I try to avoid WD and Maxtor drives, primarily based on information I read about them a few years ago that could well be wrong at this point.

it’s hard to say. it’s true that a controller can cause failure, but unless you experience consistent problems with drives failing when attached to a given array, you can’t say it’s the controller. it’s more likely to be the drives that are failing. you mentioned that both you and your friend are using WD drives.

i do agree with the comment about using linux software raid vs Promise’s solution. Promise isn’t a true hardware RAID anyways, it uses a software solution that’s proprietary to their driver if i recall. in which case you might as well go with a more standardized software solution, i.e. one offered by the operating system. that, or invest in a quality 3ware solution.

“quality 3ware solution” — a good friend of mine who writes RAID drivers claims that 3ware sucks as well, so i’m not sure that’s the answer. I have a 3ware 8-channel S-ATA card in my server, which seems to be working fine, but it fails catastrophically if I try to upgrade XP to SP2. argh.

stay way from WD drives man. just had wd 120 gb external drive knocking badly then failed. all data lost. all techies here know that wd is a waste. go with seagate.

did i ever tell you how, while i told myself never to buy another WD drive… a few months ago i bought one because it was on sale and i couldn’t resist the 8MB buffer? well… guess what, a few hours ago, that crappy WD drive died a horrible death without any warning and took ALL my data with it!!! is it irony that it should happen less than a week after your post? or did we all receive the same batch of bad WD drives???!!!!

someone at Western Digital should be shot for unleashing such horrid drives on the world… and i need to be shot for ignoring my boycott of WD drives that began back in ’98.

I had that exact same model drive go bad after 1 year of “lite” use. Sent it in under warranty and I got a 250GB one back. It was the only HD I’ve ever had go bad on me, I’m guessing those models are not so great.

I just had 2 WD1200′s in the same array fail on a promise card in less than a year.

Hey stop slagging off western digital!

Firstly, whatever configuration you are using (even if it’s RAID 1) you need to have some kind of backup strategy in place! I don’t have much sympathy for people who a.) take some cavalier attitude keep all their precious data on 1 disk and/or b.) run their disks in undercooled system 24/7, then they proceed to blame the disk manufacturer.

If someone breaks into your building and steals your RAID 1/5 how will you recover your data?

I have used lots of different brands of hard disks and I have seen all of them fail randomly at different points in their lifecycle. The fact of the matter is why would ANYONE be stupid enough to put the integrity of their data in the hands of any manufacturer’s hear-say reputation, when really all manufacturers want to do is sell a load of drives and make a ton of cash.

The reason western digital are good is not because their drives are impervious to failure, it is because they offer exceptional warranties on all their drives ranging from 3 years on OEM desktop to 5 years on the slightly pricier enterprise class drives (Like the Caviar RE/RE2s). They back their warranty up with exceptional service: they will advance ship you the drives within about 5 days, the only cost to you is the return of the drive. In addition they offer the remainder of the faulty drive’s warranty on the replacement – so it makes no difference if it is refurb or if it is new.

Of course Western Digital drives are ahead of the pack in so many other ways besides warranty. 16MB caches, most of the range are now sold with SATA 300 external interface, lots of new models now support RAFF and NCQ / TCQ natively, great seek times and transfer rates. Also the RAID editions are a great idea, because they are warranted for 24/7 operation which everyone here seems to be doing with desktop class drives.

Hey Eric, Well what you said was interesting. I have upwards of 20 drives, ranging from brand new top of the range stuff like the WD4000YRs and RaptorX 150s right down to the lowly ancient stuff like WD340s and older. I will come clean with you on this, I have just had to RMA a Raptor X which failed after about 2 months without any warning, whilst the drive was simply idling in my desktop.

It is my theory that WD drive quality may in fact be deteriorating as demand for drives increases and market pressures force per GB sale price downwards WD seem more inclined to ship lesser quality goods. Also with WD trying to lead the market with performance specs it may be placing higher demands on technology than it was designed for (eg in the case of the RaptorX at least, using a polycarbonate cover seems to be asking for trouble). It would be interesting if manufacturers were to publish quarterly returns statistics, though obviously it would never happen.

That said this does not detract from my previous statements about WD warranties and service. Also bear in mind that most of my drives are WD – so statistically it is most likely to be a WD drive that will fail on me (only the 1 has so far).

What I will say is this: 1.) From who else can you buy a 10,000RPM SATA drive? When it comes to speed everyone else is lagging behind WD. Sure you could go the Ultra320 route, but who has money for a controller not to mention PCIX mobo, ECC RAM, dual proc…. ?? Especially when all I am using it for is superfast boot times, ninja gaming speeds and burning multiple DVDRs at once.

2.) Secure Connect – such a simple feature but what peace of mind. So many people dis SecureCon as a pointless gimmick but really, who wants a disk falling out of a RAID array when you’re in the middle of something? You have to pull your rackmount server apart looking for the problem. This is especially annoying if you ever have to move your machines (eg LAN parties). Flimsy SATA cables is the last thing you need ;)

The above two features are the main points that differentiate WD drives and keep me loyal to their brand. Until other manufacturers wake up and notice the real demand for this kind of innovative thinking, I’m sticking with WD.

I have noticed that Western Digital drives have been failing at an increasing rate. About 1 year ago I received a new WD 500GB RE2 drive. That drive failed within the first week. Then WD sent a new one and that drive ran for 9 months then failed. Now I have my third drive within one year. This one is OK so far only 2 days :). The RE2 drives according to WD are the RAID optimized heavy duty drives designed for 24×7 use and should have a lower chance to fail.

I bought a set of four 500Gb Raid Edition drives 1 year ago and the first of them just failed. I’m running a raid 5 array on an intel 6 port megaraid card and although i was able to rebuild the array, the same disk again failed during the consistancy check. I paid extra for these disks instead of the hitachi’s (which a data recovery specialist I know recommended above all else) and am rather disapointed. I’ve just bought a 500Gb deskstar now which I will attempt to integrate tonight – just wondered if there is anything I should do during this sensitive time? I have the rebuild rate set to 20% as I’m scarred of thrashing the hell out of the remaining good drives (all from the same batch) but obviously this means the rebuild takes an age (approx 48 hours). If I had a power cut or crash during this time is my data gone or can a rebuild recover? If I add my new hitachi drive as a hot-spare, rather than swap the failed disk for this one is there any advantage? I know raid 3 keeps the parity data on one disk and raid 5 spreads this data between disks – is there any way to make a raid 5 or raid 3 array mirror the parity data so that I can have 2 disks in the array fail? Speed isn’t an issue here, the disk is only ever accessed through 100mb network so i’m not worried about bottle necks particularly and i do have the ability to change the raid type (although i doubt i can do this on a failed array and i certainly wouldn’t want to try it even if it let me)
I see people talking about software raid here – I was under the impression that it didn’t work, and should be avoided like a land mine.