Sandforce sucks
Last year (almost exactly 1 year, to the day), I bought a 120gig OCZ Sandforce-2 hard drive. On Thursday, I came home from work to find my system completely borked thanks to a bug in the drive's firmware that causes it to crash if your computer either goes into power-save mode at the "wrong" moment, or the power goes out at a "bad" time. The net effect is that I can boot up, and have almost exactly three minutes to read data from it before it crashes and quits responding.
The worst thing is that normal recovery tools (like Gnu ddrescue) don't work, because they make the borked firmware crash INSTANTLY.
The thing that p*sses me off more than anything is the fact that the data is there. It would be utterly TRIVIAL for someone like Sandforce to give us the option of flashing emergency recoverymode firmware onto the drive that we could use with a special version of ddrescue to rip the raw data and reconstruct it offline. But they won't, because they either can't be bothered, or (more likely), the way they implemented their drive-level security has some trivial to exploit flaw that they know about, and to protect it they make the drives intentionally commit suicide instead (google "Sandforce" and "Panic Mode" for the gory details).
The worst thing of all is the fact that RAID won't even necessarily save you. You can have two mirrored drives, and have them both fail the exact same way simultaneously due to this bug.
I wish I'd never bought that drive. The hardware itself is probably fine, but Sandforce's stupid firmware ruins it into uselessness. The fact that you can't even protect yourself from its main failure mode via RAID is the icing on the cake. OCZ itself isn't much better. Their hands might be tied, but they're officially determined to pretend that there's nothing wrong, and have employees on their forum working overtime to try and make the problem look less widespread than it really is. They keep emphasizing that the drive itself isn't physically destroyed, while trying to sweep the fact that your data that was on it is (as far as they & Sandforce are concerned) gone, and blaming their users for not doing complete hourly backups.
Put another way, if you value your time and sanity, DO NOT buy a SSD with Sandforce controller. The performance boost you see will end up getting completely destroyed 200 times over when the drive's controller f**ks itself and you have to spend a week or more in limbo recovering from its failure.
Grrrr. I feel a new special-interest obsession coming on, and it's not pretty...
_________________
Your Aspie score: 170 of 200 · Your neurotypical (non-autistic) score: 34 of 200 · You are very likely an Aspie [ AQ=41, EQ=11, SQ=45, SQ-R=77; FQ=38 ]
Well, I'm still pulling files off of it, 3 minutes at a time.
The problem isn't that the drive is going to sleep. The problem is that the computer went to sleep in a way the drive didn't like, and the drive's firmware corrupted its control table in a way that causes its internal firmware to crash ~3 minutes after startup (rendering the drive effectively unusable until the power gets reset).
I knew these drives had "issues", but I had no idea it was anywhere close to being this bad. If you want to get really scared, take a look at the forums at ocz.com, and count how many new people in the Vertex2/Agility2 forum *alone* have their drive die for the first time every single day. Then, get angry at OCZ's dismissive attitude that there's "nothing wrong", because "all" you have to do to "fix" the drive is erase and reformat it.
If a company like Seagate or Hitachi had drives failing and corrupting themselves this badly, they'd be out of business. I've read about guys who've bought several drives over the past year or two, and literally had EVERY SINGLE ONE OF THEM corrupt themselves at some point within a year. Sandforce is crap. They look good in benchmarks, but they're absolutely suicidal for real-world use as hard drives.
The worst part of all is that there's nothing you can do to really protect yourself that won't ultimately nuke most of the reasons for buying the drive in the first place. I'm probably going to end up buying a pair of 600-gig Velociraptors to run in RAID-1 config, and just use my 120-gig drive for swap and as a write-through cache (where everything gets written to the drives and SSD, then gets read back from the SSD). I have zero faith in this piece of sh!t drive anymore. If it wasn't so expensive, I'd chuck it in a drawer and never taint my computer with it again. The speed was nice, but I just can't deal with a drive this brittle and unreliable. So far, it's totally f**ked a week of my life, I'm still limping along with just my laptop, and will probably still be limping along with my laptop for at least another week or two.
_________________
Your Aspie score: 170 of 200 · Your neurotypical (non-autistic) score: 34 of 200 · You are very likely an Aspie [ AQ=41, EQ=11, SQ=45, SQ-R=77; FQ=38 ]
OCZ has serious quality control issues.
I bought a couple sticks of OCZ performance ddr ram from newegg, and the SPD eprom didn't match the sticks I bought. OCZ didn't admit to their guilt on their forums, but they did tell me the components were exactly the same on the sticks I did buy.
Make a long story short, OCZ provided a spd reprogrammer a long time ago which isn't on their website anymore but you can still download it from various websites. I finally got the nerve to flash the spd chip with a copy of the right code and now they work flawlessly.
I know it's sorta off topic, but it's a testament to OCZ's quality control. I think that's known as fraud in the legal system. Selling something branded as one product when it's really another. If they owned up to their mistake I wouldn't have said a word.
For clarity, the memory I did buy were supposed to be model number OCZ4002048ELDCPE-K, and they are labeled as such even to this day, but they were programmed as OCZP4001G modules.
After this fiasco I doubt I'd buy anything else from OCZ ever again.