Slashdot Mirror
IBM DeskStar 75GXP Hard Drive Failures?
Sean Kelly asks: "Like a lot of other people, I went out and bought myself a nice 60GB IBM DeskStar 75GXP (ATA100, 7200rpm) hard drive to put in my sparkling new computer. Boy was that a mistake! A few months after I got the drive, it failed with horrific grinding and clicking noises, plus random data loss. So I RMA'd the first one and got a 'SERVICEABLE USED PART' replacement from IBM, which died of the same death after another few months. Not getting the hint, I RMA'd that one. Last week, I got the refab. drive back from IBM and it has already died, in less than a week! This time I did some site searching and found many people are having problems with this drive. Sites such as The Inquirer, Hexus, Tech Report, Hardware One,
Sysopt, and even this PCWorld have dedicated articles, forums and user reviews to these failing and defective drives. From what I can understand, IBM is not publicly acknowledging that they screwed up here. How many other people out there have had their 75GXP (or 60GXP) drives fail? What size were they? What part number? What did IBM do about it? It is my opinion that IBM should do something about this, since I've seen an unnaturally high number of complaints about this drive now that I started looking for customer feedback. Also, here is a letter I sent to IBM explaining my frustration with them. It has more information in it."
Mine's been working fine for a full year now. I've never heard a peep about them failing more than they should or anything...*shrug*
The problem is quite simple: IBM sells Western Digital drives re-labeled. I think that spells it out, don't you?
How's the ventilation in your computer case? Is it possible your drives are overheating?
The rumour going around the people who work on the disk subsystem in Linux is that certain lots/fabrication plats have lots of problems, and others are A-OK.
I recently got to experience the latter, when I got a machine with six of these disks as a RAID. To date, FIVE of the disks have had to be replaced, thanks God that did not include the system disk...
There was some rumbling on Storage Review that these drives may just be too fast for their electronics, and once you start filling up the outer sectors on the disk you will start getting errors. My friend has a pair of the 45GB 75XPs, and at least one of them has "issues". Every so often (now that the drive is full) the kernel will spit out:
ad4s1g: hard error reading fsbn 76293856 of 26874736-26874751 (ad4s1 bn 76293856; cn 8073 tn 63 sn 37)
followed by:
ad4: DMA problem fallback to PIO mode
So far the 60GXPs that I use have had no problems (knock on wood). I've seen at least once source that suggests that the 45GB versions of this drive are the most suseptable to having this problem. I suspect there was some poor quality control on these drives and some very marginal hardware was released onto the world (bad IBM, bad!), but that's more of a feeling since I don't have much evidence to support the claim.
I read the internet for the articles.
You are much better off sueing either in Small Claims court,where the limits tend to be around 1-1.5K dollars, don't require a lawyer on your part, and tend to be settled pretty quickly.
Threats of filling a class action lawsuit are a waste of time, you are much better off going to your county courthouse, filling the paper work, doing a quick web search on where to send the papers, and hire a courier to deliver them to IBMs local legal representative. When the day is done they will pay you your 1500 because it is much easier/cheaper to do that then to send two lawyers at $200/hr to your location to fight it (and still loose quite a bit of money)
At the place I work, we had a few 18GB IBM drives fail on us. They were standard inclusions with the Sun gear we use here (Sun ships a lot of machines with IBM and Seagate drives.) We found out from IBM that there was a recall on 9, 18 and 36 GB, 10,000 RPM drives manufactured between certain dates. These drives are pretty much guaranteed to fail, period. One of our other departments had over 90% of their suspect drives fail already. Our Sun reps came out to count how many we had, so that they could replace them. My department has well over 100 of the recalled drives. Fortunately, most of them are in gear that hasn't been put into production yet.
This is not a Fugazi
We got two IBM 75GXPs almost a year ago and they've consistantly given us trouble. At least once a week now we get something like:
hdg: dma_intr: status=0x51 { DriveReady SeekComplete Error }
hdg: dma_intr: error=0x40 { UncorrectableError }, LBAsect=38535423, sector=38535360
end_request: I/O error, dev 22:01 (hdg), sector 38535360
Over and over again.
It used to be more frequent and would cause the system to completely die after a while, requiring some console-based fscking. But many months back we changed some kernel option -- forgive me as I can't recall which, though I think it was the "burst bit" or something related -- and it hasn't been completely crashing, though we still get those errors about once a week.
I've heard stories of drive completely dying, but thankfully -- knock on wood -- that hasn't happened here. Here's what our setup looks like (from dmesg):
Uniform Multi-Platform E-IDE driver Revision: 6.31
ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
VP_IDE: IDE controller on PCI bus 00 dev 21
VP_IDE: chipset revision 16
VP_IDE: not 100% native mode: will probe irqs later
ide0: BM-DMA at 0xd800-0xd807, BIOS settings: hda:pio, hdb:pio
ide1: BM-DMA at 0xd808-0xd80f, BIOS settings: hdc:pio, hdd:pio
PDC20265: IDE controller on PCI bus 00 dev 88
PCI: Found IRQ 10 for device 00:11.0
PDC20265: chipset revision 2
PDC20265: not 100% native mode: will probe irqs later
PDC20265: (U)DMA Burst Bit ENABLED Primary PCI Mode Secondary PCI Mode.
ide2: BM-DMA at 0x8400-0x8407, BIOS settings: hde:pio, hdf:pio
ide3: BM-DMA at 0x8408-0x840f, BIOS settings: hdg:pio, hdh:pio
hde: IBM-DTLA-307030, ATA DISK drive
hdg: IBM-DTLA-307030, ATA DISK drive
ide2 at 0x9800-0x9807,0x9402 on irq 10
ide3 at 0x9000-0x9007,0x8802 on irq 10
hde: 60036480 sectors (30739 MB) w/1916KiB Cache, CHS=59560/16/63, UDMA(100)
hdg: 60036480 sectors (30739 MB) w/1916KiB Cache, CHS=59560/16/63, UDMA(100)
I have undergone a similar ordeal, only with KDS monitors. The original monitor I purchased failed within a few months, and each of the three replacement monitors I was provided with died within a day of receipt. After repeated verbal requests for a refund followed by several demand letters, I have now initiated a lawsuit against KDS and its business associates for failure to provide warranty service.
I am doing this through my jurisdiction's small claims court; the procedure, while lengthy, is not so complicated that I require a lawyer. If you decide to opt for the same route, I would be pleased to provide you with a copy of my claim. You might want to examine it and adapt it for your own purposes and jurisdictional requirements.
We also need to ask how many people have had *good* experiences. Now, it is clear from reading some of these posts that many users have had good experiences with them. But posting this sort of question here, where a large selection bias probably exists and where people who have had good luck are much *less* likely to post their success stories, is going to result in a very skewed picture of IBM - or other - problems. So even if a dozen Slashdot users wrote in with their own horror stories, I wouldn't know what to make of the results, especially since horror stories of other manufacturers are not solicited at the same time. No basis is therefore provided here for us to evaluate any results.
Don't ask a question if the results you get back won't mean much !
I've used pretty much every form of hard drive out there, and in my experience the Fujitsu (whom makes IBM's drives for them) and Western Digital drives are the worst to be had.
:) I've never, ever had a bad experience with Seagate, and will recommend them to anyone who listens. The few extra bucks you'll pay over a Maxtor or similar mass market drive are worth it a hundred times over when you've got critical data floating around in a non-raid situation. That's the situation for most college kids whom are cash-starved, yet need to have their thesis survive to the end of the term. :)
For IDE applications, I recommend the Seagate 7200rpm Baracuda line.
For SCSI, the Cheetah line. I have one of the age-old original 4.5GB Cheetahs. I've low level formatted it a few dozen times, high level formatted it a few hundred times, run a dozen operating systems on it, dropped it numerous times, whacked it, smacked it, and just generally abused the hell out of it in my torture (Q&A) server for many years. Once, when all the little brushless fans in my drive coolers decided to die within a hour of each other, it got so hot I couldn't even touch it for a few hours.
And.. it's still working.
Hippies smell.
When I hooked it up and powered up the system, the drive fried. We got a replacement, and that worked for a while, but when it got moved to another system, it got fried.
We eventually figured out that what was happening was that in the systems these drives were going in, you could not easily see the power connector when the drive was in its bay. That's not normally a problem, because power connectors are keyed. However, the plastic on these connectors was not very rigid. If you tried to put a power connector in wrong, the plastic would simply bend and allow the connection. There was no noticable difference in the force required to put the connector in right and that required to put it in wrong. So, anyone using the "the right way is the one that actually goes in" method of hooking up power had a 50/50 chance of getting it backwards.
Now here's the funny part. We found out from a contact in IBM that IBM was having something like a 20% failure rate on these drives during testing at their plant, because of their own QA people plugging the power in backwards!
I have 2 75GXPs in my home box (20G and 30G) and they have both been great. I've also installed another 5 or 6 75GXPs in other peoples computers and they've been totally reliable. I've also had a few older IBM drives that have been passed on to other family members. I've had such great experiences with IBM drives that I haven't bought any other brand (for myself or anyone else) in over a year.
Kind of a glowing endorsement, I know, but we tend to find what we're looking for. If you're looking for problems with a particular manufacturer you will most likely find what seems to be an inordinate number of them. How do the complaints you're finding for IBM stack up in comparison to other manufacturers?
Also, could it be a heat problem? I always make sure that any hard drive I install has at least one adjacent empty bay for air circulation, preferably one on each side.
With that many drives in a row going bad, I would suggest that the drives aren't the problem. I had a friend who rebuilt/upgraded his computer several times last year because he was having weird stability problems. He went through 3 mobos with 2 Athlons, 2 mobo's with 2 P3s, 2 video cards, 2 sound cards, 3 hard drives, 3 IDE controller cards, and God only knows how many cables/connectors before he finally figured out it was the power supply that was the problem.
Of course, at every step he would sell the "bad" hardware to somebody at about 1/3 what he paid for it and they would have no problems with it (which is how I got my Athlon, mobo, the 20G 75GXP, and a RAIDed Promise Ultra66 without being murdered by my wife ;)
Under capitalism man exploits man. Under communism it's the other way around.
But you're conceptually correct: generally, if something stops quickly, it does so in a short distance.
OK, enough nitpicking. :)
--
chahast at pangaea FOO dhs FOO org
s/foo/dot
I've got a 40GB Deskstar 60GXP. Here's what I've dug up on these uberdrives.
The platters in the 60GXPs are 20GBs each. Apparently, this is incredibly dense for consumer-grade hard drives (the 75's were 15GB each). The problem is this: heat. Everyone knows what happens if you stick a magnet on a grittle and let it cook for a bit. If you don't, here's the skinny: there are two ways to quickly and permanently demagnetize something. One is extreme shock. The other is heat. Harddrives heat up. If this heat is not dissipated, the platters heat up. When a platter is too hot, and the read/write head attempts to magnetize/demagnetize a sector, the magnetism may occur on the adjacent track.
My tip: A HDD cooler, even a passive one (giant heatsink) is better than nothing. I have one, and it's loud and obnoxious. But my HDD is fast, and that was my point in buying the Deskstar.
Hope this helps.
and I find it hilarious that IBM says on their product page that cnet praises their new 75gig drive, and then when you go to cnet, 50% of the readers have given it a thumbs down! nice work IBM..
been consulting for 16 years. standardized on IBM hard drives long ago. never had one fail.
purchased two 40GB DTLA-305040 drives in feb 2001.
one failed within 1 month.
the other developed bad spots within 1 month.
moved my clients to Seagate until IBM proves it can act responsibly.
ok, i fibbed a bit. i once replaced an old IBM laptop hard drive. that was the only dead IBM drive i'd ever seen, until this year.
i ordered my drive from some retailer on pricewatch. After about 3 months or so it started making noises that i can only liken to a clutch slipping out on a manual transmission car. Concurrent with these noises would be long seek times, even though i had disabled powersaving spin-down. I think the individual platters spin down or something. Anyways, i d/l'd the ibm diagnostics, it said i had a bad drive. So i called ibm tech support, they said too fucking bad, bec apparently my vendor had bought the drive from dell, thus making it 2nd hand and without warranty. I, being the trusting fool that i am, had thrown out all documentation, and felt screwed. Then i put the drive back in and decided to keep using it until it completely died. Only, and here's the catch, i didn't screw it back in, just left it on the slides in my case. The plm was solved! I was overjoyed and figured i could screw it in, since it would be there for a while. Immed after putting the screws in, the plm is back. I've left the screws out and haven't had a plm since. Yeah!!!
--why?
Oh boy!
/home partition was all clingy. Some mail files went through lotsa clings and clangs. Unable to be recovered whatsoever.
Mine was working great... but it died. It is on a home computer. Linux only. Good ventilation, usual stuff. No silly overcloking nor anything wierd. Plain desktop system. I even have a fan blowing air under the HDD!
After about 6 months of home use, I had trouble with my e-mail (using Kmail). Configuration got screwed up? I checked configuration, created new user, copyed config files in new user... and it didn't work.
Bug in old KDE 2.0? I tried upgrading to KDE 2.2.1. No success. I tried uninstalling and installing back the whole KDE 2.2.1 thing. No success. Still problems. Bad ones.
Software problem? The software was OK. I tried with other users. So next I checked the partition.
Oh Boy! CLinks! Clanks! It was an ReiserFS (I like to play safe). And some nasty errors showed up on the mail files.
I managed to save most of my email, copying the files to another partition. The
So I got myself an old drive and, while trying to install a linux on that 2nd drive, inorder to burn a CD from a reliable system, I made a mistake. With all the panic and pressure, I obliterated my partition table on the IBM 60 GB Desktar IDE-100 HDD instead of backing up the partition table.
OH boy!
It has been more than a month. I am still trying to recover that partition table. parted / gparted are no good: they have issues with ReiserFS and extended aprtitions (my data is on a ReiserFS extended partition).
I am very sad. I actualy bought that disk because it was an expensive IBM disk, because I wanted my data safe.
Of course the place where I bought it have ofered to replace it. But recovering the data costs many $$$ that I do not have, specially after getting layed off by a big blue company.
Oh boy! I'm pissed!