To be fair, the review threads on Newegg and the like for popular models are basically an endless sea of “5 Eggs, 1 week-1 month of ownership, FASTEST THING EVER, CONS = COST, BUY IT NOW!” and “1 Egg, 1 month-1 year of ownership, WAS REALLY FAST, BUT FAILED AND BIOS COULDN’T SEE IT ANYMORE, WTF?!” reviews, with statistically little in between.
The most common failure seems to be something on the controller itself, because the drive just vanishes from the BIOS. It’s not a mobo issue or a cable issue; the drive as an entire whole is just unreadable. Theoretically, the data cells are still alive somewhere, but it’s my estimation that the average home user doesn’t have the kind of tech necessary to try to effect a reclamation of that.
In my case, it was more that the drive would intermittently disappear from BIOS on a couple of different SATA ports, and when I was able to see it, utilities like Acronis (a Godsend in the last month, bless those people) were unable to read or write from it (trying to backup the drive failed around 30%, then wouldn’t retry in subsequent attempts, so I tried to re-write my last backup to it, which consistently failed between 0 and 50%). OCZ honored the warranty, sure, but it’s made me a hell of a lot more cautious.
Point being that every HD manufacturing process has its failure points, and my general belief from the internets is that your SSD’s generally either DOA (a ticking time bomb) or will experience some sort of wear-related failure after a moderate chunk of time. I don’t really think “next-gen” SSDs have been around long enough to provide a useful comparison to platter-drives, but I’m fairly new to the tech, so who knows.
There’s a couple of calculators on the web for SSD life. Basically, it is more important to determine the number of writes than it is to determine the old hour-based MTBF.
Compiling happens to pile more writes per unit time than other tasks. So, more than a traditional drive, SSD life is dicated by how heavily you use the drive to actually write.
A good 'ol basic system drive with some games on it may experience a very low number of write events, making the SSD last longer than regular drives.
I’m less worried about cell life than I am about the chips as a whole going bad, whether it’s one of the flash chips or the controller or one of the other ICs. I highly the vast majority of failures were wear related.
What Armando said. SSD failures appear to be terminal, where the drive stops appearing in the BIOS.
I have had three SSDs over the past 2-3 years, so I guess my failure rate is similar to what he said. I’m not surprised at all, and warranty returns mean the cost outlay has been low, and like the OP says, the speed difference is so fantastic I’m more than willing to live with this first-gten tech.
I’d like to know more about this too. When the cells start failing is there some kind of warning telling you that the drive is having trouble and you should start getting your data off, does it keep track of failing sections and reduce the size of the drive before dying, or does it just all die with no warning?
After reading that article, I immediately setup Windows for a weekly full backup of my SSD to my 512GB backup drive. I’m bummed. I love my SSD, but that kind of failure isn’t worth it. Though, maybe by the time mine fails it will have been improved/fixed.
I hear this from people but I don’t see it. If you can’t write to the drive because that destroys it’s life span, that means it’s only good for reading. Which means that once you’re booted into Windows and have the apps running that you’re going to use … what good is it doing you? Your swap file and data files aren’t on it presumably, so … ??
People who have had lots of failures: What drives were you using? My suspicion is that the drives that focus on being super-fast, like your OCZ/Sandforce-type drives are going to be prone to dying, whereas ones that are maybe slower but focused on stability, like your Intels or Samsungs, are going to be more reliable.
This thread doesn’t scare me away from SSDs, but I think I’ll go with the smaller 120gb version for now. Like someone said earlier, if I just have Windows, a few apps, and a few taxing games on there, the only thing I’ll lose is a little time reinstalling. All the saved games and settings can go on my D:\ drive.
I take it that it’s safe to use symlinks for the Users folder too, right?
I just bought an Intel 320 160GB for use in my new Thinkpad – this thread is scaring me a bit. I guess if I have problems I can just install the 250GB HD that it shipped with it.
New technology is always a risk, but HDDs fail at stupid rates too compared to how reliable they used to be, and goes in cycles as to when you get higher than normal failure rates. The thing is you need to always back up, and with storage being so cheap and so easy to have an external backup there’s no reason not to.
With the warranties that the SSD drive companies offer the only thing you’re really out is the time it takes to get a replacement, but the speed is so well worth it. I always operate under the assumption that my drive will die at the shittiest time for me, and backup because of that (actually had two die within 2 days of each other, and it was WD HDDs).
Personally I love my SSD and I’d never go back to an HDD for a boot drive.
