Hi all, wanted to reach out about a recurring problem i've had with my server.  Occasionally i'll get Parity disk failures that show with the disk having *trillions* of reads and writes (see attached).  Previously i would stop the array, remove the parity drive, start the array, stop it again, re-add it, and the parity rebuild would work without an issue.  A couple months later, the same thing would happen.  

 

Fast forward to this week, it happened again, and i thought "okay this drive is probably on it's way out".  I hopped off to best buy, grabbed a new drive, popped it in, precleared it (which went through without issue), and added it to the array as the new parity.  During the parity rebuild, the exact same thing happened with the brand new drive.

 

Previously i've tried moving the drive to another bay in the chassis (it's a supermicro 24-bay) but it doesn't seem to make a difference.

 

Has anyone seen this before?  What are the next troubleshooting steps?  The attached screenshot is for the brand new drive.  I've also included a diagnostic packet.

feezfileserv-diagnostics-20201230-1037.zip

Check the cables/connections, and make sure you don't have too many drives attached to one power line back to the power supply

I'll double-check this but i'm not certain it's a power issue.  It's a 1000w power supply (Overkill i know but i had it laying around) going to a backplane with 5 power inputs, i have 3 on one line and 2 on another.  When the parity was initially having issues i swapped it to a location that would've been powered by the line it wasn't initially on and the issue persisted.

If it's only dropping during high stress loads, I'd start with power. Even a slight sag will reset the drive

Good point.  I just double-checked and it's actually 6 connections back to the PSU, 4 are plugged into one line and 3 into the other - maybe switching it to 3 and 3 will help.