JorgeB

Cache filesystem is corrupt, you should backup and re-format.

May 8 21:10:17 ThreadRipper19 kernel: BTRFS: device fsid 067b1393-440b-497e-b688-fb11e9c6611d devid 3 transid 645077 /dev/nvme3n1p1 scanned by udevd (2760) May 8 21:10:17 ThreadRipper19 kernel: BTRFS: device fsid 067b1393-440b-497e-b688-fb11e9c6611d devid 2 transid 291831 /dev/nvme0n1p1 scanned by udevd (2792)  
As you can see the transid of one of the devices is way off, it should be the same for all pool members, note the other pool below:
 
May 8 21:10:17 ThreadRipper19 kernel: BTRFS: device fsid 565f4d7d-1b62-4a67-9414-ac108e2553f3 devid 3 transid 31253 /dev/nvme1n1p1 scanned by udevd (2741) May 8 21:10:17 ThreadRipper19 kernel: BTRFS: device fsid 565f4d7d-1b62-4a67-9414-ac108e2553f3 devid 2 transid 31253 /dev/nvme2n1p1 scanned by udevd (2727)  
The difference is so large that it couldn't happen if a device lost a few writes, so likely some corruption happened here, also note that Ryzen with overclocked RAM like you have is known to in some cases corrupt data, so suggest fixing that.
 
You can first try restoring the superblock from backup to see if it helps, do it both both pool devices:

 
btrfs-select-super -s 1 /dev/nvme3n1p1 btrfs-select-super -s 1 /dev/nvme0n1p1  
Then reboot and post new diags.

If it's still slow also a good idea to run a test with the diskspeed docker to make sure all disks are performing normally.

Them I would suggest trying with a new flash drive config, you don't even need a key, just to see if you can access the server that way, if yes it's a problem with the old config, and you can then reconfigure the server or try copying a few config files to test and identify the culprit.

Free space with an odd number of devices won't be correct, it's a btrfs issue, but the total space is usable and the free space reported will get better as you fill the pool.

Make sure nothing is accessing the disk during the diskspeed test (or parity check), and if it's still slow it's the disk.

Problem with the onboard SATA controller, quite common with some Ryzen boards, BIOS update might help or use an add-on controller.
 
You need to reboot to see actual array status, there might be a disabled disk.

Not while it's still assigned to the array, you can unassign it, start the array without it then format the device directly (not the emulated disk) with UD or using the CLI.

Click on settings in front of the disk you want to enable that:
 

 
Then enable "Share"
 
P.S. next time please use the existing plugin support thread.

This, you don't even need the same order, just start the array without any pool device assigned so the config is "forgotten", then re-assign the original members only, not the new one, start the array and Unraid will import the existing pool, then add new device.

https://wiki.unraid.net/Manual/Storage_Management#Removing_disks_from_a_multi-device_pool
or
https://forums.unraid.net/topic/46802-faq-for-unraid-v6/?do=findComment&comment=480418
 

Click on the first pool device, then scroll down until this section:
 

 
Click on "perform full balance", chose the profile you want to convert to then click on "Balance"

There's filesystem corruption on the emulated disk, once you fix that by running xfs_repair without -n it should mount and you'll see your files, I don't know how to explain that more clearly to you.

That is to fix a corrupt filesystem, not a disable drive, disks look healthy but diags posted are after rebooting so we can't see what happened, to rebuild the disks you can see here:
 
https://wiki.unraid.net/Manual/Storage_Management#Rebuilding_a_drive_onto_itself
 
P.S. LSI is using the MegaRAID driver, that's not recommended, see if it can be flashed to IT mode.

If it's a possibility I would return it.

Not sure about the error but you started the array with only one pool device assigned, so just need to wait for the balance to finish, which it already did by now, so you can just add the new device.

That controller uses SATA port multipliers, those are not recommended, you can cancel the rebuild, replace the controller then re-start rebuilding the disable disk.

Make sure UEFI boot is enable in Unraid, if the flash has a EFI- folder rename it to EFI

Seems unlike, other than trim possibly not working they should work fine, unless there's some firmware/compatibility issue, and it's using a very old firmware, so you should update to latest.

You can use the other Ethernet ports, but it must be set as eth0, you can just delete network-rules.cfg from the flash drive and reboot.

Apr 20 05:36:49 Tower kernel: ahci 0000:00:17.0: Found 2 remapped NVMe devices. Apr 20 05:36:49 Tower kernel: ahci 0000:00:17.0: Switch your BIOS from RAID to AHCI mode to use them.  

Run it again without -n or nothing will be done, if it asks for -L use it.

Some data corruption is being detected causing the balance to abort, you can run a scrub to list the corrupt files in the syslog, then delete/replace those, also a good idea to run memtest.

It's known issue with the scrip and current releases, it should not be used, you can still do it manually:
 
https://forums.lime-technology.com/topic/61614-shrink-array-question/?tab=comments#comment-606335
 
 

Drives might be formatted protected see below:
 
https://forums.unraid.net/topic/93432-parity-disk-read-errors/?do=findComment&comment=864078
 
if that alone doesn't help see also here:
 
https://forums.unraid.net/topic/110835-solved-help-with-a-sas-drive/?do=findComment&comment=1011689