Squid

Mover adds in ".partial" during the move to solve various problems that users have on occasion
 
This makes your filename too long.  Right now, your source is 255 characters on the filename itself.  This is the maximum length of any filename.
 
I'd simplify your naming convention that you've got (presumably) sonarr naming the file to.  It seems to be a tad much, having both the [S01E12] and the episode # 012 separate, along with all of the file info x264, audio etc in the filename itself when usually it only needs to be done in the associated .nfo file unless you want to see at a glance everything about the file.

Nothing super obvious, but based on the tasks currently running, see if restarting Shinobi before you go to bed makes a difference.
root 12317 0.0 0.0 0 0 ? Zs 16:05 0:00 | \_ [ffmpeg] <defunct>  

I'd say it's bad because it is currently being corrected by virtue of it being ECC.  IMO, ECC is purchased so that you know when one is beginning to return errors so that it can then be replaced rather than waiting for the module to get bad enough that it can no longer correct the errors.
 
The SEL should tell you exactly which DIMM it is (or mce says that it's DIMM0, but I'd trust the SEL more than what mce says for identification)

By deleting Theme Engine and rebooting do you mean that you uninstalled it and then rebooted?
 
Try in safe mode.  Also try using incognito mode in the browser if an extension is interfering 

Tools - New Config
 
Reassign the drives accordingly.  Since you're not using any parity device for redundancy, that's everything you need to do.

docker inspect nameOfContainer  
Will tell you ever path, every Env, every port that you've got.
 
If you didn't set up environment variables, then easiest to work out the paths and ports from the Docker Tab (Advanced view)

Apr 13 11:48:03 Gardenhill root: nginx: [emerg] cannot load certificate "/etc/ssl/certs/unraid_bundle.pem": PEM_read_bio_X509() failed (SSL: error:0908F066:PEM routines:get_header_and_data:bad end line)  
Try deleting /config/ssl and /config/ssh reboot if still no go, then post a new set

Remove that xmod line you added to go and see if that makes a difference
 

No driver is being loaded.
 
6.8.3 is rather ancient.  Update to 6.9.2 or 6.10.0-rc4+ and see if that fixes it

2 possibilities can cause this:
Certain combinations of hardware will issue an mce upon processor initialization (probably related to firmware update) and can be safely ignored. Your wife has asked you to sweep the floors and you've lifted up the carpet and swept the dirt there. Under either scenario, you're safe (except for the wrath of a woman)

You have a docker container referencing "CamelCase" in it's path mappings

Disable mover logging in Settings - Scheduler
 
Also, generally the appdata share should be set to use cache: Prefer. 
 
Mover keeps trying to move it to the array (and the logs fill up because most of the files are in use)  You will get far far better performance by setting it to Prefer, running mover (with the entire docker service disabled in Settings Docker).  Use the appdata backup plugin to make weekly backups of it to the array

Post your diagnostics

If you're running 6.9.2, then you need to delete the /config/BTRS.key file
 
 

There's 1.3G which the containers have written directly into the image.  A bit on the high side, but still within the realm of possibility.  All in all I'd say just expand to 30G and not worry about it.

Delete it.  I'm not doing an exception for a backup of a backup

super.cfg is not a file which is used by the OS (it uses super.dat)  I'd just rename it to be an extension other than .cfg and check to make sure that nothing else was potentially using it.

This is the workaround here:
 
 

Reseat all the cabling to the drives at both ends.  

Try deleting /config/plugins/corefreq.plg from the flash drive and reboot
 
If that doesn't solve it, then try safe mode

the appdata share vs any other random share is simply a convention and nothing your particularly have to stick with  Change the default appdata location in settings - docker.  This will NOT however change anything already installed.
 
The docker image does not exist on the flash drive, so if you boot with another flash and the image location etc is the same, then your apps will all show up (but won't be editable etc without the applicable files on the flash drive.  This includes references to the icons (Admittedly, this is a scenario I never tested with the various changes to the way icons are handled). 
 
The docker image is very safe to delete (Settings - docker with the service disabled) and then recreate (via Apps - Previous Apps), assuming you did an actual backup of your original flash drive which had the templates stored on it.
 

Just an error from the controller basically stating "Huh! The drive was here and now it isn't"

By using Krusader, you are effectively bypassing the "system", resulting in this behaviour
 
The way that every OS in existence operates on a move is that first the file / folder is attempted to be renamed to the new location.  Only if it fails does the copy / delete process (physical move) take place.
 
Because Krusader is working from the single mount point (/mnt/), the rename will always succeed, resulting in files being location in apparent violation of the rules.
 
Options to avoid this is once 6.10-rc3 is released to install a File Manager plugin at the same time, or 2 do file operations via Windows and move between the shares, or map the shares to Krusader individually and move between those mounts

rename the file to be docker.img instead and try again.  I seem to recall an issue with that.

Which ever gets you back to 6.9.2 Alternatively you can always download the zip file from the webpage (downloads) and extract all of the bz* files from the zip and replace them on the flash drive and reboot.