Kilrah

Don't use 127.0.0.1, that's be the local address of the container. Use the actual network IP of where the service is running.

You need to move unraid's gui to other ports so you can run the NPM container on ports 80/443, just like what was answered to the post you quoted.

Or the Python plugins in CA.

Do you have a monitor plugged to the GPU? If not try one, if it runs better you may need a dummy plug.

Diags are just after boot so don't show any error, but that looks like filesystem corruption, Check filesystem

And you're doing that from windows explorer? Because I just put some dummy files on a share, deleted some with del and some with shift-del, in both cases I got the "delete permanently?" confirmation, and both were kept by the recycle bin on the server. I seem to be having the exclusive share issue though.

Shift-del bypasses Windows' recycle bin but it already doesn't exist on network shares, when you delete whether shift or not you get the confirmation box saying that's permanent. This plugin is separate and acts on the server once the OS has sent the delete command regardless.

Parity is not a backup, it only protects against drive failure, if you delete something parity is updated in realtime to reflect that. Unless you had the recycle bin plugin installed you'd have to know which physical drive the files were stored on, take it out and use file recovery software.

Depends on what you want, either make a redundant pool setup or ignore the unprotected warnings.

This means the share has files on storage that is not redundant, i.e. if you wanted it all green your pool(s) would need to be multi-drive with redundancy.

It's easy to go into dev tools and paste another string in place of an identifier, wouldn't be noticeable in the rendering

Shares always include all array drives and pools even if not assigned in the share settings, those only decide where new stuff is placed. You must have put something in /mnt/Scratch/data yourself at some point.

All you should need is enabling them in config.php. The screenshot is from the settings of the "memories" nextcloud app.

Don't see anything specific to this but there seems to be a "cache" pool defined and used in shares but with no disks assigned to it?!

Make sure there isn't a partial file from a previous copy attempt with the same path/name on any drive, if there is it'll continue to try to put it on that drive. Also to avoid "clogging" a drive the share min free space should be more than the size of the biggest file you expect to transfer to that share... which is a bit of a problem when you're often transferring such huge files.