jgreco
Resident Grinch
- Joined
- May 29, 2011
- Messages
- 18,680
That along with a lot of other horror stories on their drives. Not to mention that it was pretty short lived. I am used to drives lasting ten years and beyond, I just sent off an older 120GB IDE Seagate drive to my father after running badblocks and smart tests and I know it was at least 8 years old as my wife had bought it before we got together. I know sometimes drives will go bad but I rarely have a high turnover rate. I still have drives that are in the 10 to 20 GB range that have spun for many years that still work fine if I need them for something.
And there's lots of horror stories about Western Digital drives - something that's gone on for years:
http://www.macobserver.com/tmo/article/french_retailer_data_offers_ssd_failure_rates/
Note this isn't actually a comprehensive failure rate list, it's JUST drives that were RMA'd to the retailer. Presumably more were returned to the manufacturer directly for service, or failed after the retail RMA window.
Look at the WD Caviar Black 2TB with its nearly 10% RMA rate, or the Hitachi Deathstar at almost 7%, or the WD Green 2TB at almost 5%. Seagate's first entry is at 4.35% with the Barracuda LP, which looks tepid by comparison.
This is all very consistent with our local experiences; I've got a pile of dead EARS drives that fairly consistently failed shortly after the warranty expired, and I knew a company who had made ZFS arrays out of WD Blacks and actually lost a pool because of a manufacturing run defect (IIRC) that killed several drives in rapid succession. Just helped reinforce my belief that heterogeneous arrays are a great thing.
So, again, I ask, what are you suggesting? The drive manufacturers all suck. Best advice is to plan for failure and go for massive redundancy (and I like heterogeneous arrays as well), along with proper burn-in testing.