0

SERVER: Dell T410 Raid Controller: Perc 6/i

What is the best way to diagnose root cause of failed drives?

Back story:

So I had a 4-drive Raid 10 + 1 hot spare setup. This setup worked for years, about 3 months ago, one drive died, and the hot spare took over. In the past week another drive failed, replaced it, and another failed during the rebuild. We lost data.

With 4 drives remaining, I created raid 6 and restored from a backup. I let the server run over the weekend and 3 more drives have failed.

  • When the server was running, OMSA showed that the raid battery is dead. I plan on replacing it, but as far as I know, that shouldn't cause of my issues, perc 6 runs in a special mode, and we have a UPS backup. Ran a self-check on power-supply (pressing hardware button - led shows green).

All these drives are refurbs...but I have a hard time believing 6 drives went bad about the same time without some external cause. Looking for direction to diagnose. My guess is it's the power supply or raid controller - but how to best diagnose?

Improve this question
5
  • 1
    All these drives are refurbs - Honestly, that would be my first suspicion.
    – joeqwerty
    Nov 28, 2016 at 18:25
  • @joeqwerty me too...I'm just worried to get new drives and kill them if that's not the case. The fact that I just added 2 drives in (supposedly tested), and then 3 drives fail is weird.
    – webaholik
    Nov 28, 2016 at 18:27
  • Your best chance of diagnosis if you want root cause would be to export the controller logs using OMSA. That will give you event history including SENSE key events from the drives, and might give you an idea of whether there's some kind of communication problem, or if they're simply experiencing "normal" failures.
    – JimNim
    Nov 28, 2016 at 20:13
  • What did SMART attributes show on bad drives?
    – Mikhail Khirgiy
    Nov 29, 2016 at 5:46
  • @MikhailKhirgiy SMART didn't indicate any errors, ended up purchasing all new drives, everything is working great
    – webaholik
    Dec 1, 2016 at 0:05

0

Reset to default

You must log in to answer this question.

Browse other questions tagged .