Hi everyone,

I wasn't able to take care of this issue before, but it makes three weeks already 😫

It started with an issue Disk5 and I'm pretty sure this a cable/controller issue, but then I messed up.

I replaced the failed drive with a spare one, but during the rebuild, I tested the drive and cable. It froze the server and I had to reboot (after waiting a while)

As a result, the 2 parity drives AND another disk failed !!!

Can anyone confirm the disks are okay ? there is no SMART error, I'm quite certain the data is still there (most important data is backed up).

My plan is to test every disk individually, every sata cables with and without the SAS expander.

Should I try resetting the config ? I can't rebuild the array as there are more than 2 "failed disk"...

In the mean time, I'd like to "export" some of my docker container to a VPS (eg. Vaultwarden, immich, radicale...). I have a backup of my appdata share and some container specific shares (immich) but I don't really know how should I proceed. Anyone could tell me the main steps ?

Thank you so much for you help !

kus-diagnostics-20260121-1814.zip

It won't disable more disks than you have parity. The problem seems to be it doesn't recogize either of your parity disks, so you have too many missing disks to start the array.

Post a more complete screenshot of Array Devices.

Thanks, I'll try to restart and take a screenshot again

Thanks for your response.

A reddit user suggested the problem was from RAM errors. I ran 10 tests of memtest82, but all were fine (10 pass / 0 errors) ! Maybe the ram needed some holidays ^_^

I really don't know what caused the problem, but after shutting it down for a few days, the 2 parity disks were good, so I swapped the Disk 2 and Disk 5 with 2 spares I had... The parity rebuild went fine (took 8 hours). So far so good. Here's the screenshot you asked, even though it doesn't seem relevant anymore.

I checked the 2 "faulty" disks on my laptop using an USB dock, and there doesn't seem to be any issues with the disk. All the data is still there and SMART values are good too (even when marked as "failed").

Thanks again :-)