Squid

Docker apps: You'll have to readd them and set up the paths, etc again. The actual data within the apps (eg: Plex's database, the mySQL database etc) will all still be intact as its stored on the array (usually the cache drive) Plugins: You'll have to re-install them. Generally most plugins have very little set up that you did to them. If you can still read the contents of config the config folder on the original flash, then you can just copy that to the new flash, and everything (plugins and docker) will be there ready to go (just don't copy over the replacement key file) If not, you have a screen shot of your disk assignments - redo them, your shares will be automatically picked up (Although if you had security on them you'll have to redo that) Net result is that even on a complete redo of the flash drive, setting the system back up the way that it was is fairly painless and won't take that long.

No idea. I use winrar for everything

I've been running it on my production server with no problems. I don't however use the new features of encryption / SSL access though (don't trust them yet in a production environment )

The rsync flags are fine for everything. Only thing I can possibly think of is something getting lost in translation between the source NTFS and the destination XFS/BTRFS. Normally I'd say bite the bullet and start it all over again and see how it goes, unless your custom artwork / watched status / etc is very important to you. If the latter, wait for the developers of Plex to give you a response.

What flags are you using with rsync?

Yes, but that's if the cable doesn't make any contact. If the cable makes intermittent contact, then you'll have problems during a rebuild. Entirely up to you, but as @bjp999 stated, loose / intermittent cabling is the #1 problem on any computer. Just a bad design of a cable, and affects all computer systems equally.

This. Since your original config was from a Windows version, there may be slight incompatibilities between the two versions that are causing all of your issues. Delete the existing appdata folder for Sonarr (or give a new install a different folder to write to)

Spin up/down wouldn't have affected this. This is purely a function of the drive itself.

disk.ini

Probably. I just did this, and one share disappeared immediately. The others took up to 10 minutes to disappear

Many reasons for the docker image to get filled. The Docker FAQ covers it.

Tons of hits for doing this over the network from Windows / Mac

Any background process can create any file it chooses within /tmp whenever it wants to with adverse effects if its not there when it needs it. Just delete the dup* files

Its more than that. What it appears is that the script by @jbartlett makes extensive use of mdcmd to determine everything (which is now deprecated). I can't get it to even test and come up with the correct drive being tested by selecting a single drive. The graphs, not quite sure by may be related. I do know that he is working on a plugin himself that is supposed to be far better than this hack on his rather old original script. What I'm going to do is mark the plugin as being incompatible with everything over 6.4RC7a and wait for the replacement.

If the share is set to not use the cache drive, then it doesn't matter if you use /mnt/user or /mnt/user0 Generally you reference everything as user to make everything simpler to understand. Just set the share to be use cache: no

Is the stick formatted as FAT32? Try the USB creator tool https://forums.lime-technology.com/topic/59025-unraid-usb-creator-tool-version-01-beta/

Why does ARRAY STARTED STALE CONFIGURATION appear at the bottom of the webUI? See here: https://forums.lime-technology.com/topic/59845-unraid-os-version-640-rc8q-available/?page=11&tab=comments#comment-588602

Are you running the tail via SSH (putty) or directly on the local monitor / keyboard?

Its all about copying from /mnt/diskX to /mnt/user (or user0) and vice versa (ie to/from a disk already assigned to the array to/from the user share system). That will corrupt the files. That's fine. Specifying /mnt/user0 forces the files to go directly to the array instead of going to the cache drive if its enabled for the particular share. unRaid has rules to deal with that. Generally, the version on the lowest numbered disk takes precedence, but it can certainly be confusing. If you only copy to /mnt/user, then this situation will never happen. The Fix Common Problems plugin when running extended tests does look for duplicated files. 0777 / 0666 and root:root are perfectly fine. Both have identical results. Either way if you do use either the CLI or the utility, ensure you don't run them against your appdata share if you have docker containers. You may impact there ability to run. If in doubt, Fix Common Problems includes "Docker Safe New Permissions" in the Tools Tab.

When you add shows, you have it search in whatever mounts you've passed to the app for them. Pass through both /mnt/user/... and /mnt/disks/... and add TV shows from each.

There're all page allocation stalls. https://forums.lime-technology.com/topic/59858-trace-error-found-635/?tab=comments#comment-587518

After selecting a custom activity stream, assuming that there's a mistake in the filters when creating the stream, when changing the filters you're given an option to save the stream with the updated filters. That no longer works.

so it's showing the wrong series in the logs? Implies you selected the wrong series when you added it, unless I'm reading your post wrong

Is the keyboard on a passed thru Usb controller?

Pretty much. unRaid is hardware agnostic, so it won't particularly care so long as all the drives are recognized. Doesn't hurt though to take a screen shot of your current drive assignments just in case. They won't care either If you're using hardware passthrough, then you will have to adjust the VM settings