5punk

You may remember that a while ago one of the hard drives I had steam installed on died a horrible death. Today, Windows merrily informs me that the hard drive that replaced the dead one has "failed SMART". Which seems bad, I dunno.

Its not the end of the world as the only things on there are Steam and Diablo, both of which I can just re-download if the drive all of a sudden goes tits up. I will be getting a replacement for it ready anyway, but it occurs to me: What is more likely, that I'm just unlucky or that there is something killing the hard drives? I wondered if they were overheating, but there's another hard drive in the same bit of my case that has been merrily chugging along fine the whole time. Could it be the power supply or something? I ask because although a new hard drive is not going to cost the earth, I would rather not have to keep replacing the thing every couple of months, especially not if I can fix the situation by just bunging a few quid down on some additional part now and be done with it.

PSU is about the only common factor there really. Do you have a different Molex to try? If it works you'll probably never know, but it's worth a shot.

If it's a few months old I'd warranty the fucker. If you're molex powering them, try a different one like pants suggested or try switching to a sata power connector?

SMART is unfortunately not overly so when it comes to communicating what the problem is. It could mean it's detected bad sectors on the disk that will no longer write, or it could mean it's about to fire an actuator out the side of the case .

Do you have logging enabled? The test log or possibly self-test (can't remember the name properly) should show you a more precise indicator of which bit is playing silly buggers.

HereComesPete wrote:If it's a few months old I'd warranty the fucker.

Its only been in my machine for a few months, but before that it was in a mates server as part of a RAID for god knows how long. So there is a reasonable chance the thing has just failed from old age to be honest. I just want to cover my bases.

I'll try a different power connector, see if that helps.

Oh...ive just installed a trial of HD Tune to see if that can shed any light on the matter, and it tells me that the drive is running at about 60 degrees. That seems rather warm. I should probably do something about that.

A tad warm yes, but more drives fail from cold that heat apparently and even high temps don't tend to make much difference. Google did a study on their data centre -Kinda interesting.

Joose wrote:Its only been in my machine for a few months, but before that it was in a mates server as part of a RAID for god knows how long. So there is a reasonable chance the thing has just failed from old age to be honest. I just want to cover my bases.

Ah. I'd go with old age then.

One of my old drives?

I'd just RMA it, also hdtunepro (I think) is pretty good at actually telling you the SMART fuckup.

If it's been part of a RAID array for any length of time, it's seen more use than a lot of disks see in their lifetime, RAID disks are in almost constant use so do tend to wear out - that's why the 'I' initially meant Inexpensive.

Get a backup of it, and do a full format of it. Personally I'd replace it and use the formatted one as a spare, under light use (like in a caddy as a USB disk) it'll probably last a while. If new isn't an option, restore everything and keep your eye on it, keeping the backup up to date.

I have four 1TB disks that have shown errors at some point while inside my machine, which work fine as backup disks - I just put them in a caddy like the one pictured below and swap them around every few months and none of them have failed since coming out of the machine. For backups I just use xcopy.