I have an external 1TB drive, but it’s RAID 1 over three separate internal disks and quite sturdy, so I’m more worried about running out of space than I am of it failing anytime soon.
I’ve got a few of the 1.5 TB drives; they’re all working fine. Of course, there’s no data on long-term reliability yet.
That said, sure I’d trust them. Drives are more reliable now than they’ve every been.
I always use fairly high-capacity drives (video and other high-bandwidth hobbies), and usually buy the cheaper ones. Aside from a couple of first or second-day failures, I’ve had only one drive go bad on me with less than five years of service life – and it made it four.
That seems to correlate with the data, which is surprisingly limited for a topic that affects so many people. This article talks about (and references more detailed) results of a couple of reliability studies: KHB: Real-world disk failure rates: surprises, surprises, and more surprises [LWN.net] Mostly, they seem to find that capacity isn’t a specific predictor of failure (and that some of the predictors they though were, like temperature, aren’t either). This of course ignores the specific Seagate bad firmware problem with 1.5’s, which was so widespread that it would skew the data if included (and has been fixed in any case).
I always buy hard drives 2 at a time (pairs) so that one can be a backup of the other. It’s a passive (manual) mirroring system – I just periodically sync the contents of one drive to the backup drive.
When my friends asked me which external drive to buy, I told them not to buy just 1 but to get 2 of them so that one can backup the other.
I have about 20 2TB drives – none have failed yet. I also have 7 1TB drives and one of those crashed a few months ago. Fortunately, I had an up-to-date mirror copy of that drive.
I used to do RAID but abandoned that approach because it is extremely difficult to set up a stable system with the newest drives (such as the 2TB drives).
If you plan (budget) for some type of redundancy from the start (whether mirrored pairs or RAID), then there’s less anxiety with a 2TB drive failing and losing your data.
I’ve been using a Seagate Barracuda 1.5TB drive since the beginning of the year and all is well. I back up vital data on multiple drives. About 660GB has been used up. Almost time to start looking for another one!
It’s not really 47TB because much of that is redundant data (mirror copies, backups). More like ~20TB.
I shoot hi-def video and it consumes a lot of space. But my storage is miniscule compared to others. Some wedding videographers have hundreds of terabytes of video footage.
Google stores PETAbytes (thousands of terabytes) of data.
I suppose I’d trust a two terabyte drive, but at that point the problem is that backups are difficult. (Even a recordable Blu-ray Disc can store 50GB at best, and forget trying to use CDs or DVDs for that much stuff.) I suppose you could copy the data to a second drive, but how long would that take?
Fortunately, I don’t need that much storage yet. My most critical files (Outlook PST file, the Quicken data file, letters, etc.) amount to about 1.5GB. And the other stuff (mostly digital media files and software installation files) is about 300GB.
I have a friend who worked for EMC (which manufactures enterprise storage systems) for years. I remember when he told me how they put dozens of consumer-grade hard drives in a rack enclosure to provide a whole terabyte of storage for clients like airlines and banks. It seemed like such an unimaginably big amount, but I suppose I’ll need it soon enough.
Hang on a second, do you mean that you are using 1[sup]1[/sup]/[sub]2[/sub] disks for data and 1[sup]1[/sup]/[sub]2[/sub] for the backup? I didn’t even know that was possible with RAID 1.
If the middle disk fails wouldn’t that mean you lose a chunk of your data AND the backup too?
Are you certain you’re not using RAID 5? This uses 2 disks for storage, and the third for a parity check.
Note quite Pete, the drives are set up so that the parity is spread out amongst all three.
(see here: RAID - Wikipedia )
We’ve got about 3.5 Tb on the fileserver in the basement. For the most part, it’s backups of the laptops, so if those backups die, as long as the laptops re okay, we’re golden. The rest is offsite netflix backups so if we lose them, we just add the movies back to our queue.
I have a theory that a drive will run almost indefinately IF it’s got good clean power. I’ve had, probably, 20-30 hard disks in my personal care, and can think of maybe two that died.
If your fileserver is on a good clean UPS, and you back up once a week, and the drives spin down when idle, you shouldn’t HAVE a failed disk. If your laptop gets battered about, and it’s plugged into the same kitchen circuit as the Fridge, dishwasher, Microwave and food processor, it’s power may not be quite as clean, and the drive might be subject to more mishaps.
Beyond infant mortality, modern drives (especially with emergency head parking and sudden shock sensors) are AMAZINGLY sturdy, considering the tolerances involved.
While “good clean power” won’t hurt, it’s clearly not power problems that cause the vast majority of drive failures - witness the number of RAID arrays that have a single drive fail, while all the other ones are fine. All the drives are being fed from the same power. The power supply in a decent computer has excellent transient rejection - my G5 will ride through 1/2 second of power failure. The power is further regulated and filtered on the drive itself.