meep

Welcome, and good luck!

Adding a USB adapter won’t impact performance, and will enhance usability for your setup.

It's odd that the mouse shows up like that, but the keyboard does not. I would strongly advise that you figure out which physical ports are attached to the other controller and pass that controller through. First, you'd need to plug a uniquely named device (say, a usb thumb drive) into each port in turn and observe in your devices list what controller its appearing under. Then, assuming there's nothing else on this controller, and its in its own IOMMU group, use the VFIO-Config plugin to select it for passthrough and reboot. You will then be able to pass the entire controller through to the VM, and plug your KVM into one of those ports. You'll likely find this a much more reliable set up. Bear in mind that some of the ports attached to the controller may be your internal motherboard USB headers.

So you have a mouse and keyboard attached to a KVM, which in turn is attached via a single USB cable to your unRaid server (with a similar single USB attached to your standalone PC). Is that topology correct? In the listing above, can you please point out which lines relate to they keyboard and mouse attached to the KVM? Can you point out if ione of the hubs listed is the KVM itself? That information would be useful. Do you have the KVM switched to your VM when you are booting unRAID? When you are booting the VM? What exactly are you attempting to pass through to the VM? (a screengrab of your VM config would be useful, particularly the bottom section that shows USB and PCI device passthrough). I use KVMs with VMs myself and they work fine, but I always have a dedicated USB adapter passed through to the VM and have the KVM attached to that. This facilitates native device hot swapping inside the VM and works much much better than trying to get an unRAID device working.

This is what I set up, but honestly, I reboot my MacOS VMs so infrequently, I forget how much this covers up! Also, I neglected to mention I run an RX570, and it sounds like that works OK for you, so I might not be able to help much with the RX5700.

Does it truly hang? I've had cases where it just takes a very long time to progress through the first boot. Keep a close eye on the loader bars to see if they are moving at all. That 4x progress bar is fairly normal (you can 'hide' it through a setting in Clover that throws up a full screen image). I'm running 10.15.4 with Clover and it boots OK. Happy to share anything you'd like to see. You might find some nuggets in this post where I describe how to get HEVC working; https://mediaserver8.blogspot.com/2020/01/enabling-hevc-on-rx570-in-macos-vm.html

Are you able to remote in to your vm? Did you install something like splashtop or teamviewer after os install and before you added the GPU? If yes, what is the status of the gpu in device manager inside the vm? if no, I’d suggest removing the gpu from the vm and changing back to vnc. Then install splashtop and ensure that’s working. Then, add the gpu back in and try to get some visibility on what’s going on. of course, splashtop will only be useful if the system is booting to desktop. You’ll need to use netplwiz in the vm when youre in vnc mode to allow auto-logon. You don’t want the system to be sitting at a logon screen that you cannot see. a further wrinkle would be if the system has a boot error. If this happens, Windows will go into recovery mode, but you won’t be able to see that. I’ve found it useful in the past to install virt-manager docker. This allowed me easily add a temp. vnc display alongside the gpu to observe the boot process but without all the pfaffing around I’d have to do using unRAID UI which destroys some xml customizations. Virt-manager is non destructive, so that becomes less irksome when your swapping devices in and out. good luck.

Yes, of course, except for the pro license which has no such limitation on attached devices https://unraid.net/pricing You get get sort of used to that.

And why you might want to wait a couple of weeks to make a decision;

Hi, welcome to the forum, and unRAID. I’ll take a stab at your questions; It’s recommended to use a USB 2.0 device for the OS. These are found to be more reliable. There is no real performance penalty, as the system boots from the drive but th OS resides in memory and the boot drive is rarely accessed thereafter. Your license is tied to your USB key UUID, so if it fails, you need to go through a process online to re-license the new key. I believe you can do this once every 12 months for free, but lime tech are a reasonable company to deal with. UnRAID includes a facility to manually back up your boot drive contents and configs for easy restore. I believe this will become automated in future versions. how you configure your shares will dictate how long data stays on cache. Also, if you store your docker data, vm data there or not will impact size requirements. I’ve found multiple cache drives to be a little unreliable in my setup so reverted to a single SSD. UnRAID counts drives in the array, cache and parity towards license. USB key and unassigned devices are not counted. Your parity drive must be at least as large as the largest drive in your array. That’s the only requirement. Parity does not store any of your data, that’s all on the array drives. The parity drive simply stores a parity bit that allows for the recalculation of data on any failed / removed drive. You can add a second parity drive to protect against 2 simultaneous drive failures. parity is NOT backup. You should plan to store additional copies of any valuable/irreplaceable data on separate systems as well as on unRAID. memory is allocated 100% to a VM while its running and cannot be used by any other processes. Once the VM is stopped, it becomes free again. A share set to prefer will have its data stored on cache and it will not be parity protected. You could set up a scheme to duplicate it to your array, but this is not inherent functionality. Further, any share set to prefer will actively move its data from the array to cache, if there’s space. Here’s the specific help note on this share type; Prefer indicates that all new files and subdirectories should be written to the Cache disk/pool, provided enough free space exists on the Cache disk/pool. If there is insufficient space on the Cache disk/pool, then new files and directories are created on the array. When the mover is invoked, files and subdirectories are transferred off the array and onto the Cache disk/pool. Hope that helps! Do grab the free time limited version and have a play with it. good luck!

Hi, welcome to the forums, and unRAID. It’s nice to see prices in €€€ for a change!!! your use case is quite common, but without more specifics, it’s a bit of a challenge to help. one good place to start would be the gaming VMs as these will dictate a lot. How would you like to spec these out? Let’s say you’d like 16gGB RAM, 8 CPU cores and a decent GPU. Assuming both VMs will be in simultaneous use (?) you’ll need to budget for at least a 16 core, 32GB base system, but that doesn’t leave any headroom for unRAID itself, your dockers etc. while all need memory and CPUs. so, a more reasonable VM config might be 6 cores, 12GB, leaving plenty of headroom in that 16 core 32GB system. of course, you could scale this back even further, Windows will run quite happily in 8GB on 4 cores. Would that be sufficient for you? Or maybe you don’t envisage running the VMs simultaneously, in which case you could share all the resources, including GPU. in any case, your VMs will work best if you can dedicate GPU cores to them by pinning them so the vm has exclusive use. you also don’t mention if you have a preference for an Intel or AMD based system? AMD seems to be where the price/performance is these days. When speccing your system, I would suggest planning for growth. Make sure you can expand memory, upgrade the cpu and have a few spare PCIe slots. consider NVMe storage, especially for the dedicated VM drives. These are blazing fast, and easy to passthrough. You’ll benefit from the performance, and save a few SATA connectors for your other drives! when you make a decision on intel vs AMD and narrow down your specs, cast around here for people with similar setups. You’re looking for a MB that is easy to work with in terms of IOMMU grouping. You want to be able to get your passthrough devices into discrete groups as easily as possible. Some setups cooperate better than others. thats all I have for the moment, come back with more details and someone might give more specific advice. good luck!

The easiest way is to install the VFIO-Config plugin from Community Applications, mentioned by @T0a, above This provides a visual listing of devices, allows you click checkboxes to add them an generates the .cfg file for you.

I think I've been fortunate with AMD and I don't see the reset bug. Likewise, I think you'll find people will have smooth sailing with nVidia. My trouble for me was the -43 error. It seemed that no matter what I tried, I could not get nVidia drivers to install and in my most recent bout of attempts ended up installing Win10 about 10 times, trying different combos of machine type, bios etc. (nVidia drivers do not like to install if there's any hint that the OS is a VM) I finally found the magic formula, but it was hard fought. YMMV, as they say. Batteries not included. Investments may fall as well as rise.

You are welcome. Good luck!

Hi You are running CA Backup/Restore at 18:00. This is stopping your doclkters to allow the appdata backup take place. It tells you this in the logs; May 11 18:00:02 Tower CA Backup/Restore: ####################################### May 11 18:00:02 Tower CA Backup/Restore: Community Applications appData Backup May 11 18:00:02 Tower CA Backup/Restore: Applications will be unavailable during May 11 18:00:02 Tower CA Backup/Restore: this process. They will automatically May 11 18:00:02 Tower CA Backup/Restore: be restarted upon completion. May 11 18:00:02 Tower CA Backup/Restore: ####################################### It seems to be scheduled for this time every day. Are the dockers coming back up for you when this is completed? I'd suggest scheduling CA Backup for a more convenient time - 3.00AM?