Hear Ye! Since 1998.
Please note: This post is at least 3 years old. Links may be broken, information may be out of date, and the views expressed in the post may no longer be held.
17
Dec 01
Mon

I lost my hard drive over the weekend. 45 gigs – 30000 emails, over 100 megs of icq logs, 5 years of school and uni documents, all the web sites I’ve ever made since 93/94, a fair chunk of my mp3 collection… all gone. It came without warning – an unrecoverable head crash (the drive is making a clicking noise) during operation. Motherboards won’t detect the drive, so using software to recover data is not possible. IBM 75GXPs are notorious for their unreliability and I guess I found out how true that is. Mine was only 15 months old. The computer still boots, I have another 45 gigger in there where the OS is installed, it’s just that I lost my data drive. God, it’s fucking painful.

I rang up IBM who referred me to Digiland (their hard drive distributor), and of course they don’t repair drives, only replace them. They couldn’t refer me to a data recovery service, but those guys are exorbitant (about $100 per gigabyte recovered I think.)

Of course, people will say “backup backup backup” but how do you backup 100 gigs of data? There was a Slashdot thread on this recently, but even if I bought another hard drive, it would be soon filled with new data, and not backup data.

Companies on my hard drive blacklist now include Quantum, Fujitsu and IBM. Next time a Western Digital or Seagate will be the way to go.

Responses

I had a similar experience with a 6Gb (or so) Western Digital a few years back. It was not such a large loss because it was only 6Gb, but scale back the time, and the amount of space each document and item takes up. I lost like 6-8 years of accumulated data, it was devastating.

From what I understand, Western Digital has cleaned up their act, but I still don’t take chances. I have a Seagate, Quantum and Maxtor in my computer, and it gets backed up onto the server every evening in chunks, it takes one week to complete a backup cycle, with all “recent” docs getting priority… Once a month I use my CDRW to burn a “hard” copy… it’s worth the effort, trust me.

A Canadian reader,
Kevin

—–

Sorry to hear about it, but after what I’ve seen recently with my home machine it doesn’t
surprise me. I lost a 40 gig IBM drive about 3 months ago, and with it 4 years of Quake
demos and screenshots. At work we had a bad batch of 36 gig SCSI drives in 4 new servers.
In the end we forced our supplier to replace all 15 new drives with Seagate drives.

The new seagate barracuda drives are damn quick and very quiet. They don’t produce
anywhere near as much heat as the IBM drives either.

I’ve gone through the same thing as you by the sounds of it. I used to be a big fan of
Quantum drives, had 3 fireball drives until a bad experience with one of them. IBM were
good for a while, but I wouldn’t touch them now. Seagate is my choice for now :)
– Fuzzy

—–

I suggest you forego some drive space and set up a RAID if you want
reliable data storage. You may have to shell out for a RAID controller,
if your mobo doesn’t have one built in (ie. you didn’t decide to spend
an extra $40 or so to have the feature “just in case” when you bought
the mobo), and your 2 45 gig drives will give you considerably less than
90 gig of space (how much less depends on the configuration), but it
will be reasonably reliable. The point of a RAID is that the data is
duplicated on other drives to minimise data loss when one drive dies …
of course this is more effective with more than 2 drives but hey. Also
I’ve got a Maxtor drive that I still use that’s well over 6 years old
(it’s 1.3gb, what does that tell you) that still doesn’t have any
problems. And 2 Quantum Fireballs, one is a 1gb SCSI about 5 years old
(only now beginning to play up, but works well enough that I got all
important data off it just in case), and one 6 gb (6.4 if you use 1gb =
10^9 bytes, as quantum does) that’s about 3 years old and working fine.
Although the cable jam in my case has caused some of the cables to go
dodgy and occasionally I have to open up the case and jiggle them to get
windows to boot. So the way I see it, you’re just plain goddamn unlucky.
The only piece of IDE hardware I’ve had that ever went so bad I had to
replace it was an old Creative CDROM drive, and even that put in about 4
years of faithful service first. So maybe you should consider putting
the warranty replacement drive in the centre of a pentagram with candles
and sacrificing a goat or something.
– Victor

Well, I’d be inclined to agree with you about the drive’s failure as being a freak accident, but 75GXPs are not held in good esteem. Check out Storage Review’s Reliability survey (registration req) and see how IBM’s 75GXP range fares. It’s not good. In fact, in the US there is a class action lawsuit underway, suing IBM for not delivering a product as advertised – the false advertising is, you guessed it, reliability. My cousin owns about 5 GXPs, and he’s experienced failures in 3 of them (not complete failures, but significant data corruption/recoverable clicking).

Quantum has given out on me twice before. I have a couple old 2 and 6GB drives (Maxtor I think) that are years old now and they are still chugging away fine.

The only problem with RAID1 is that you have to double your expenditure on hard drives to back up the data. You have, in effect, 50% of unusable hard drive space. I also assume that if the hard disk controller dies, you’ve lost your raid array (unless you have disk duplexing, but you need another controller for that). RAID5 would alleviate this problem only wasting one drive (if you have more than two drives), but most onboard mobo raid controllers only support RAID1. Nonetheless, for consumers, having to buy extra drives for RAID1 is expensive, and when you have 100gb or more that requires mirroring, RAID1 is not cheap. Furthermore, most data on the drive is not important – only documents and media require backing up (as opposed to program files). So, I guess the alternative is to buy a separate physical hard drive and use it partially for backups of data, and use the rest of the space for installing programs. For automated directory backups I can recommend Second Copy. Because of Second Copy, I have a mail archive that was made last August. Unfortunately, after August, my registry’s software hive got corrupted (dodgy IBM hardware again!) and I had to reinstall all my programs. I got lazy and didn’t set up Second Copy again.