JonathanM

Where, exactly, are you changing this?

That's my vote.

Why so large? Normally a VM would be much smaller, with most of the data it needs to use stored on other drives or pools so all VM's could access it. Ideally each VM should be as lightweight and purpose built as possible.

Attach diagnostics to your next post in this thread.

Only bridge mode has any meaning with port assignment and redirection. Since rutorrent is routed through the gluetun container, there are no ports redirected to the host, everything flows through gluetun.

That will work, however, it's explained more fully here. Obviously it doesn't only apply to WD. https://forums.unraid.net/topic/73717-solutions-shucked-wd-8tb-easystore-33v-pin-issue/

Make sure you have good plans in place if there is a leak or a pump failure. Things can rapidly get out of control if either of those things occur, generally you have seconds to minutes to respond to keep from damaging components, so unattended water cooling needs special attention and monitoring.

Read the help description on the cache share settings. "yes" means write new files to the specified pool, and move them to the array on schedule. You want "prefer", so the data is written to the pool and stays there. You will probably need to disable the docker service (NOT JUST STOP ALL THE CONTAINERS) then run the mover manually after changing the setting.

Easiest solution is to temporarily set up a trial version of Unraid on another machine and mount the drives there.

NPM is intended to proxy OTHER servers, not host a site. The SWAG container is a combo package that has tutorials for proxy of other sites, but it also contains a capable NGINX install with built in certificate management from letsencrypt that is easy to set up with a standalone website. It's what I use, and the folders it maps to appdata contain the common folders that you will find in any NGINX website tutorial. If you don't want to use SSL for some reason, there is a plain NGINX container from binhex that works well.

USB is problematic for array drives, inconsistent identification and random disconnects are just a couple of the issues that are prevalent with USB enclosures. If you can't use internal SATA connections, the next best option is SAS, it's much more suitable for external drive connections. Some people have successfully used USB, but figuring out which controllers and drive bays work at all, let alone consistently, is a real problem.

Diagnostics taken after the error has occurred may hold a clue.

While we are discussing the new notifications viewing system, I'd like to propose a small tweak. Instead of only changing the color of the bell, why not change from a green bell to a yellow exclamation point to a red asymmetric star, like a cartoon splat or crash symbol? That way colour blind people can still tell at a glance which type of notification is showing. The highest level of unread notification would determine which icon was shown on the header. (BTW, I find it amusing that last year's "ditch ALL the color and go to flat white icons" is slowly going away)

Can't speak specifically to that card, but typically RAID cards with batteries are there to keep power interruptions from corrupting data in flight, as they typically have a not insignificant amount of RAM on the controller for caching writes. If the card is in IT mode and all transactions are handled by the host, I don't think it's needed. Question is, can the card really be put in IT mode, or is it always going to be "in the way" of Unraid managing the drives?

nomachine works well for me.

Lots of little changes, but the major 6.12 feature was ZFS, because there was a small but very vocal group that insisted ZFS was the solution to all their issues. In reality, very few people currently using Unraid really benefit much from ZFS, and it brings with it a good deal of complexity and adds more issues. It will take a while to get things calmed down.

Another thing to try would be a different VGA cable. Some cables don't have the pins connected that transmit the monitor ID back to the board.

Does it still act up with 6.12.2 in safe mode?

This looks like it may be useful to you. https://fohdeesha.com/docs/perc.html

See here or here Depending on which container you are running.

Disable docker service.

Try here, that's where the template support link leads. https://github.com/monicahq/monica/issues