PDA

View Full Version : IBM knew the Deathstars would fail.



speedracer120
02-17-2004, 12:58 AM
So it seems, they knew even before it all happened and blew up in their faces and crapped out all the data on three of my purchases. :rolleyes: :angry:

http://techreport.com/onearticle.x/6292

IBM GXP fiasco secrets revealed
by Scott Wasson - 12:58 pm, February 16, 2004


The possibility of serious reliability problems with certain of IBM's GXP-series hard drives has been well documented. High failure rates may have forced IBM's decision to sell its hard drive business to Hitachi, and anecdotal evidence has been overwhelming, spawning a massive thread that will not die here at TR, as over a thousand posts have rolled in over time detailing users' experiences with drive failures, failures of replacement drives, and struggles to get IBM to honor its warranties. The fiasco led to the filing of a class-action lawsuit against IBM on behalf of GXP failure victims. However, IBM has consistently denied unusually high failure rates on its GXP drives.
Now, the law firm behind the suit, Sheller Ludwig & Badey, has updated its news page about the case (thanks Lo Yuk Fai) with a PDF version of a bombshell report from the Watchdog section of Maximum PC magazine's February issue. (Reprinted without permission?) The report is based on IBM e-mails and documents made available recently by a California court in connection with the class-action suit. The Maximum PC folks have dug up some sensational info in these documents, including internal IBM discussions revealing failure rates 10 times as high as competing products from other manufacturers, an apparent determination to address the problem through marketing denials rather than manufacturing changes, and a decision to sell a batch of faulty drives into distribution (and thus to consumers) after its rejection by a larger customer due to high failure rates. These practices seem to have raised significant concern inside IBM, based on excerpts from internal communications. A sample: "We have woven a story based upon half truths and misinformation that now places IBM in a position that is almost untenable."

In short, this is one case where the anecdotal evidence, which is all we had at the time we first reported on this story, proved to be pointing us in the right direction. IBM was quite apparently engaging in a pattern of deception and knowingly shipping faulty products to customers while hiding behind rhetoric about standard industry failure rates and the like. Obviously, IBM's decision to bail out of the hard drive business was an indicator that GXP failures were a serious problem. Now, thanks to the class-action suit, we seem to be learning the truth. Let's hope the court sees fit to punish IBM and reward consumers appropriately.

Jeffbx
02-17-2004, 05:08 AM
Damn... I hate to see a class action suit with something like this. A team of lawyers will split $50 million, and each consumer will get a coupon for $5 off their next HD purchase.

They should save everyone a lot of time & money and just issue a recall for every deathstar drive & swap it with a different model. This would probably be cheaper for IBM than the class action.

attgig
02-17-2004, 07:13 AM
yeah...but unless they get that much money...no lawyer would actually do it.

these lawsuits are more against IBM to slap their wrist for doing something bad and as a warning to the rest of the industry, saying this better not happen again, or else we'll really be after you next time.

bachviet
02-17-2004, 07:44 AM
Originally posted by Jeffbx
Damn... I hate to see a class action suit with something like this. A team of lawyers will split $50 million, and each consumer will get a coupon for $5 off their next HD purchase.

They should save everyone a lot of time & money and just issue a recall for every deathstar drive & swap it with a different model. This would probably be cheaper for IBM than the class action.
That's so true. Class action lawsuits don't do consumers any good. They only make the lawyers richer. :mad: I have two DeathStars and one of them is a replacement.

slaus
02-17-2004, 10:37 AM
I had a deathstar die on me last summer.

LegendKiller
02-17-2004, 11:08 AM
This shouldn't really come as a shock to anybody. Having worked for a disc drive company, I know that they do rigorous testing and statistical tracking on failure rates. They KNEW that there was a blip in the return rates, it is impossible to miss given the amount of press this was getting and with SR having their failure database.

When I was at Seagate I talked to some failure techs about the Deathstars, they had a nice chuckle. It was kinda known within the industry that there was a problem, the 75GXP's were the first major consumer IDE line to have glass+high density+high speed wrapped up into one package that wasn't stressed+heavy duty fault protection like SCSI drives. In my discussion with them we kind of settled on the issue being heat+glass platters = thermal expansion beyond what error correction could compensate for. This theory kind of settled with what was going on the real world. People would have less than optimal cooling, the drive would be physically fine, but the head would keep seeking, it wouldn't find the FAT table and would try and restart, thus the spin up/spin down cycle. If you stuck the drive in the freezer, it would be perfect for 10 min, then go to pot again.

Furthermore, if you Fdisked it, it would be ok for a while, until, according to our hypothesis, the drives tables would get outta whack again. This jivved with what they were seeing with their own glass platter drives. Whether its right or not, who knows.

It's sad that companies just dont come out and say they f'd up. However, they do make a decision based upon current trends, expected cost is a big calculation

cost*odds = expected cost


If the cost for coming out and recalling all drives is $100 per drive @ 1m drives, thats 100m dollars. They dont count goodwill gained since it is not a tangible asset.

Then they consider the likliness of a CA-suit succeeding, thus if it has a 80% chance that it will cost 50m, they figure that it will only cost them $40m. Thus, even with the lawsuit, they have saved themselves 60m dollars.


Ala, fight club, but it isn't just a movie fiction, this is the way companies operate.


LK

Kevster
02-17-2004, 11:22 AM
I have personally had 2 60GB GXPs fail on me and I have never bought another IBM product since.

I work with Nortel Passport switches every day and three years ago when we were upgrading our switch control processors from Series I to Series II, we had a huge failure rate with the new processors. Specifically, we had onboard hard drive failures in 98% of the defective processors. Guess who made the hard drives?

Jeffbx
02-17-2004, 12:17 PM
Heh, I used to use a Nortel Passport & an Accelar 1200. I finally got rid of the Accelar when Nortel wanted to charge me $7000 for a year of hardware maintenance, but would only offer $500 on a trade in for a new model. Hmmm, they admit they're only worth $500 in trade, yet they charge customers $7000 for 12 months of maintenance? I don't think so.

Now I have shiny new Cisco 4507's.

Kevster
02-17-2004, 01:50 PM
Originally posted by Jeffbx
Heh, I used to use a Nortel Passport & an Accelar 1200. I finally got rid of the Accelar when Nortel wanted to charge me $7000 for a year of hardware maintenance, but would only offer $500 on a trade in for a new model. Hmmm, they admit they're only worth $500 in trade, yet they charge customers $7000 for 12 months of maintenance? I don't think so.

Now I have shiny new Cisco 4507's.

We have the carrier class Nortel Passport 7540 switch on our network. Having one is one thing, but when you have over 400 of them deployed around the world, you have a bit more bargaining power with Nortel. We use them for our Frame Relay, X.25 and FRATM backbone transport. They do what they're supposed to do, give us very few problems and make those services very profitable for us. :)

Airencracken
02-17-2004, 02:55 PM
I thought this was funny...

http://www.littleboyinc.com/uploader/uploads/Funny Coincidence.JPG

Jeffbx
02-17-2004, 07:07 PM
Originally posted by Kevster


We have the carrier class Nortel Passport 7540 switch on our network. Having one is one thing, but when you have over 400 of them deployed around the world, you have a bit more bargaining power with Nortel. We use them for our Frame Relay, X.25 and FRATM backbone transport. They do what they're supposed to do, give us very few problems and make those services very profitable for us. :)

Man, we had nothing but bad luck with Nortel. Our parent company had a good relationship with them, so we also used them for frame relay & x.25... however, after multiple failures in many offices around the world, when we switched from frame relay to MPLS all of the network hardware switched to Cisco.

Joshua
02-18-2004, 03:52 AM
Nice thread hijack guys.. :P

Anyone IBM sucks for the Deathstar fiasco and should feel plenty of pain for it. ...'sides, their stuff sucks with the exception of Thinkpads which should be spun off to it's own company, this way whne IBM tanks they don't take Thinkpad with them.

Anyone remember the IBM Aptiva home pc primarily from Radio Shack?