RMA
A few weeks ago, ANOTHER SSD silently died. I just submitted my RMA ticket. I thought I’d repost my review of the OCZ Vector.
Pros: This drive is insanely fast – like an extra 256 GB of ram. If they made a cheap ~32GB version, I’d consider buying one to use as a dedicated swap drive. It looks fantastic, and generally radiates high build quality. Scores an “Absurd” in the windows experience index “Disk data transfer performance” category. Sometimes I even think I hear a faint grinding sound – as if OCZ installed a tiny speaker to mock their mechanized magnetoresistive brethren.
Mine challenged The Most Interesting Man in The World to a race. Guess who won?
Cons: I’ve had two fail (I’m about to RMA the 2nd) on me in 4 months. Both gave zero warning. Both failed silently. Both failed absolutely.
Drive failure is inconvenient, painful, annoying, and frustrating – even if backed up. However, it is far more frustrating (for me) when I haven’t the slightest idea WHY.
Both drives suddenly became undetectable – unresponsive to commands. At least with a HDD I can hear the read head stuck or something otherwise horrible happening to that very delicate disk inside. At least, even if the drive is FAR beyond repair, it may emit a dying error message. At least, if its minor, I might be able to debug myself with a logic analyzer and a soldering iron.
OCZ, couldn’t you embed an array of SMD LEDs somewhere to display an error code? Or maybe even a *GASP* JTAG port of some sort?
If I understood WHY these drives were failing, maybe I could understand the incredible difficulty of engineering a solid state drive to a truly reliable standard. I would be more than sympathetic if I discovered that OCZs engineers were trying to tackle some great unsolved problem in embedded systems design – I’d be curious, and hope the engineers who solve the aforementioned problem win awards & are remembered therefor.
Oh, and this second drive failed before I finished backing up to crashplan.
Other Thoughts: The OCZ customer service department is fantastic. Their support forums are filled with dedicated employees who do all they can to help revive a presumably failed drive. The RMA process is very smooth. I thank them for their past work and in preemption.
For those interested in SSD reliability, I highly recommend “Understanding the Robustness of SSDs under Power Fault” by Mai Zheng & Feng Qin of The Ohio State University; Joseph Tucek & Mark Lillibridge of HP Labs.Lastly, if you plan to buy (…and replace, and replace) an SSD, I’d recommend treating it like BTRFS – expect surprise data loss. The best thing to do is to implement an automated daily/weekly drive imaging system – something I’m planning for my next try. After all, the third time is the charm 😉
If you’re interested, I highly recommend checking out that paper (mirror).
Mirror of other paper versions: SSD robustness mirror
Alexander Riccio 