Seeing as the HDD is one of the few mechanical parts within a PC, it's pretty obvious that the potential is there for them to fail, just the same as any other mechanical part. I have had my fair share of HDD failures personally, and see them frequently enough at work to appreciate the problem. It's a gutter when the backups have been put off longer than they should have been too, and the HDD fails. I just recently lost a boatload of data from 2 HDD fails.
If these are server drives or commercial workstation drives they'll be on for long periods at a time. Google might not be buying joe-schmo consumer grade drives that we buy, which would make comparisons with what we have nd what they have difficult. Regards temperature, 40 degrees isnt really 'hot', although i can believe that some temperature is needed to keep bearings healthy. They are mechanical devices and so usage will be a factor to a small degree. I believe the most common cause of failure is spindle drive failure. Case fans dont last forever and eventually the bearing does go, so i can understand an analogy to hard drive motors. The only hard drives that i have had some contact with that have died have been cheap ones. My samsung 250gb drive will be 4 years old in november. If it failed today it would have had a very good run. EDIT: im a retard and should have read the google study referenced. I still believe usage is a big part of failure rates but im not going to pick apart the wording of the article now just after work.
Yes they fail. I had a almost new WD Green fail in my RAID array. If it wasn't a RAID, I'd have lost data. Well.. I wouldn't have, because I'm an obsessive who mirrors his RAID to another RAID... but most people would have lost data. /thread.
If the data is really sensitive, it may be cheaper to just back them onto more permanent forms of storage than maintaining a RAID array.
I've had three harddrives fail in the last 4 years. Two were IDE drives in a server (both same make & mode, failed within a week of each other ) and one was a brand new SATA drive that didn't even last the installation of Vista 64. They do fail...protect your data!
Ok.. first: What is more permanent? Optical disks? Afraid not. These have terrible archival qualities, and are very easily damaged. I would not trust optical. Second: If you were to use optical, how often could you reasonably back up? I back up each drive every 48 hours. Doing that with slow, low capacity optical drives would be impractical, and you would also have to be present during the back up to change disks. Third: RAID IS safe... that's the point of it. It can withstand drive failures. Just remember, RAID isn't back up, it's redundancy. You should always have your back up in more than one place. A dedicated drive, and then a mirror of that on something with redundancy. If you're as paranoid as me, both would have redundancy. Optical disks for back up suck.
+1 on this, I couldn't agree more. The day that you decide that you don't think you need to back up your important data is the day your data is at risk and, keeping sod's law in mind, is the day that your hard drives will start to get flakey. Copy your data onto more than one spindle, using RAID or not, is the first step in protecting it. The more places your data is the better protected you are.
I've never had a drive fail but I'd hate if one did and I didn't have it backed up BTW RAID isn't backup and shouldn't be treated as such
In my experience, when PCs break down hardware-wise, it's usually for the following reasons, sorted by descending likelihood: 1. Hard Drive 2. GPU 3. Motherboard 4. PSU 5. RAM 4. Processor. (I've only seen it once.)
Ooooooohhhh I'd agree with that but put motherboard up to 2nd. Especially on cheap econo-boxes. GPUs can be very flaky, they are soooo not made to last. Short but frantic and hairy lifespan. You've actually come across a dead non-overclocked cpu? If it did work and then just failed, well, that kind of thing is frankly extremely rare. What type was it do you remember?? It actually died during an install?? LOL! how did you react? I think I would have started laughing. Reading these comments I must be extremely lucky with my drives (nervously eyes 6 month old Caviar Black).
I haven't really had a drive fail, and when I do, it's not a big deal since Most of my photos and such is stored in multiple computers and Some SD cards.
As noted before though, my problem's on a different magnitude to this. If I took out the mirrors, one drive failure would mean losing about 850GB of data. I don't think I could live with myself. Like someone said a bit earlier, I'd rather buy 2 1TB drives and mirror them than by one 2TB drive and risk losing 2TB of data in one go.
I had 2 maxtors die on me for no reason, my WD and Seagate drives have been good so far but i don't trust the seagate.
I haven't had a single HDD fail yet. I mirror important data across multiple systems on my LAN to ensure better redundancy than RAID could offer, with offline and off-site backups as a final touch. Note that keeping a HDD running 24/7 is better than starting it a few times a day. HDDs are rated for N number of spin-ups, each of which slightly damages the bearings (as happens with all bearings). I run all my HDDs 24/7 for this reason.
yes constantly (24/7 use) and in my opinion its heat the 2.5 issue is wife having laptop on sofa etc....
As I work at a Data Centre I see failed hard disks quite regularly, 15,000 RPM 3.5" SCSI types as well as SAS 2.5" drives ranging from 7,200 to 15,000 RPM that are becoming more popular on the newer HP servers (Blades etc). We use HP and Sun servers and they supply their own branded disks as replacements so can't positively say who the manufacturers are (I assume HP / Sun don't manufacturer their own drives). Hardly ever see RAM or CPU failures, if they are in a failed state then its usually dead on arrival. Motherboard replacements are quite regular but again no more on one manufacturer than the other. Other than those I've never had a failed disk in any of my machines (desktops or laptops). I thought I had a failed disk in my previous laptop (Sony Viao) a couple of years ago but it turned out to be a IDE controller failure as opposed to the disk itself so was able to rescue the data from the drive. Then had to buy a new laptop (it was overdue anyway and I was eyeing up a Dell M1530 in anycase!).
I've had one DOA and one failiure after a good, long life. As has been said, they're one of the only moving parts a pc has, and as such one of the more likely to break. They're also the most catastrophiv to lose. If I lost a CPU (or anything else for that matter) I could get an exact replacement tomorrow. Not so with a HDD full of uni work, priceless photos, CVs, music etc.
