Failed Drive Green Again - Now What?

Three days ago I got a failed drive notification on my Gen-2 Drobo (bay 2). After restarting my computer, the Data Protection process started. It took two days. Now all four drives are green again. I don’t know what to trust. Should I replace the drive that was previously blinking red as a failed drive?

You may have a failing drive, I’ll recommend you replace the drive previously shown as red.
But before doing so, check the usage on your Drobo unit. Try reducing your data to about 70% (if not already) if you can before replacing it.
Only replace it when the lights are not showing orange and green; not in the rebuilding phase.

Thanks for the reply Kelvin. I’ll take your advice. Thankfully, I did reduce the amount of media on the Drobo before the rebuilding process started. Pretty much, as soon as the drive went red.

Hello,

Do you know about disabling the auto-park on those Green drives? If you do not do this, the drive will “Park Itself to death” in about 9-11 months, give or take. WD has a utility that will do this, it increases the park timeout to 300 seconds, instead of 6 secs.

David

I have never heard of this. What is “part timeout”, and why do you want it longer rather than shorter?

Hello,

2 things to do when using ANY drive in a RAID array.

#1) Disable parking, let the RAID controller handle the power savings. On WD green drives, they park (not PART) every 6 seconds, and with a 300k parting MTBF, that is not a long time, can be as little as 18 months. So, the thing to do is use the WDidle3 utility, from the WD site, and disable parking or set it to 300 seconds instead of 6 seconds (default).
Here is an article about this:
http://lime-technology.com/forum/index.php?topic=21007.0
WD’s stance on parking disable:
http://wdc.custhelp.com/app/answers/detail/a_id/5357/session/L3RpbWUvMTMxNTYzMzQwMC9zaWQvNmwxX1ZJRGs%3D

#2) Disable error reporting, and again, let the RAID controller handle this. For WD drives, they call this TLER (Time Limiting Error Rate). Here is an article about that:

http://en.wikipedia.org/wiki/Error_recovery_control

If you don’t disable TLER, the Drobo can mark a drive as bad when it actually is just doing it’s internal drive correction, as well as any RAID controller. I have heard Drobo firmware can handle this, but why take the chance?

For the parking, if you don’t disable it, you are just wearing our your drive prematurely. A standard, non-green drive usually waits about 3-5 minutes of inactivity before putting the drive “to sleep”, aka spinning down the motor and parking the heads. What WD did on green drives was increase the time to sleep and park to 6 seconds. The problem is, drives park 700-1000 times per day, instead of much less when set to a few minutes. The extra parking causes wear on the drive’s arm motor, which it would not be doing if parking was not so frequent. WD themselves rates the number of parks before a fault may occur at 300k, which if you divide 300k by 1k that is 300 days, less than a year if running the drive 24/7. I have had WD green drives running for over 3 years without any errors, but I disabled TLER and parking from day 1.

Enjoy!