JonathanM

Pretty sure the pool name can't be the same as a user share. Try a different top level folder name.

Parity swap only applies when you have a failed drive, the steps you outline are normal drive upgrades, which is what you want. The only thing you may want to do is a parity check after 5 and before 6.

If you can find your *.key files, you can automatically transfer the licenses to new USB sticks, but if you don't have the license files you will need to get in touch with Limetech directly.

Start here, you should be able to use the keywords to further your education. https://forums.unraid.net/topic/51703-vm-faq/?tab=comments#comment-557606

If you use the sparse copy from the array to the cache drive it should shrink. However... What you are doing, over allocating the images, will likely shoot you in the foot if you can't manage them. Keep in mind that the file size WILL eventually expand to the size it is right now through daily use of the VM, unless you take steps to keep that from happening. This event should serve as a wake up call to either take the time to read the information on this forum and apply it to managing thin provisioned disks, or redo the VM's with smaller disk allocations so that even when they are full they don't cripple your server.

If there is not a ready made container, chances are it doesn't work well or at all in that environment, so the next best thing is to set up a VM and host it from there. For example any games that don't have a native linux setup are going to be much happier in a Windows VM. If there is a native linux port available, and there aren't any ready made containers, then @ich777 might be convinced to develop a container if you can provide him with a license to run the game for testing purposes. If it's not a linux native but uses wine, it's less likely to be able to run well or at all in a container.

Maybe @ghost82 could help?

Except for the *.key files.

Try setting it up manually instead of using the creator.

Several, worst is random disconnect / reconnect events that kick the drives from the array.

Yes, although if you are keeping all replaced drives intact, you could possibly recover the data from the old drives if a rebuild goes south. Sorry if I didn't clear things up much, but it really is your choice, based on how safe you want to be.

use bookmarks not currently possible AFAIK

Use nomachine as the remote control tool, install on host and client. Much better, you even get VM sound.

Do you have any containers with their own IP addresses? I noticed macvlan listed in the trace, so that would be my first suspicion.

Read here for how Ubuntu handles it, tldr: they use microsoft signed code https://wiki.ubuntu.com/UEFI/SecureBoot As I understand it, the docs you posted for slack would entail loading the keys into the UEFI firmware, since they only come preloaded with microsoft keys.

https://<new ip>/ Accept the warning and continue.

I believe it's a byproduct of how the filesystem works. In my experience when the last application with an open handle on the deleted file finally closes, they will disappear. Do they persist through a clean shutdown and restart? (Disclaimer, what I've said is general linux / Unraid, not specific to this container)

Benchmarks aside, I'd call moving from AMD to intel an upgrade. Just my personal opinion and experience from working with both companies from back when AMD started making IBM compatible chips. AMD really does try hard, and every once in a while it seems like they've got it, but on the whole you will spend more time dealing with AMD specific issues than intel issues. There is a whole thread right now on how to set up an AMD system to make it stable. No such thread for intel. In the server space, where you just want to set it and forget it, intel is the winner right now.

Not who you were talking to, but yes, it's still hanging for me. I was so hopeful we had a solution, and at first it seemed to make a difference, but it turned out to be confirmation bias, I've had to clear hangs at least half a dozen times since making the change.

In general, if you have no passed through hardware, and are using IT mode disk controllers on old and new systems, it will work OOTB with no work necessary. If you are passing hardware resources directly through to a VM or container, or using a RAID controller on old or new, it can be more complicated.

This is a bad idea. Manually mounting things into the /mnt/user path isn't likely to end well.

Unfortunately there are a good number of things that the GUI just can't handle correctly when it comes to VM XML editing, so you end up stuck in raw XML mode for advanced stuff. All the little nitpicks I'm aware of have been pointed out and listed as bugs, but they seem to have fallen by the wayside. One glaring example is boot order. I've never been able to get the GUI to produce the results I need to change the boot order, I always end up manually swapping the 0 and 1 for boot sequence and then editing it back when I'm done. Be aware, if you make edits in XML mode, they will get blown away if you save a setting in GUI mode. Always been that way, the rationale is that everything you normally need to edit should be handled by the GUI, and if you are advanced enough to edit the raw XML then you should just stick to that mode. I don't like it, complained about it, didn't get anywhere because a solution that keeps custom XML while manipulating the parts specified in the GUI is too hard to code properly. The solution for now is to either keep your VM's simple enough that you don't need the raw XML view, or learn how to parse VM XML to produce working results without involving the GUI.

It's kind of a universal thing, some edits in the wizard end up with a "stuck at updating" condition. It's not just the share config, I've had it happen when messing with video and network as well. Unfortunately it's not easy to reproduce, at least for me. Some things work, some don't.

This is a valid concern, some OS's use information specified or referred to in the XML for licensing (MS / Apple) so it is a good thing to be aware of the content of the XML so you can transfer the important bits if needed.

Booting the server isn't the delineation point, starting the array is. Many people have the array set to auto start, so the difference isn't apparent there, but stopping the array with extra drives over the license limit attached will keep the array from starting until those drives are disconnected, and / or the server will boot just fine with extra drives, but the array will not start.