[drive failure] tell us which drives have failed on you

[keyclone]

i get quite a number of PMs from people needing to transfer their licenses after re-installing due to a hard drive failure. this has always struck me as odd, since the drives are supposed to last a VERY long time and be VERY dependable.
(i've only ever had one drive fail on me in over 25+ years)

maybe this informal poll will help people steer clear of the bum drives

so my question is, and i will fill out the poll possibilities as i get responses, which make/model of drive has failed on you?

[algol]

None of those, but I just had a Raptor 150 go down.

[Hachoo]

In the 15 years or so I've had my own personal computers, I've only ever had 1 hard drive fail. It was a Maxtor (But ive had a ton of other maxtors that never died).

[keyclone]

if you can remember/find the make and model of the drive(s), i will put it on the poll

don't want to bash the wrong drives... please only vote on drives that have failed for you personally

[Hor]

In the past 10 yrs I've had 1 Seagate die (out of over 10) and maybe 2-3 WD drives fail. EVERY Maxtor I've touched has died. I will not use Maxtor at all anymore, if someone's paying me to put a pc together for them and they ask for Maxtor I strongly recomend against it.

[Starbuck_Jones]

Only computer hardware Ive personally had fail were two buffalo link stations. First one lasted about 8 months and they replaced it with a newer and bigger one that lasted about 4 months. I just tossed it into the trash at that point. I had a fan fail on an old G-Force card but never found out till went to upgrade it. That's all my personal experiences.

However working for Dell the last 7 years of my life, it doesn't matter what vendor it is, they will fail. About once or twice a year a manufactuer will put out just a crap run of equipment from hard drives, to motherboards, to optical drives. Some can be fixed with firmware up dates, but few and far between do people do that before the problem occurs. Other times the fail rates are so high the stock has to be simply purged. Where that purged stock goes is anyones guess, but you get what you pay for on thoes discount web sites.

So eventualy your golden, can do no wrong, favorite brand will go to crap. And the brand you swore you would never buy will be the only company thats making what you want that works. Keyclone, if you think your seeing a trend then you prolly are and it will be gone by the next run of equipment. But god knows how long it will take for bad stock to get out of the enviroment.

[Oswyn]

Starbuck hit it on the nose. It's not a matter if drives fail, but when. I've had pretty much every brand go bad on me (rough guess of 20+ drives total). The only brand that I seen more failures than normal were the infamous IBM Deathstars. I've heard of issues with newer high density Seagate drives, but have not experienced any failures yet.

[Buefurd]

The only drive I've ever had fail on me was a Hitachi 80GB 7200RPM drive from like 4 years ago. I'm currently using a Maxtor 250GB and Seagate 80GB that I've had for approx 3.5 years and they are still alive and kickin. Although I need more space :O

[Sam DeathWalker]

I use hatachie for my OS drives and maxtor/seagate for big data and WD for WoW/EQ........

Ya drives fail but I have everything backup so I just take it out, try it a 2nd time, if it fails again, toss in garbage and put in anther.


Sometimes you will get head crashed that make it look like the drive is no good but if you reformate its fine again.

[Starbuck_Jones]

Im going to disagree with your last statement there Deathwalker. If a head on a HDD "crashes" it will leave scratches on the platter. No amount of formatting is going to make that drive good again and the cast off from that scratch is going to just make things worse if the drive is even functional at that point.

[algol]

Im going to disagree with your last statement there Deathwalker. If a head on a HDD "crashes" it will leave scratches on the platter. No amount of formatting is going to make that drive good again and the cast off from that scratch is going to just make things worse if the drive is even functional at that point.
The Raptor and some Seagate 200GB I have around both make this horrible scraping noise when you power them up...then there's a 750GB that scrambles data you put on it in short order, and a 1GB that does nothing at all.

Drives die. If you're lucky, not for a long time. I have lots, some fairly old, so I get a dead one a few times a year. Sometimes a few weeks apart, sometimes many months apart.

[Sam DeathWalker]

Dont drives mark bad sectors when they format?

[Herc130]

The seagate 1.5TB drives suck major ass. I bought 4 of those drives. Of the 4, ZERO of them would take loading the OS. I found this out after my seagate 1tb bit the dust after 9 months of use. Then read forums full of post after post of people with the same drives and problems. Normally I use WD and yes I did have 1 WD go bad, so am not locked down on what brand to get. Went with Seagate because of the deals going on and thought it would be as good as WD....but yah..5 out of 5 drives bad. 1 died totally after 9 months, 4 will work as storage but for whatever reason will not load an OS. Keeps on failing durring install. Ordered a new WD and OS install went flawless after that.

[keyclone]

Keyclone, if you think your seeing a trend then you prolly are and it will be gone by the next run of equipment.i'm guessing you are probably right, sorry to say.

the poll will be mostly worthless then, as the same model can have good or bad runs. what it would point to is the quality control systems the companies have in place... but that can be variable as well.

hmph

[Starbuck_Jones]

No they don't Sam. A quick format just erases the index information on a drive. Think of the HDD like a text book with a table of contents/index. If you erased that info only, you wouldn't know where anything is in the book but all the pages are still full of txt and pictures. But for all useable purposes its now blank. Thats where all the rumors and like about even if you erase/delete a file the info is still there comes from. Its just removing the pointers to that info is all and eventualy will get overwritten by new data. Now a slow format like when you install windows and it gives you the option between a quick or slow. Is the same as a quick format, but it will run a chkdsk as well. This still doesnt erase any real data and yea chkdsk can fix minor problems with file structure, as well as help map out bad clusters and stuff. But it doesnt realy make sense for software to control bad blocks as it wont be relevant to anything unless that software is running.

There are a small handfull of apps out there that will do a low level format and acutaly write 0's or 1's to every block and most SCSI and Raid controlers can do this as well.

Most mapping out of bad blocks is done by the firmware of the HDD itself. Simple thresholds, like if a block fails to write, it tries again and again. Eventualy at some magic number of failed read attempts or write attempts the firmware flags that block as bad to the bad block list. If that list grows to a certian point then you will get things like SMART or predicted failures on post. Now if you have real physical damage to a drive like a head crash that leaves a nice scratch around the platter, theres no fixing that.

[Natch]

I've had a lot of WD drives die over the years, everything from 80gb drives to 500gb ones (yet I kept buying them for some strange reason, I've stopped now though :))

[Ughmahedhurtz]

It may not have been obvious (unless you're in the PC tech industry) but we just went through a major shift in drive platter density at the same time as a shift to faster interfaces (data rate and "percentage-volume" seek times) here in the last 2-3 years. Any time you go through dramatic changes like this, I'm never surprised that failure rates spike. Of course, you could also accurately assert that increased outsourcing of drive manufacturing to cheap options such as Shenzhen companies (e.g.: Foxconn) could also be a contributing factor.

I've had 7 PCs here in operation since about 2000 with various drive sizes and a few from each of the brands listed in the poll above, with zero failures whatsoever. Granted, none of mine are over 150GB in size and only two of those are the new raptor 150s. I generally buy at the "75th percentile" in PC equipment so I'm always a few revs past the "bleeding edge" and companies have usually had time to debug things when I choose to buy 'em. :P

My 2 copper, at any rate.

[Nitro]

I've had a few raptors of diff sizes go down on my personal machines but at work I've seen just about every make and model go down sooner or later.

[Zzyzxx71]

It may not have been obvious (unless you're in the PC tech industry) but we just went through a major shift in drive platter density at the same time as a shift to faster interfaces (data rate and "percentage-volume" seek times) here in the last 2-3 years. Any time you go through dramatic changes like this, I'm never surprised that failure rates spike. Of course, you could also accurately assert that increased outsourcing of drive manufacturing to cheap options such as Shenzhen companies (e.g.: Foxconn) could also be a contributing factor.

I've had 7 PCs here in operation since about 2000 with various drive sizes and a few from each of the brands listed in the poll above, with zero failures whatsoever. Granted, none of mine are over 150GB in size and only two of those are the new raptor 150s. I generally buy at the "75th percentile" in PC equipment so I'm always a few revs past the "bleeding edge" and companies have usually had time to debug things when I choose to buy 'em. :P

My 2 copper, at any rate.

This is the main reason I clamped my hands around 80gb drives (from pretty much any manufacturer) for about the last 4 years... for whatever reason (in my humble opinion) 80gb drives seemed much more solid than other higher capacity drives.

I've since moved on and have gathered relative faith in 500gb drives.

My current main PC runs 2x500gb striped, and a 2x80gb mirror for critical data. *shrug* - Starbuck hit it on the head.. they come in waves.. some good, some bad.

[bbj]

ive been building systems since the mid 80's.... Here are a few tricks ive learnt and this is applicable to almost any electronic device.

If it didnt fail within the first month then its gonna last years ( if you look after it ). Any electronic device will last longer if its kept cool. Not freezing, just cool ( like 20 deg C ) or close to that. Ive proven this time and again with 2 machines - one in hot enclosed space, and anther in ventilated cool place. some of the machines have uptimes of over 3 years ( the OS has a lot to do with that as well )

my latest 2 failues were 2 360G seagate drives. They were both the same age, and both lived in the same enclosure with zero ventilation. I burnt myself when i removed them form the box they were in. They did run for 2 years like that though.

[Clanked]

Back in 04 I lost 3 80GB drives. Two were WD's the third was a Seagate.
At the time I attributed them dying to the motherboard. (This was the second computer I had ever built, so I have a better understanding now) I swapped the mobo+new drive, and its still running smoothly.

Thinking about it now, I can't picture a motherboard actually eating drives. But they wouldn't even recognize in a second computer I had.
Great, now I am sitting here pondering what actually happened.

[Redbeard]

Not to change the subject of your post Rob but maybe if dealing with the reactivations is time consuming / annoying we could help you come up with an alternative system?

I assume youre doing some reactivation / deactivation coding, maybe you could make a web interface where people could enter their info and automate it. Then you'd only have to personally see to cases (such as mine, a few days back) where people didnt have all their info.

Hmm.

Or maybe thats more work than just doing it manually haha I dont know.

[keyclone]

@Redbeard
the license switching isn't too big of a deal, and helps keep me on top of the questions (as i don't wander too far from the computer). it does eat my non-real-work time, thus killing my keyclone coding time... but i'm dealing with that.

i do have a plan i want to implement for the next release, and that is holding up the new version (that and the new visual studio requirement to push .net into the install of even the most basic apps... very annoying). the plan will allow the user to drive their own transfer, removing the time it takes for me to respond (i try to be fairly quick, but sometimes i sleep). hopefully i can resolve this soon.

[daviddoran]

I've had every brand fail on me at least once. I just buy what has the longer warranty now, and try to remember to keep everything backed up


*note to self, back up wow folders....

[Kaynin]

I have had several Maxtor's die on me (in the ~10Gb era) whilest all my WD's from that time kept working. Had 1 or 2 seagates too that I didn't have problems with.

I haven't had a harddrive die on me anymore in like 6 years or so though.

[Kaynin]

Back in 04 I lost 3 80GB drives. Two were WD's the third was a Seagate.
At the time I attributed them dying to the motherboard. (This was the second computer I had ever built, so I have a better understanding now) I swapped the mobo+new drive, and its still running smoothly.

Thinking about it now, I can't picture a motherboard actually eating drives. But they wouldn't even recognize in a second computer I had.
Great, now I am sitting here pondering what actually happened.

I had an ISA videocard way back (some 1mb vga card) that, when slotted on any type of motherboard just once and started up, it would destroy the mobo. I fried 2 mobo's with that videocard. Then figured it was the videocard that was doing it. Later on I wanted to test it to be sure, took an old working comp that wasn't worth anything anymore and installed that videocard. Start up, blackness. Swap for previous card, blackness again.

So, I wouldn't say it's impossible for a mobo to fry a harddrive. I've seen similar things happen. :P

[Hor]

Add a 250g Seagate to the mix ... I've had it for 4 years though, and it's been in constant use. So, I can't be too upset. It came the day I bragged to a friend about that computer being "ole reliable" lol. A couple of cheapo Rosewill Fans died and the drive overheated and failed. Now it boots, then freezes after a few seconds. So ... today I'm spending my day off putting in new fans (thermaltakes this time) new hard drive (a 500g Seagate) and my OS and software. This is going to be a long day.

[heffner]

Never had a failure in 20 years of using hard drives. Most likely due to buying new computers every couple of years though. The largest HD I have owned is 500GB. I typically buy Seagate or WD, but owned pretty much every brand at some point.

Everything else has failed on me though.

[aboron]

The most annoying failure I had was both the WD 36 Gig original model raptors I bought failed within the same week - i hooked them up individually to a different pc and ran WD's analysis tool and it confirmed them as both bad. Both were still under warranty, and the return location was down the block from work, so i saved shipping and dropped them off for repair. They are now running in a Raid1 array to boot my primary linux server from, hopefully they don't fail on the same day.


Second most annoying failure was 2 of the 5 160GB drives in my old raid5 array failed, they were Seagate 7200.8 series drives, but i was transporting them as bare drives in carriers not installed in the server and they failed when booted back up, so something may have physically happened to them on the drive to their new home.

[Hor]

Eww, addendum.

Looks like I won't be boxing for a while, unless I can actually box on my laptop. (which I doubt) Looks like I've had a big meltdown in the old rig, and it's an AMD Socket 939 so getting replacement parts is next impossible. RIP AM rig and your 7950 GTX!

[elsegundo]

i've been told by one of my CS teachers to never buy two drives that's of the same production, especially if i want to put those in a raid array. the reason being is that you dont want all of your drives failing at once, and buying them at different productions, even from different companies, can mitigate that risk.
is this still the case?

[Souca]

I've been buying and replacing drives for most of my life and I can say that there isn't a brand who is going to be superior until you get into enterprise class drives. The one thing you should look at is the replacement policy. At one point when I was buying a good deal of drives in bulk I ran into a bad batch from Maxtor. The thing that kept me buying from them was that I got replacements the next day via FedEx along with a mailer to return the defective drive in t he next 30 days. Not sure if they are still doing that as it was about 7 years back. I can't give any particular reason I have for the brand I chose other than warrenty and general trends in platter density and drive offerings. That said, I have never liked Western Digital. Likely there is no real data, but they just feel inferior to me. I liked Quantum until just before Maxtor bought them. Maxtor was good for a while, but then I switched to Hitachi or Seagate depending on the model. Lately I've been using Seagate. Can't comment onthe 1.5 TB for booting as I don't use large drives for the OS, too expensive to mirror; use mine for fraps file storage.

In all this time I have never seen a real head crash. Most drives die due to heat, bearing failure, or electrical component MTBF. I look at hard drives like socks, eventually you should replace them and a little while later they will have holes too big to ignore.

As for buying drives in the same lot for a RAID, if you are hitting the MTBF average for the drives then you have had them in the raid too long and should have had backups ;) Or a service contract; always love when the guys from EMC or other drive companies show up within 30 minutes of a failure. For most home users you aren't going to see a difference IMHO.

- Souca -

[Chorizotarian]

I've run my last 2 computers off dual Raptors in RAID 0, which is basically begging for a fatal hard disk error, and never had a single problem. IMO the quality of Raptors is >>> than most cheap gignormous hard drives. I've had lots of Caviars, etc. fail.