[SOLVED] (6.8.3) Billions of writes/reads in seconds on two drives - Related to ACS override? Dead controller?

[TurkeyPerson]

So i've been having 2 drives basically die together constantly and both happen to be attached to a SATA pcie card. The two drives get a billion read/writes and errors suddenly, but as far as I can tell the SMART reports look fine. I think it might be getting messed up again each time I load a VM - but i'm really unsure. I tried switching the PCIE slot but it happened anyway. Could the controller be dead (it's only ~2 years old I think), should I try some different ports? 

 

To make matters worse, at some point I tried to revert to an older Unraid version using UnraidDVB plug in and had to restore everything by just keeping my config folder - i'm not sure if that could be part of the issue so i'm mentioning it.

Thoughts? 

(For clarity the screenshot below was when this happened after I was rebuilding parity from this happening the first time. Shutting down server till I know what's going on.)

EDIT: OH! And I checked for loose cables and replaced one of them just in case - did not help.

tower-diagnostics-20200730-2309.zip

Edited July 31, 2020 by TurkeyPerson

[JorgeB]

You are passing through the Marvell controller to both the Ubuntu and Windows VMs, so as soon as you start either that controller will no longer be available to Unraid, dropping both devices.

[TurkeyPerson]

I didn't realize it was passed through. I'll check that out. I guess I probably did it in the XML and then changed the slots. Very helpful information, thank you!

Edited July 31, 2020 by TurkeyPerson

[TurkeyPerson]

Update for readers: yep it was in the XML only, hidden away. Thanks again!