Looks like my WDCWD20EADS drives are aging prematurely because of the green technology. They spin down after 8 seconds apparently and in a Drobo, this means they keep spinning up and down and effectively wear out faster. There are a few posts on the forums about this on the web. So, the moral of the story may be to avoid WD green drives for Drobo or other RAID type applications.
Have you found any alternative for the green drives without replacing them? Perhaps forcing a longer time period before spinning down? Also, how do you monitor this with the drives in the drobo? I have a very similar configuration as you, using the WD20EADS drives. While I haven't seen any problems yet, I want to start monitoring to see if I'm experiencing similar issues, indicating a possible failure coming.
Posted by: Davedraeger | February 21, 2010 at 10:34 AM
There is supposed to be a widdle3 utility which can change the settings for head park and the error recovery maximum time but I'd need to connect the drives internally which I can't do with an iMac. Drobo doesn't expose any smart metrics to see whats happening with the drive. The Drobo is still rebuilding and the docs say don't turn it off in that state. For something which takes days, that kind of warning is pretty scary.
The Drobo S doesn't look an better, same basic issues, under powered and poor admin software/diagnostics. I'm probably ditching this and buying a proper NAS like a QNAP 459 or something. Faster and they have a console that can display something besides a red pie chart... I'll attach it to my iMac using ethernet and using iSCSI.
I'll probably move to enterprise SATA drives also and cycle these out over time.
Posted by: Billy Newport | February 21, 2010 at 06:03 PM