JorgeB

Start shutting everything else down, one by one.

Yep, most likely some board compatibility issue with Linux/Unraid.

If you continue to get read errors, and it's not a slot issue, best to replace it, after you might try it in a different computer if available.

On the diags posted there's something writing to disk2.

right.

Yes.

Most if not all should be fine for now, you can run an extended SMART test on all to confirm.

With single parity you can only replace one at a time, but possibly all those warnings are not critical, you should post the diagnostics: Tools -> Diagnostics.

I would set the RAM cache to no more than 4GB.

Same SASLP controller? SMART report is still not correctly generated, you can try a different controller, but best bet would be to replace the SASLP with one of the recommended LSI controllers.

Macvlan call traces are usually the result of having dockers with a custom IP address, see below: https://forums.unraid.net/topic/70529-650-call-traces-when-assigning-ip-to-dockers/

It means to VM will start as if recovering from a crash/power loss.

If it's not appearing in lspci it can only be a hardware issue, try the 3090 on it's own or on a different slot.

Auto check after a power loss is non correct. Should be fine. Yep.

CRC errors are a connection problem, just acknowledge them and as long as it doesn't keep increasing your fine, if it does start by replacing the SATA cables on the affected disks.

You can run Memtest from the Unraid boot menu.

Kingston DTSE9 (first gen USB 2.0) is known to work and also be reliable.

Use a trial flash first, without your config, just to see if it boots.

Jan 18 12:37:23 Tatooine preclear_disk_S3YJNS0N704931R[7562]: Signature: verifying unRAID's signature on the MBR ... Jan 18 12:37:23 Tatooine preclear_disk_S3YJNS0N704931R[7562]: Signature: Unraid preclear signature is valid! Jan 18 12:37:24 Tatooine preclear_disk_S3YJNS0N704931R[7562]: Post-Read: post-read verification started (1/5).... Jan 18 12:37:24 Tatooine preclear_disk_S3YJNS0N704931R[7562]: Post-Read: verifying the beginning of the disk. Jan 18 12:37:24 Tatooine preclear_disk_S3YJNS0N704931R[7562]: Post-Read: cmp /tmp/.preclear/sdb/fifo /dev/zero Jan 18 12:37:24 Tatooine preclear_disk_S3YJNS0N704931R[7562]: Post-Read: dd if=/dev/sdb of=/tmp/.preclear/sdb/fifo count=2096640 skip=512 iflag=nocache,count_bytes,skip_bytes Jan 18 12:37:25 Tatooine preclear_disk_S3YJNS0N704931R[7562]: Post-Read: fail - beginning of the disk not zeroed This time the MBR was valid but the disk wasn't correctly zeroed, or is now returning non zeros, since it still happens with a different device it points to a hardware problem, like bad RAM, controller, etc.

Disk2 dropped offline so there's no SMART, check connections and post new diags, also note that you're using a SASLP, those are not recommend for a long time due to several issues, including dropping disks without a reason.

Unlikely, looks more like a board compatibility problem, you could try a different flash drive.

See here: https://forums.unraid.net/topic/46802-faq-for-unraid-v6/?do=findComment&comment=819173

Turn off what? If you want that data on the array you need to set cache to "Yes" for those shares and run the mover (docker/VMs services need to be stopped first).

Basically this is the problem: Jan 18 00:22:45 Tatooine preclear_disk_WFL33WEX[8763]: Signature: failed test 1: MBR signature is not valid, byte 0 [10741] != [00000] Disk is returning different data than expected, this also why you get "invalid partition", Unraid checks for the MBR, can't tell you why though, try a different disk.