonymous Coward]

Some time ago we built a system which had roughly 150 IBM Travelstar, 20GB notebook drives in it.

Whenever we turned the system on, there would almost always be some drives (roughly 3 or 4) that made 'clunking' sounds for about 20 seconds. Consequently, the system that one of those drives was in would not boot because it couldn't read from the drive. It wouldn't always be the same drives, but some would do it more frequently than others.

Originally we ran these systems with a in-house written BIOS, but in the end we where able to reproduce the problem without a BIOS chip at all (that is, the clunking would happen, of course the system would never boot). We looked at the power up voltage and it was well within spec.

IBM engineers came over to look at the problem and took a drive with them to analyze it. Nothing came out of that exercise and we ended up swapping all the drives for Toshibas, after which the problem never occured.

What amazed me was that IBM recognized the problem and never came through with an explanation, let alone a fix.

not entirely their fault

[epine]

Please, somebody post the link to the story about the lawsuit between Fujitsu and the supplier who added phosporous to the molding compound. I've bought a fair number of Fujitsu disk drives mostly they worked great. Of the four drives I bought affected with the recent problem, 2.5 have failed. I don't think it was Fujitsu's fault. That said, Fujitsu has done a miserable job of owning up to the problem once they realized what had happened. The other day I heard that a local school had to return 40% of all drives ordered from this drive series.

Compaq Deskpro EN's have these little buggers

[fialar]

They were failing in record numbers here where I work. I was one of the victims but fortunately, I backed everything up before it died.

Fujitsu released a firmware upgrade for them, but it didn't work and the drives failed anyway.

We had a lot of pissed off people with lost work, but there's not a lot you can do about that.

Now the newly-outsourced IT dept. here wants to switch us all to IBM Netvistas.

Um

[runderwo]

I hear people swearing allegence to a particular brand of consumer Hard Drive all the time, but I don't buy it.

Sure, but if you'd actually read the parent post instead of having an emotional response to it, you'd see that he wasn't swearing allegience to a particular brand at all! The point was simply that his new (large) drive failed, while his old (small) drive has been working for years.

I would have to say that he is lucky though; those particular Maxtor drives (850 meg to 1.6 gig) are extremely failure prone.

You can't make blanket statements about one brand versus another, but you can take past data into consideration when buying a new drive. Some manufacturers have pretty consistent failure statistics (WDC, Seagate). Others produce good drives most of the time, but have bad spells from time to time that alienate a lot of customers (Maxtor, IBM, Fuji).

All it really comes down to is the level of honesty and support that you get from the company you buy from. IBM and Fuji show an astounding lack of good faith when it comes to dealing with quality problems. Maxtor, WDC, Seagate not only go out of their way to bring problems to their customers' attention, they also have advance RMA policies, even for OEM drives in the case of Maxtor and WDC, to get you back on your feet ASAP.

Cheap RAID in five steps!!!

[Newer Guy]

1. Buy a Promise PCI HD controller. The ATA 100 one is available everywhere for 49 bucks. Maxtor sells these branded as their own too. 2. On the way home, stop at Radio Shack and buy two 120 ohm resistors. 3. Do a Google search to get the instructions on how to convert it to a RAID controller. If you are able to solder 4 connections, you can do this mod. in 30 minutes. It's beyond easy. 4. Get yourself a HD the same size or bigger then the one you want to mirror. Brand doesn't matter. I bought a Maxtor 60 gig for 99 bucks that had a coupon inside it for a $50.00 rebate to get their controller card free. 5. You're done. Okay it was only four steps. The ATA 66 Promise card can also be modded and doing so is even simpler then the ATA 100 one. I've done many of both and never had a single one go bad. The ATA66 card can be found as cheap as 20 bucks.

Re:Trends (bad correlation)

[gosand]

As drives have gotten smaller/increased data density, they've become increasingly unreliable. I'm pretty sure this coincides with the new 1 year warranties (versus the older 3 year standard warranties).

I think this is a bad correlation. At the same time drives are getting more dense and/or smaller, more people are using them. The use of PCs over the last 4 years has greatly increased. There are more reasons to need more drive space, I have a 30 GB and a 120GB. I wouldn't have needed those 4 years ago, but now they are about 60% full. Hard drives are used a little harder now. People are modding cases, OCing their systems, and generally getting more out of the PC than they have in years past. I had a 4 GB drive fail 3 weeks after the 3 year warranty expired. Now you would be hard pressed to find a 4 GB drive. I think that manufacturers realized that 3 years is a LONG time in the tech industry. Compare the number of drives sold 5 years ago to the number sold today.

I don't know if there is an increase in unreliability of hard drives over the last few years, but I know that instead of 1 computer I now have about 5 running at home. Of course, all this applies until one of my drives crashes, then I'll be convinced that hard drive manufacturers don't give a damn about quality anymore. :-)

The funky jumper settings screwed me

[swb]

I have a 6 gig Fujitsu that has the goofiest jumpter settings. In one mode you set one jumper across a pair of pins like any other drive, in another mode you set two jumpers, one in the normal fashion and another *horizontally* across one of the pins used for the other mode, in a manner normally used to park jumpers on drives that have all jumpers open for some modes.

I neglected to do this properly -- I couldn't believe it worked that way -- when adding it as a slave drive and it corrupted the master drive, sinking my system.

It's the only drive I've ever seen that used jumper settings in this manner. I haven't used the drive much, so it hasn't failed...yet.

Cost

[wytcld]

Thanks for correction about RAID-1. But wouldn't the extra cost only be for the extra drive, presuming the CPU is lightly enough loaded that Linux kernel RAID shouldn't much affect the responsiveness of the system, as is the case with most home office systems? 60-gig 7200 rpm drives are around 80 bucks - or 100 if you don't shop around. So with two of 'em you're at 160-200, less than a single 10 gig drive cost not too long ago. And at costs this cheap, how could they not be intended to be disposable?

Re:Cost

[jred]

I don't consider a hard drive for $80 to be disposable. An $80 car, yes, not a hard drive. A $20-$40 drive would be disposable, but just barely.

I guess it's been said enough times.

[cecil36]

If you want a decent hard drive, buy a Western Digital. I've used Maxtor and Western Digtal drives in the systems that I've bought or built, and between the two drives, Western Digital is more reliable. The Maxtor drives I've had in my lifetime started to fail after a couple years, but my WD drives are still running just fine. My first Maxtor drive that died on me I ended up taking apart just so I could see the interior of the drive. While looking for all the hardware to rebuild an old Athlon 750 system, I stumbled across this drive, and I plan on hanging on to it until I decide to quit being a g**k (censored because my girlfriend doesn't want me saying it, along with "n*rd").

Number from a hardware reseller

[Taurim]

Here are the failure rates from a big French hardware reseller (LDLC) : IDE 7200 rpm 20 Gb : Seagate : 1.3% (448) Western Digital : 8.8% (1506) IDE 7200 rpm 40 Gb : Seagate : 1.6% (7643) Maxtor : 1.9% (8052) IBM 120GXP : 3.1% (4790) Western Digital : 7.2% (1726) IBM 60 GXP : 22.9% (1068) !!!!!! IDE 7200 rpm 60 Gb : Seagate : 0.7% (284) IBM 120 GXP : 2.5% (722) Maxtor : 2.5% (1791) Western Digital : 8.6% (490) IBM 60 GXP : 16.1% (932) !!!!!! IDE 7200 rpm 80 Gb : Seagate : 2.4% (1248) IBM 120 GXP : 2.8% (2131) Western Digital : 3.1% (1676) Maxtor : 3.3% (2060) IDE 7200 rpm 120 Gb : Western Digital Special Edition : 3.0% (132) IBM 120 GXP : 3.1% (708) Western Digital 100 Go : 4.3% (470) Western Digital : 5% (120)

Single Drive RAID

[superid]

Highpoint chipsets are cheap, I've got two motherboards with them built in (and unused)

In a heartbeat I would buy a 40 GB drive that was actually internally mirrored 40's. Yes, I will pay a significant premium for integrity.

So, manufacturers, build me a single drive form factor hard drive, with 1 ide connector that is in fact a RAID 1 array!

Re:Single Drive RAID

[dsb3]

> So, manufacturers, build me a single drive form factor hard drive, with 1 ide connector that is in fact a RAID 1 array!

And, when one drive fails, you want to replace both together instead of just replacing the bad one and remirroring??

Sony has issued a recall

[swv3752]

http://www.ita.sel.sony.com/support/news/hdd.html

Yes, I've noticed problems with Fujitsu drives

[Faramir]

I cannot give you any kind of meaningful data, except this: in the last three years, in environments that are probably equal mixtures of Fujitsu, IBM, and Maxtor (in terms of IDE drives), I've seen far more Fujitsu drives die than anything else. At my current company, I've had 75% of my Fujitsu drives die, without a single other failure.

The link for the modifications...

[Newer Guy]

http://www.tweakhardware.com/guide/raid100/

Trolled...personal mod points?

[mmol_6453]

Sorry, but the parent post was a troller's comment. It said precisely the opposite of what most of the people here will tell you they've experienced.

I've personally (Though you'll have to take my word for it) lost several WD drives; the last one made a frequent clicking sound for about a week before it went kaput.

I'm told by several of my friends (who build and sell computers) that IBM drives have usually been untouchable, while one of them said he'd shoot himself before putting a WD drive in a customer's machine.

Minor lesson:

How are you going to tell the difference between a troller's comment and a factual (or at least honest) comment? You really can't, without having been around to absorb the general opinion.

When I don't know much about a subject, I usually depend on the anecdotes and hyperlinks. Up to this point, even this post, by that criteria, is highly suspect.

Wouldn't it be neat if Slashcode had a sort of Bayesian (sp?) filter that tried to predict whether a particular user would reject or accept a post based on his past reactions?

While in training, it would merely tell you what it would have predicted, while you train it by performing personal moderations on comments. (personal mod points would be granted depending on the availability of CPU time to run the Bayesian analysis.) These personal moderations could even be made visible to friends and fans.

Granted, it's the worst form of censorship: I'm willing to bet that a lot of people would simply reject anything they don't agree with, and thus silence the opinions of any dissenters to their private world. It also has a large possibility for abuse, if you consider Slashdot admins to fulfill the "big brother" role.

But people who reject dissenting opinions automatically probably wouldn't listen to the opinions even if they did have to see them. So I think it would still be a nice, and useful, feature.

I'm going to put this in my personal journal, for those of you interested in following this topic.

Keeping your drives cool

[Virtex]

the easiest is to put the drive as far down as you can get in the case

Interesting idea. I hadn't thought of that. For my computer, I mount my 3.5" hard drives in removable 5.25" drive bays. The bays are made of aluminium to help dissipate the heat, and they have a small fan in the back to help circulate the air away from the drive. Of course, the only 5.25" drive bays in my case are at the top of the machine.

I originally bought the drive bays years ago because I noticed how much heat there was between my two drives. Given that there was only a couple millimeters of space between them, the heat had a difficult time escaping. I wanted to put more space between the drives, but my only 3.5" bays were taken by the hard drives and a floppy drive.

The drive bays cost me about $50 each (I bought two), which seemed expensive, but as I think about it, I've never had a hard drive fail on me. These days, you can get similar drive bays for $10-$20 each.

Record numbers, compared with what other cases?

[Observer]

It's a bit difficult to do comparisons when both the manufacturers and their major purchasers have their own reasons to be, er, less than transparent about the actual figures.

Fujitsu do seem to be attracting a lot of attention recently, though. And the place I work has had 2 division-wide replacement programs for Fujitsu HDs for their Compaq ENs in the last couple of years - for most cases a precaution, with data copied successfully 1 for 1 to the replacement device.

Just a single data point, of course.

Quality control changing...

[Andy Dodd]

I have to agree... Manufacturer QC yo-yos so much these days that by the time one manufacturer has an established rep, their QC changes and people buy loads of drives they think are reliable that die in droves.

IBM used to have a stellar reputation for drives, and I would never buy from any other vendor. Now, IBM's name is mud and no one who has a clue is going to buy a Deathstar.

Likewise, it took me 4+ years before I heard enough testimony indicating that Maxtor and WD had shaped up their act before I would even think of buying another drive from either of them. (I now have a Maxtor 80G drive, and I'm pratying.)

Back when I was in high school administering my school's network ('97-98 is when I did most of my admin work there, as we got our 'net connection in erly '97), we had Western Digital 540M drives failing on a regular basis. Out of 20-30 PCs with WD 540s, we had an average of one failure a month. Our web server used 2 gig Maxtors. Over the course of 2 years, we had two of these units fail.

Supposedly WD and Maxtor are much better now... I hope so. I have a Maxtor and may be getting another 120G unit.

Seagate has always had a stellar reputation, especially their SCSI drives.

Re:Trends (bad correlation)

[Matey-O]

You're also neglecting the appliances: ipods, Xboxes, PVRs all have Harddisks too.

Between my laptop, fileserver and workhorse, plus the other oddball products, I've got 7 drives a spinnin. and three or four in a box somewhere that were too small to continue using.

That said, i've NEVER had a drive fail that I didn't addicently cause myself. I've had a few with niosy bearings, but have found that as long as I didn't power cycle the machine they were in, they continued to run faultlessly.

Link To October's Warranty Adjustment?

[limekiller4]

I wonder if the relatively recent (October) change in hard drive warranties was a pre-emptive move on the manufacturers part realizing that they could ship these just-good-enough drives with severe early mortality rates and get away with it.

I don't see Fujitsu in the lineup but I know very little about hard drives, so far all I know one of the manufacturers Tom's Hardware reviews actually applies to them.

for onboard promise controllers

[squarefish]

http://www.stormpages.com/crazyape/mbraid.html

Wouldn't this be in breach of

[myowntrueself]

the DMCA?????

Q. How do you spot a troll?
A. With very *very* large knives.

Compaq & drivers

[toby360]

With Compaq aka The new HP computers in our office, which almost all had fujitsu or maxtor drives, the fujitsu drives have almost all died out from our 2001 computer batch. Computers prior to 2001 seem to be far more reliable. I would say 80% of our 2001 fujitsu's have needed the hard drive replaced.
Here is a lawfirm with a class action lawsuit regarding several models:
The Fujitsu hard disk drive model numbers that are a subject of this litigation include, among others, MPG3204AT; MPG3307AH; MPG3102AT; and MPG3409AH. Continue to monitor this page for the addition of other model numbers.

Two good Fujitsu's here, and six failed IBMs

[Dr. Ion]

I've been pretty happy with my sample of *two* Fujitsu MPG3409AT drives. They're silent, run cool, and serve up 40GB each without hassle for about three years so far.

My beef is with the IBM Deathstar GXP drives.. the 60 and 75GB drives last 1 to 6 months, and then get read errors. I have one drive that has been RMA'd four times. I don't dare install the replacement drive.

FUJITSU only HD I've ever had fail...

[blakespot]

I've been using computers for 20 years now and the only HD of my own, and I've had a lot of machines, that has ever failed was a 512MB, SCSI Fujitsu purchased in 1995. It failed a few months after I purchased it--they promptly replaced it for me. The replacement got a good bit of use and now sits with pretty light use in an enclosure tied to my Mac Plus (it was originally used on my 486-66 used at the time to run NeXTSTEP for Intel v3.2).

blakespot

Never say "never", but.....

[King_TJ]

I think there's a grain of truth to some of the complaints.

You have to look at things over the LONG term though.... not just an isolated batch of complaints from around the same time period.

Honestly, I know relatively few people with complaints about Maxtor drives. Like everyone else, they occasionally released a bad batch. Still, you'll consistently find people relatively satisfied with their products over the years.

Fujitsu, on the other hand, I never had a good feeling about. I heard some good things about their rather pricy SCSI drives, back when they competed with Micropolis and built drives that took 2 full-height 5.25" drive bays. Whenever I looked at their IDE drives though, I just got the idea they weren't striving for "top quality". They cut corners on the little things, like the IDE connector itself. (Instead of surrounding the pins with a plastic guide, they typically went without - making it harder to plug in the ribbon cable properly.)

IBM always had a great hard drive reputation, until they trashed it with the horrible Deskstar issues. It's going to take a lot for them to dig back out of that hole.

Western Digital is probably the one drive vendor that's that hardest to pin down. I've generally liked their drives a lot - yet I can't deny they have a lot of drive failures. From using their products over 10+ years now (in the workplace and at home), I get the idea they generally have a lot of RMA "out of the box". If you get a good drive that doesn't make any weird noises, it'll probably be a good drive for years to come. If it seems a bit "flaky" when you first start using it though, look out. It'll probably be a dud in the long run.

On negative product reviews, etc.

[Corvaith]

That they're there isn't necessarily a big deal, but quantity can be telling. And how many people you see saying good things. And so on.

This is true of all sorts of things. The first computer that was completely mine was purchased from a company that eventually went belly-up, had all sorts of shady business practices, and tons of bad things said about them on the 'net, but I didn't find that out until after I bought it.

Let the buyer be informed. It's always a good idea to read up before making major purchases.

Re:What utility software?

[morgajel]

I've had 3 10gigs, a couple of 5 gigs, a 6 gig a 2 gig and a 20 gig all RMA'd back to WD.

don't make blanket statements. wd is all I buy, and they're not immune either.

I wish I had mod points!

[Dman33]

I would mod you up +1 Funny.

Nice one!

No Surprise - We Get what We Pay For

[GroundBounce]

The downward price pressure on hard drives has been extreme in the last few years, and now wer're paying the other price - in reliability.

I worked for several years for a company which designed and manufactured ICs for hard drives (I worked on read channels, but the company made other chips as well, such as preamps and servo controllers). There has always been competition and downward price pressure in this market, but early on, both the ASPs and the product lifetimes were somewhat reasonable.

Over the last 5-10 years, things have changed a lot. The lifetime of a drive product is very short (sometimes as short as 6 months), and each new generation is so much faster and denser than the last that many of the critical components require a from-the-ground-up redesign with very little being borrowed from the previous generation. This, combined with lower ASPs than ever, have made it more and more difficult to be highly profitable as maker of chips for hard drives. Companies that are successful have engineers working very long hours to do it. Several companies have left the market entirely, or have taken on other product lines as well

And this is just the ICs. I'm sure manufacturers of other drive components (platters, heads, etc.) have seen similar erosion of product lifetimes and ASPs.

The end result of all of this it that there will be an inevitable hit in quality and reliability. There's really no other choice. When customers are once again willing to pay $200-$300 for a current technology drive, you will see the quality go backup. Even today, SCSI drives, which are generally more expensive then IDE drives are also more reliable, as many posters have pointed out.

IBM and "SMART"

[Jon Abbott]

IBM had a pretty crazy rate of failure and was telling people to turn off smart mode.

Where and when was this said? I have been keeping track of the recent Deathstar fiasco and have never heard IBM mention anything to this effect. Is this a real workaround?

Re:IBM and "SMART"

[DuBois]

...turn off smart mode

Is this a real workaround?

Well, I own two 75GXP 30GB IBM drives. One came built into my Mac G4/450MP, and the other I bought from my hole-in-the-wall PCclone supplier. I needed more space for video, so replaced the Apple-supplied drive with a WD 80GB drive, but during the copy from the IBM to the WD, a few files had too many errors to copy (fortunately, not files I needed).

Smart mode wasn't of much use back in MacOS 9, so it didn't tell me anything about the impending doom of my data.

I took the drive out of the Mac and put in a PCclone and ran IBM's Drive Fitness Test which showed a boatload of errors. Since I didn't care about the data on the drive anymore (it had all been copied to the WD), I ran the low level format utility from the DFT and erased everything down to the bare glass. It came back perfect. I have since had the drive running for about six months in the PCclone on FreeBSD without problems. I'm about to upgrade this machine to 4.7 (or maybe 5.0, if it comes out on November 20 as promised by the FreeBSD webpage), so before I do that, I'll check it again with DFT. I suspect it'll come back fine.

The other 75GXP (the one I purchased from my PCclone store) hasn't given any problems.

They're both made in Hungary. I do use one of those drive cooling fan panels underneath the drive in the PCclone, since I've heard the rumor that the 75GXP failures are often heat-related.

Bottom line: turning off smart mode is NOT a solution. Reformatting to the bare glass IS a solution.

Finally I have found others with this problem!!!

[compugeek007]

The company I worked at purchased a bundle of 130 PC's about 14 months ago (I didn't work there then). They were fully propritary (yeck), major brand name, and "Legacy Free" (Everything integrated and all USB, no serial, ps2, parallel, or PCI slots.) The moron who bought them didn't buy an extended warranty either (smart eh? Buy a bunch of PC's you can't possibly service and do not buy any warranty with them. Needless to say that guy is long gone.) Luckily I argued one out of the company before and bought an extended warranty before the orignal expired.

Anyhow, almost 60% (probably more, the PC's are spread all over the country) have had Fujitsu drive failures. These are all domestic (U.S.) Pc's with Fujitsu 10 GB drives. I am so glad that this is a real problem - I feel like sending this link to the pain in the ass support guys that have made my life hell everytime a PC dies and needs a hard drive replaced.

Gramma Issues

[vandan]

If people can't write complete sentences.
Then I don't know.
Which is more important.
Whether hard drives don't work.
Or whether our language skills don't allow.
Us to.
Talk about it.
In some previously agreed.
Formal manner.

And Seagate still give 5yr warranties

[myowntrueself]

I just replaced a drive for a client, not due to failure but due to need for increased storage.

When I looked around for drives only Seagate were prepared to give 5yr warranties on their products.

If manufacturers don't stand by their products, how do they expect consumers to.

Interestingly, this axiom *should* apply to most commercially available software which usually has a complete disclaimer of incompetence, warranty etc in the license agreement.

How long before we see such disclaimers on hardware?

RAID only protects against media failure

[anonymous cupboard]

First of all as you suggest, RAID will faithfully write whatever screwed up data that the host system tells it to. It only protects against a bad drive.

The problem is if the PSU goes on the blink, this can trash multiple drives. If you are unlucky, it can trash your entire RAID-5 set. The PSU on many machines is selected purely on a cost basis and even if can deliver the watts, it may send spurious voltages in the process of self-destruction.

The only way is a second system and to synchronize, i.e., with rsync, the data on each system. Regrettably, disks have expanded beyond the capacity of all but the most expensive tape drives.

Re:Fujitsu makes hard drives?

[jedidiah]

Microsoft is actually an anomaly in this regard. Prior to them, a "big name" in operating systems typically mean a good product. It implied a certain level of engineering. Smaller isn't always better either.

The better comparison would be between Apple/Sun/IBM, Microsoft and Linux/FreeBSD.

On one side, you have SERIOUS big names. On the other, you have generica. The genuine big names have good engineering practices, solid product and highly effective support organizations. The generica beat all other players when it comes to price. The players in the middle resolve to "why bother".

Why bother with any brand name product unless it is some "big name" that you expect to get something more out of?

Re:huh...

[jedidiah]

I too can say that about drives in general. However, I stick to the major players and avoid the bleeding edge. I suspect that many people avoid problems simply by avoiding the situation where they are paying to be beta testers.

All of my personal harddrives have lasted at least 5 years before being taken out of service.

Fujitsu > WD

[CowbertPrime]

I have a Fujitsu 6gb laptop drive that has been running great since 1998. I also have a Fujitsu 3.5" 20gb that has been running equally well since 2000. My brand spankin new WD-800BB 80GB Western digital drive failed in 5 months after purchase (and the only reason i got it was because of their $75 rebate on a drive that was almost $200 in Feburary 2002).

Western Digital drives are excellent.

[Futurepower(R)]

No problems here with Western Digital drives, which have been perfect in recent years.

Fujitsu's customer service is great, BUT

[krinsh]

the hard drives still suck. I had this same sort of failure problem with them 3-4 years ago when we had put in a bunch of new Compaq workstations. They were very nice about it even though they made us go back to Compaq, who then promptly sent us Western Digital drives and did not even bother to ask for the Fujitsu ones back.

Re:Cooling Bays != Cooler Drives

[Jucius Maximus]

"I've also cooked 2 Maxtors to death in a Lian-Li removable IDE Rack for similar reasons. I will never again put any 7200 rpm+ drive in a single enclosure with 40mm fans - my data is too important. "

Interesting ... I stopped using a Lian Li removable rank (with 2 x 40 mm fans) because the HDD kept getting messed up!! I would lose entire partitions and the drive refused to be reliably accessible.

When I put back directly on the IDE channel (using the same cabling btw) everything was fine again.)

But I am going to take a look into the S.M.A.R.T. temperature monitoring so I can see if the temperatures in the bay are really higher than the temperatures on the regularly mounted drive. Thanks for the heads up on that.

We had a problems with IBM Drives

[puppetman]

They got excellent performance reviews on Storage Review but no reviewer can travel forward in time a few years to see how the drives hold up.

The result is that we have IBM drives in all our production database servers, web servers, and app servers. And I picked the hardware. Who said no-one ever got fired for buying IBM? Guess it's true, as I didn't lose my job over it.

AnandTech had an interesting article here on the drives, and why they went bad (poor microcode that handles the interval between tracks as the drives heat up).

It has happened to me twice

[IRNI]

One of our developers machines locked up. He rebooted and there was a bunch of garbled text where it was trying to detect the IDE drives. Nothing would fix it. I tried the drive in other machines and nothing got it back up. So the guy had to recode everything he had been working on for the last few hours after I got a new drive installed for him.

Then one day I got called out to a client site. They had a Linux server I installed for them. They said it was locking up. I told them to power it off and bring it back up. They did that. I SSH'd into the machine and it locked up about 30 seconds in. I told them to reboot it again and it wouldn't come up. They said "it is printing some really weird stuff on the screen." I dropped everything and headed over there. It was the same exact problem. I couldn't get the data off the drive. No bios would accept it. I noticed it was the exact same model. It was a 9.1 gig IDE drive by Fujitsu.

Needless to say, as a person with a lot of influence on what my clients buy, they never buy Fujitsu ANYTHING anymore.

Yes - Fujitsu 20Gb is bad news

[yem]

I've had FIVE of these die in the last year. Five out of about Eight. Model number is MPG3204AT. Someone in my department had his drive die on Monday.

Re:Trends (bad correlation)

[swordboy]

I think this is a bad correlation. At the same time drives are getting more dense and/or smaller, more people are using them.

I work in tech support for a company where the population has been largely fixed (so it doesn't matter if the rest of the world is using more than usual - I have my own data). I have LOTS of hard drives going through my hands so I'm familiar with failure rates. They have been increasing. Certainly, there are lots more drives out there, but they are failing at a higher rate.

In years past, it was easier to deal with tech support if you could let the drive "speak" to the technician on the other end of the phone. Usually, the techs were button monkeys that didn't realize that *I already knew* the drive was bad and needed to be replaced. So in the end, I'd usually just power up the drive and give it a few good whacks on the counter. Then I'd call up support and put the phone up to the drive. This reduced call times to only a couple minutes rather than the typical 20 - 30 minutes that it took the monkey to run through the flow chart.

Me: Here THAT? It's broken!
Tech: Your shipping address, sir?

Today's drives don't take much whacking as they are much more delicate. This is also evident by IBM's new Thinkpad Shock Absorber (page 2, feature #5). With my old Thinkpad, I once (forgive me...) had a near car accident while it was powered up. The damn thing flew across the car and smacked into the dash with nary a problem. It still works today.

Tip: for the new one year warranty's, just buy two drives and mirror them. Whack one at 10 months and the next at 11.

Cheetos,

swordboy

Re:Trends (bad correlation)

[jafuser]

I've never had a hard drive fail on me at all.. What is the "straw that breaks the camel's back" when it comes to drive failure, and how often is it reasonably unrecoverable?

I've never met someone who lost the drive instantly and permanently without using some simple technique to temporarily revive the drive long enough to get the data off.

Usually it's just an old drive that won't spin up anymore. I tell them to get a replacment drive, install it, then power up, tap the "stuck" drive on the side gently with a rubber hammer (or a regular hammer wrapped in cloth), then copy your data off of it to the new drive.

Re:Trends (bad correlation)

[WowTIP]

I'm personally hoping that a more expensive drive is released that is more reliable. I'm willing to pay extra for a critical component in my computer needs.

Yes, me too.

Anyone know which is considered the most reliable drive manufacturer out there? Is there any brand that is famous for not crashing?

fujitsu hard drives

[y0bhgu0d]

my friends and I all have a 100% failure rate with 10 and 20gb fujitsu hard drives... i got about 4 of them for free from an auction (a computer training business went bottom-up, and i was working with the auctioneers about technical stuff), and none of them lasted more than a week. these were brand new drives. since then, i have had friends that have acquired these same model drives through similar means, and none of them have worked for more than a month or so.

the problem that would appear:
things would start to get unstable... acting flakey; so you run scandisk. bad sectors start at the beginning of the drive and continue for at least 5 gigs. not something i want to sit through... i have a friend that did a full scandisk, and it took ~6hrs for a 10gb drive. when it was through, he had ~4gb of usable disk space left.

just my experience...

Exec. Summary: Fujitsu's Been Good to Me So Far

[ewhac]

It was because of recurring problems with IBM drives that I ended up replacing them with Fujitsus about six months ago.

My system is all SCSI, all the time. As a result, I end up paying in the neighborhood of $200 for an 18G drive. With prices like that, failure is simply not acceptable. Some people say that all hard drives are crud and are going to fail, so one should simply plan for it. Well, then why are the hard drives in my 12-year-old Amiga still working fine?

After enduring my most spectacular failure to date, I resolved to change drive brands. A couple of years prior, I upgraded the drive in my laptop computer. The first drive I tried was an IBM Travelstar, and it made the most gawd-awful racket. I could hear the thing two rooms away over the fans in my main rig. So I sent it back and took a Fujitsu instead. It's been perfectly quiet and reliable ever since.

After this happy experience, I decided to put a couple of Fujitsu MAN3184MP SCSI drives in my main rig. So far, they have given me no trouble at all.

I can't imagine what the heck's going on in the hard drive industry to cause so many failures. I can only hope one of the manufacturers will spill the beans at some point.

Schwab

Re:Trends (bad correlation)

[gosand]

Certainly, there are lots more drives out there, but they are failing at a higher rate.

I think part of it is that there is a bigger demand for drives, so the companies that make them are pushing the technology. It does seem that there are bad "batches" or "runs" of drives. In the rush to get to market, I can see where quality and perhaps durability has slipped. But several years ago, there didn't seem to be this big of a push for bigger drives. I have a feeling that the quest for larger drives will slow a bit, because currently we are running out of ways to fill them up. Not everyone needs a 100 GB drive. When the average user could fill 10GB, demand for drives seemed to jump. For the average user, a 20 or 30GB drive is PLENTY of space, at least right now. If you do video editing, or keep a digital music collection, or run a server of some kind, you need more space.

At least by pushing the speed/density frontier, manufacturers are advancing the technology. Hopefully that will drive for more stable storage technology as well. I am pretty sure that the platter style hard drives we currently have has its days numbered, but they aren't going to find the replacement until they push the limits.

Yes, they did

[gallir]

Last week I've lost my disk, after failing for months. I thought ReiserFS was the guilty. When the disk was unrecoverable, I checked it with badblocks -f, it was the hardware.

Sorry Hans, and respects to your mother :-)

I had a same problem years ago with fujitsu disks

[wtarreau]

Several years ago, I worked for a little company which sold PCs for professionals and individuals. During several months, we have sold fujitsu drives ranging (1.6, 2.1, 3.2, 4.3 and 6.4 GB). These models were based on exactly the same hardware, and we had lots of drives failing because they suddenly confused their type ! 1.6GB often became 2.1, and 3.2 or 4.3 became 6.4 GB. Of course, there weren't enough platters to make this work, so not only our customers lost their data, but we had to send the disk back to warranty. It was a very embarrassing situation because the hardware was OK, but we couldn't get the data back. The disk started, detected the error, then stopped. Sometimes, waiting several days allowed the disk to recover its original model. I always wondered if they stored the model on the medium itself instead of an eeprom.

That was a bad experience unfortunately, because except for this problem, these disks were relatively quiet and really fast !

Willy

I'll put it this way....

[NerveGas]

After maintaining a good number of web servers and office machines for over three years, I have a nice pile of dead hard drives. About 30% of the machines got Fujitsu hard drives, yet every single dead drive but one is a Fujitsy. I stopped bothering with replacing them under warranty, because a new Fujitsu would just mean another dead hard drive in less than a year.

steve

Oven chips ;)

[Epsillon]

OK, I have a Fujitsu MPG-3102AT date coded 2001-03, right smack bang where the problem occurred. It's also dead, an ex-drive, if it wasn't screwed to the drive-bay it'd be pushing up the daisies... The problem is reported to be with the controller chip, one Cirrus Logic's CL-SH8671 batch coded 450E on mine. I contacted Fujitsu (being unfotunate enough to have purchased mine from a computer fair, silly sod) and found that they DON'T hono(u)r the warranty for end users! The b@stards! Last time I buy a Fujitsu drive. The problem with the chip is that Cirrus, in their infinite wisdom, changed the material they use to encapsulate this huge QFP IC without telling anyone (so Fujitsu's story goes) and subsequently the reflow ovens in the SMD process were not reprofiled to take into account the new properties of the material they used. So the *chip* ended up either cooked to the point that ingress of moisture became possible during heat-up/cool down cycles or didn't reflow properly so ended up with dry joints on the legs because the new material leeched the heat away from the joints. I tried reflowing mine on an SMD rework station and no joy so I suspect the former. Can't believe they say there isn't a problem, especially when they're rumo(u)red to be currenly in dispute with CL over this batch of ICs which they claim were sub-standard. If they were so sub-standard, how di they get through QC, humm? IS ther a QC dept.? Draw your own conclusions!

Re:What to buy then?

[DuBois]

Seagate.

Their IDE drives are *quiet*.

Their SCSI drives are *reliable*.

Re:huh...

[Sylver Dragon]

The reason I mentioned it, is that a large part of my job is system building and integration, between 3 and 5 computers a week on average. That's a lot of hard drives going through. Mind you, we try and keep off the bleeding edge of technology, in our industry its a bad idea, but we have enough volume that I think I've seen a good sample size.

In a few hundred drives...

[Hyped01]

Our noted failure rates, highest to lowest are:

Fujitsu

Seagate

Western Digital

Samsung

HP

Maxtor / IBM (tie).

The IBM and Seagate drives (and a Western Digital) all run WebBinaries XXX Thumbnailed Newsgroups and WhoreBoyz so they get pounded (no pun intended) pretty regularly.

We dont run any IBM SCSI or IDE over 40GB (36GB & 40GB largest respectively), ALL the Seagates seem to suffer bearing or motor burnout - so far 37% of them died all while running in well cooled $6,000 Tricord ES Series filtered dual redundant power supplied drive arrays (ie: they werent abused heat, air quality or power wise).

The IBM, btw, was gotten operational again with a little WD40 in the bearing hole under the label (woulda used lithium grease or something better, but no time, and the drive is very old so other than backing up a whopping 2GB of data, I really dont care about it).

The Maxtor was a little newer.

The drives aren't the problem

[MrResistor]

If you've had failures that consistently the problem is not the drives. That's one of the first lessons I learned as a professional troubleshooter: If the replacement part fails, the part you replaced was not the problem. That doesn't mean the part you replaced wasn't bad, but it does mean that you're going to have to keep replacing that part until you take the time to get to the root of the problem, which is something else.

Power supplies are the most often overlooked problem causer, but it could easily your RAID controller or a poorly made data cable. It could also be that your drives are simply overheating.

I've built several computers for various friends, family, and small businesses over the last 6 years and I've used Deskstars exclusively for the last 3 years. I follow a few basic rules when mounting them: (1) always leave an empty space between the hard drive and any other device for air flow, and (2) never stuff excess cable or other crap into that empty space. I have never had a hard drive fail in any of those systems.

If your case doesn't have space for you to do that, then your case is too small. It's that simple. Having parts fail due to inadequate airflow is much more expensive than getting the right case.

As for second hand drives, that's always going to be a dodgy issue. The manufacturers do test them, and they don't send them out if they don't pass their tests. The problem is, though, that none of the drive test utilities really stress the drive enough, IMO.

I torture-test SCSI RAIDs for a living, and we've found that the standard test suites were letting too many drives through that couldn't hold up to the demands our customers placed on them (we're talking top-end digital video production), so we had to develope our own tests. I typically push a drive/LUN/controller card at the threshhold of it's advertised performance for at least 2 days before I call it good. The funny thing is that spending more time doing a more serious test actually reduces my work load, since I very rarely get any of those drives back, and those I do get back are usually because the customer was moving their equipment and wanted to have spare drives on hand just in case.

Not everyone has $15,000 realtime MPEG encoder cards lying around to hammer their drives with, but I'm sure there are other ways to achieve similar results.

Re:Trends (bad correlation)

[isorox]

had a near car accident while it was powered up

Perhaps if you werent surfing slashdot when you were driving you wouldnt have had the accident

Re:Trends (bad correlation)

[G-funk]

Yet more truth to the fact that "if it ain't broken, don't fix it, and if it is, hit it with a hammer".

Poll?

[Openadvocate]

Isn't this something you could put in a poll?
-----------
(o) I own a Fujitsu and no problems.
(o) Fujitsu 0wned my harddrive
(o) I don't own a Fujitsu and no problems.
(o) I don't own a Fujitsu but many problems.
(o) I want a Fujitsu so I can get problems.


etc etc

Re:Trends (bad correlation)

[perky]

With my old Thinkpad, I once (forgive me...) had a near car accident while it was powered up. The damn thing flew across the car and smacked into the dash with nary a problem. It still works today.


Interesting you should say that. I worked for IBM a few years ago and on the way back from a conference wrote off my car (ford fiesta) with 6 thinkpad 600's in the boot. My car suffered a direct sideways impact on the side of the boot from another car (don't ask); and then spun round and hit a traffic light post on the other side of the boot. All six worked perfectly afterwards, and I contrinued to use one daily for the following 6 months. Thank fuck.

I still think they are the best laptop you can buy and am about to pick one up second hand.

drives have a 100% failure rate; matter of time

[madbrain]

Every hard drive I have owned has failed at some point, it doesn't matter which brand. The difference has been how long it has taken for them to fail.

Over the last 12 years, I have owned in my various computers such brands of hard drives as Miniscribe, Seagate, Quantum, Maxtor, Western Digital, IBM, CDC. I probably owned about 25 drives.
You may not recognize all the names because some of the manufacturers are defunct. During about 4 years, I ran a BBS 24/7 and kept the drives running. I remember maxing out the capacity of the narrow SCSI card (7 devices).

I have not resold any of the hard drives, rather, I have just kept using them. All of them ended up dying, except the 3 I'm currently using, which are all less than 2 years old. Most of the drives failed between the third and the fifth year. Since they were nearly all SCSI drives that carried 5 year manufacturer warranties, they were eligible for free replacements, but of course by then the capacity of was ridiculously small.

Sheesh... Everyone listen: BACK UP.

[aussersterne]

Most tech veterans from the pre-Windows era will already know this but for everyone else:

Hard drives fail.

They always have, they always will. They're mechanical, and even better than that, they're magnetic. That's how it is. You should plan to have to replace your hard drive every 1-3 years at least, if not more often than that, depending on workload and conditions. That's how it's always been. For the people saying 'but... I've never had a hard drive fail...' -- you're lucky as hell, but someday your luck will run out.

Even backup solutions fail.

Removable storage is mechanical as well. There are a lot of variables. Years of experience taught me the following lessons:

• If you don't maintain removable storage backups (i.e. not just RAID-0 sitting inside your computer), you will lose data, sooner or later. Period. Some component on your power supply (even your high-end redundant power supply) will blow and take the entire system out with it. Or someone will acidentally bump into the case and cause a head crash. Or you'll have an earthquake. Or the sprinklers will come on because of a fault in the fire system. Or something. Trust me, it will happen.
  
• Never use a single magnetic backup solution as your only solution, or you'll come running to your one set of tapes the day your hard drive crashes only to find that the secretary has been holding notes to that particular file cabinet with a nice magnet... it never fails.
  
• Always keep at least one relatively recent set of backups offsite. Hell, put them in your desk at work, or in your locker at the gym, or in your safe deposit box, or at your mother-in-law's house. If your house or business burns down, you will be glad that you have your data saved somewhere.
  
• If someone else manages your server (i.e. you're paying big bucks to some hosting company), be sure you pay extra for good backup service. Trust me, they will happily lose your data through some dumb operator error, and they'll be happy to tell you that they won't be held liable when you're yelling at them on the phone.
  
• Make sure you back up often, but also keep a number of backups stretching back several weeks at least. Trust me, at some point you'll be glad you can go back and restore that 'old file' you deleted, or the old system, before the root kit that you haven't noticed for two weeks was installed.
  

This may all seem excessive for the "home" user, but if you're anything like me (these days I'm a writer/photographer), being a "home" user can often mean that your entire livelihood and household are tied up in your data.

As for me, myself, personally, right now I keep my nightlies on a rotating group of 14 8mm tapes using an Exabyte 8505XL drive. I use only data-grade tapes from major manufacturers. I run drive diagnostics often. I never use a tape through more than 10 passes. For my really important data, I also use 9.6GB DVD-RAM for redundancy. I would never consider working without backups, simply depending on this brand of hard drive or that one to not fail. I've lost too many hard drives over the years (ever seen a platter on an 10" drive crack and bits go flying everywhere, cracking the other platters and half the windows in the room?!) ever to be naive enough to trust one again.

Point of post: BACK UP YOUR DATA. Never think of a hard drive as anything other than short-term storage. Never think of any magnetic media as anything other than short-term storage, or you'll be crying sooner or later.

Re:Sheesh... Everyone listen: BACK UP.

[TheLink]

Yes backup backup backup, but the point was the excessive failure rates within a short time - within a year or so.

I don't know about you, but for me it would be unacceptable to have to do restores and deal with failures every few months. Or even yearly (barring exceptional circumstances). Backups are like insurance, it's bad not to have insurance, but it's almost as bad if you have to keep claiming.

Acceptable failure rates to me:
0-2% is ok.
3-10% is avoid personally.
Above that - recommend other people avoid too.

Heck if a consumer drive older than 3 years fails on me, I'd find it hard to find proof of purchase/warranty, and it's probably time to upgrade (if possible) anyway.

But 1 year warranties are way too short- we're being screwed. I think 2 years is tolerable - grumble grumble, open wallet, buy a newer better one.

Back to the backup: my personal data isn't as voluminous as yours, so I use CD-Rs for backups - CD-Rs allegedly have longer lifespan than RW media. Plus they are fairly cheap. But not going to go overboard and make a landfill a remote site ;).

Not going DVD till the DVD industry starts getting _real_ standards. And I personally am averse to tape - expensive, and I'm just biased against the idea of reusing tape. AND, I'm pretty sure I'd be able to easily get a drive that can read my CD-Rs even 5 years from now. Don't have to pay lots for a niche item. That is unless the RIAA/MPAA/TCPA screw things up even more.

IBM is number 2?

[autopr0n]

I think that page may be somewhat out of date. IBM doesn't even make hard drives after their disastrous 75gxp...

Reliable HDs

[delta407]

The problem with the IDE market is that everyone and their grandma uses IDE disks, so it doesn't matter which drive is more reliable; people buy whatever is cheap, and manufacturers have no economic reason to invest in making their drives last longer than 2-4 years.

If you want reliability, grab a nice high-end 10,000 RPM SCSI disk. Actually, pretty much any SCSI disk will do -- I've got a set of Seagate 'cudas that have been spinning for eight years and are still spinning as I write this. I also picked up a dual-133 IBM box at an auction a few years ago that's also running SCSI. It's a great little file server, and as far as I know it still has the original hard drives (from back in the day when dual-133 was cool).

Remember: the server market is designed for reliability. Grab an HPaq Proliant or some other box designed for the datacenter. Yes, they're far more expensive than workstation models with similar specifications, but they aren't going to break any time soon.

You get what you pay for.

So, anyone try to calculate an MTBF?

[Hollinger]

Anyone calculate a Mean Time Between Failure?

Fun with hard drives

[freeweed]

I guess you've never taken apart a hard drive before :)

The magnets in them (at least in any model I've ever opened up) are some of the strongest permanent magnets I've ever seen. And yes, they're located on the pivot of the drive head arm.

Also, the platters make way nicer decoration than boring cd-roms.

YES, 100% failure on a batch in Norway

[Graabein]

I've seen 100% failure rate on a batch of 10 and 20 gig Fujitsu drives installed in the summer of 2001. They started failing after 10 months or so. This was in Norway.

You might have purchased an older batch.

[TheLink]

And so not affected. Or the 10GB batch you got didn't use the affected chips.

Fujitsu blames Cirrus for close to 6 million hard drive failues:
http://www.theinquirer.net/?article=6074

Fujitsu's filings suggest Cirrus Logic had supplied it with defective chips in the summer 2000, and started to receive complaints about failures from May 2001. The filings confirm what our Eva discovered here: that a change in the constitution of the resin used to attach the controller chips caused a failure in up to 5.9 million drives.

---
Still, backup backup backup.

Nah. They're just screwing us.

[TheLink]

They're reducing warranty periods, raising prices etc.

It's just most people don't have good access to relevant data. I think many dealers would be "punished" if they published return/failure rates. Just so the master distributors or manufacturers don't get punished by reduced sales of stuff below expectations.

Right now all most see is measured performance vs _anecdotal_ reliability reports.

Given better access to information I'm sure many would pay a fair premium for reliability, though not an obscene SCSI premium (at SCSI prices one might as well mirror two or more drives from different manufacturers).

Someone here has posted failure rate stats allegedly from a french hardware reseller.

I find 1-2% failure rates acceptable in a major consumer level component.

I'd avoid stuff in the 5%-10% region (like WDs). Not totally crap for some, but still crap to me.

Anything above that region I almost feel a responsibility to tell other people to avoid them. It's practically civic responsibility to advise people to avoid buying something that is 20-100% likely to fail within a month or two.

So what if they provide good warranty or RMA. I want to spend time using the product not returning it.

Sure, manufacturers sometimes can't tell what impact a change in process would have on actual failure/return rates, same goes for totally new models - those MTBF figures are mostly bullshit, I only pay attention if they are unusually low. That said, they can't totally use that excuse for they do know how to make reliable drives e.g. most SCSI drives are ok.

But without proper statistics, there can't be market pressure to get manufacturers to produce reliable drives at lower prices.

The Target Market has Changed.

[GroundBounce]

There is a reason that the warranty periods have been reduced. Warranty periods are not being reduced simply to screw us, they're being reduced because the "us" has changed. The target market for IDE drives is now heavily consumer, where 10 years ago it was heavily business.

I have been away from the industry for around five years now, but having worked directly with drive designers, I know that at least at the design engineering level, they used to be genuinely concerned about reliability and MTBF. Now that disk drives have been pushed into the commodity consumer market space, manufacturers have to cut more corners than they used to to meet the low ASPs

What's happened, however, is that over the last 10 years or so, disk drives (and computers in general) have largely gone from being considered primarily business tools, to being considered primarily commodity consumer items. These two categories come with very different ASPs and expectations of reliablity.

One thing that you can't escape from is that if you target the same functionally identical product for $300, you can design and build in more reliability than if you target it for $150. You can do this by using higher quality components and sub-assemblies, by spending more engineering time evaluating, testing, and beating up the design, and by more thorough production testing.

Now, its not uncommon for manufacturers to offer multiple tiers of product quality for different markets. Unfortunately right now in disk drives, the two main choices are IDE (low end) and SCSI (high end); there isn't much middle ground. I agree that the SCSI premium is too high for most consumers to consider.

Perhaps if the true reliability data is known, as you suggest, then this might help create pressure on the manufacturers for a "middle-ground" tier of drives - perhaps a line of IDE drives that would be a little more expensive than current IDE prices, but less than SCSI prices with corresponding "middle-tier" reliability.

Re:The Target Market has Changed.

[TheLink]

I don't really think the consumer market has low standards, they're less forgiving if anything - if the average home user's VCR/TV/DVD player failed within a year of purchase, they wouldn't be happy. If it failed within 3 months, others are likely to hear of it. And the typical home user isn't going to say "BigBrand Model XXX product ID YYY was crap, the SCSI models are great tho", they're just going to tell everyone "I'm never going to buy BigBrand again". Techheads are more likely to forgive given evidence and explanations, whereas nontechs could hold an irrational grudge for years.

Consumers may not use reliability as a consideration for purchase, but they could be requiring/assuming reliability after the purchase.

Which is worse for manufacturers? So I really don't know if cutting corners that fine really makes as much sense for that market.

A few bucks buys you a fair bit of protective electronics or better electronics. Just tell the engineers you want X years at least, and listen a bit less to the nickel and dimers, who think everything runs quarterly - and who may be running off to a different company in a few quarters anyway (can't keep increasing those figures).

One of the companies I worked for used to sell modems and had shelves of returns. So they bundled in lightning protection (gas discharge tube + assembly costs probably USD2-4?) for free and it dropped to one or two every month. Definitely worth it.

Fujitsu really screwed up. They must have seen the signs earlier on in 2001. Yet they kept denying there was a problem. This sort of screw up is a management problem. They're still not even sorry - they're just pushing blame to others. So even tech fixes won't be convincing.

Re:Repeat after me: mirroring is not backup.

[kbielefe]

I agree with you completely about businesses (large and small). I disagree about the home user. Home user's have very different needs from a company. They have only a few files like their budget, the latest school paper they are working on, etc. that would be painful to lose. They don't need every revision of their work. And they don't work their hardware as hard. If their computer room burns down, the data on their drive is the least of their worries. Mirrors are an easy and appropriate solution for this kind of user.

WD, Quantium, IBM, DEC, Maxtor, & Seagate

[McFly69]

With Western Digital, I have the best luck. Every drive I own, never died under 5 years of age. I currently own about 12 drives, in functioning systems. Two of them are still under 500 meg. I really think they make a high quality drive. Back in 1998, I had one 1.6 gig drive that I coudl nto get working, and I really did not want to be without a drive. Tired using maxtor utils, fdisk /mbr, and even mounting under unix. Drive was fine but could nto boot to it. Use my 18" speaker, let it sit on it for about 5 minutes, formated it worked great... go figure.

Quantium - I delt with 4 of their drives before. One died after 4 years, 1 6.4 gig BIGFOOT died after a year (got a replacement within 2 weeks of call), and the other two (under 2 gigs) I am still using till today (both are SCSI from macintrash units).

IBM - I only had the pleasure of owning one. Bought an UltraStar 9.1 gig SCSI 80 pin drive and died after 3 months of usage. Got a replacement after two weeks but never really depend on it 100% after that. Worked fine now since 1998.. go figure.

DEC - Dec drives from DecStations are the best friggin SCSI drives. So abusive and do not want to die. I have 4 of them;2 one gigers and 2 500 megers. They are full height but are sweet.

Maxtor - Lets nto even go there. Peice of shit. Owned 10 gigier and a 40 giger, both died within 3 months and go replaced twice.

Seagate - My favorite. The only company that gladly still replaces your drives over the warrenty period. Deal with them alot. Let me tell you this one story. A few years ago, I bought 12 ~1.2 gig drives dirt cheap. One of them died, under warrenty, so I called them up and the sent me a repalcement unit of a 4 giger. I was surprised. Knowing that my 1.2 gigers would be unusable in a year and the warrenty was comign close, I did something pro-active. I removed them and paper-cliped* them all. Two weeks later, I got 9 four gig drives. I was happy! :)

*paper-cliped - a process to short curcuit a unit, in which you use a paper clip to touch various open circuity with each other. AKA you fry it :)