JorgeB

Cache is complaining of a missing device, and at the time this happened it wasn't fully redundant, so while you should be able to mount it read only most likely some data will be missing, still you can try some recovery options here.

Most files are empty, including the syslog.

You'd need to backup the data first as the formatting wipes the disks.

Unlikely that you'd be able to boot from a flash on an ad-don USB controller.

Your are overclocking the RAM for the CPU you have, with Ryzen it's known to in some cases corrupt data (resulting in sync errors), see here.

Yes, backup current flash first, it's safe to restore super.dat (disk assignments) and your key, both on the config folder.

Like mentioned we need the diags before rebooting.

Can't see which device, try doing the same as this one.

Yes, that's a bug, note that only 480GB can be used.

Good theory but autodefrag is disabled by default for any btrfs mount, you'd need to use that option at mount time to enable it, and it's not used by Unraid.

No idea, if you read the rest of the thread you see that the command that was going to be used doesn't work for everyone.

Then best not to use the mover, it's not efficient for many small files, use cp, mc or your favorite copying tool.

Best bet it to backup current data on cache, re-format the pool and restore the data, some recovery options here if needed.

Is the current cache btrfs? With the btrfs you can upgrade a device with the pool online. https://forums.unraid.net/topic/46802-faq-for-unraid-v6/?do=findComment&comment=480419 If xfs a dd or similar clone won't work since the partition won't use the new device full size and Unraid will complain.

It's likely one of the installed PCIe devices, you can remove one by one to try and find out which one, then try it in a different slot if possible.

I have no idea if your board has 4 or 6 PCB layers, you can try to find out or just use 2667 for now, and if it's stable then you can try higher speeds.

Try installing a docker.

RAM is @ 3200Mhz, max supported speed with your CPU is 2667/2933Mhz depending on the board.

Cool, thanks.

Try booting in safe mode.

Flash drive continues to drop: Jun 13 23:40:43 plex kernel: usb 1-3: USB disconnect, device number 2 Jun 13 23:40:43 plex kernel: print_req_error: I/O error, dev sda, sector 2049 Jun 13 23:40:43 plex kernel: Buffer I/O error on dev sda1, logical block 1, lost async page write Jun 13 23:40:43 plex kernel: print_req_error: I/O error, dev sda, sector 31413 Jun 13 23:40:43 plex kernel: Buffer I/O error on dev sda1, logical block 29365, lost async page write Jun 13 23:40:43 plex kernel: print_req_error: I/O error, dev sda, sector 468180 Jun 13 23:40:43 plex kernel: Buffer I/O error on dev sda1, logical block 466132, lost async page write Jun 13 23:40:43 plex kernel: usb 1-3: new high-speed USB device number 3 using ehci-pci Jun 13 23:40:43 plex emhttpd: error: put_config_idx, 619: No such file or directory (2): fopen: /boot/config/shares/Media.cfg Jun 13 23:40:43 plex emhttpd: error: put_config_idx, 619: No such file or directory (2): fopen: /boot/config/shares/appdata.cfg It's either a flash problem, a USB/board problem, or someone is removing it of the server.

The only way I can think of this being a configuration problem would be if you were using VMs and by mistake passing-though the USB controller where the flash drive is to the VM, other than that this is a hardware problem, flash drive would be the prime suspect, but since you already tried a new one... I guess you could try a different one before trying a different board.

Since the diags are after rebooting we can't see what happened but almost certainly not a disk problem, if the emulated disk1 is mounting correctly rebuild on top and if it happens again try to grab diags before rebooting, number one suspect would be the AMD Ryzen SATA controller which as known issues with older kernels, upgrading to v6.9-beta1 should help if that is the problem.

Best bet is to use the existing plugin support thread:

Possibly if you post the complete diagnostics.