steve1977

This is fantastic. Only realized now that this is possible. So, I don't need to buy a second GPU and worry about space and power consumption. Excellent! One question before I get started: my GPU model (GFX 1050 by Colorful) is not listed by "GPU-z/ Techpowerup". Does any other brand work?

Great, thanks!

Thanks. I did a bit more forum search. From what I have read, my setup may not work to pass-through the GPU to the VM: 1) I don't have an on-board GPU and only one Nvidia GPU. So, I'd need to buy a second GPU to make this work? 2) To use the GPU headless (no screen attached), I'd still need to get a dongle as linked above? Do I understand this correctly or any other suggestion to get this working? Thanks in advance!

One disk became unmountable. It still shows "green" and the array is up and running. Disk content is not available. Any thoughts what to do? Diagnostic attached. tower-diagnostics-20171027-2252.zip

Does RDP use the built-in VNC protocol? I actually don't use the built-in VNC viewer, but access it from a Win notebook via RDP. However, I fail to passthrough my Nvidia card to the Win VM. Is this related to RDP or did I do something wrong in setting up the pass-through? Also, I noticed in a related thread that it may not work anyways without a dongle (https://forums.lime-technology.com/topic/54292-nvidia-gamestreaming-on-headless-windows-10-vm-help-needed-please/). Is this still the case or has there been some form of work-around to this?

I don't have to use RDP, but it is very convenient and I like how it works. Do you imply that I cannot pass-through my GTX1050 and use RDP? No work-around?

I changed my setup and unfortunately I am facing the same issue again as I did two years. Never change a running system... Anyhow, it is done now and I have a new mobo (Asus X299) that supports two M.2 devices. I have installed my "old" Plextor M6E and also added a Samsung PM961 M.2 NVMe PCIe SSD. I moved my VM to the new Samsung drive. Unfortunately, the above mentioned error is back (Error 0x8007003B). I have no clue how it was fixed two years ago, so am back to this thread. I ran lspci -k as advised before. Please see below results. Any thoughts appreciated: https://pastebin.com/jsG6Ukk3

I have it somehow working now. Nice to have MacOS on Unraid. I am now left with two issues though: * I cannot pass-through the audio card. Once I do, I get the error message two posts above * I haven't succeeded yet to install/migrate Clover to my local harddisk. My issue starts in min 16:01 of the youtube video. The labels in disk utility don't match with what you show in the video. Also, only one of the two EFI partitions includes clover. Any help appreciated!

Thanks for your replies. Besides the inconvenience to do backups, is there any other strong reason not to have a 1TB visk on the cache drive? Does it negatively impact performance or bring higher chance of corruption? I doubt I can get two vdisks to work with MacOS. And you are right, that MacOS doesn't allow to put my iphoto library on a mapped network drive (same as what you describe for your temp file). My SSD/NVME is the cache disk. It carries two vdisks (one is the VM for MacOS and one the VM for Win10). Is this not a good setup? With my setup, I don't think I can even passthrough teh SSD/NVME as the vdisks are on the SSD/NVME?

Making baby steps forward, but not really moving to get the Mac VM installed. Above originated as I was passing through the audio card. While I'd eventually would love to get it working, I deactivated it for now. This allowed me to at least start the VM and see it throgh VNC. However, I am now facing the same issue as described before (Clover 4259 hangs). I had set up using a USB created using the latest youtube video (with fakesmc file). I then followed the suggestion below and manually added the OSK key to the XML file of the VM. However, this didn't change the issue. Any thoughts appreciated!

Made an attempt this weekend, but failed to get it working. I followed the two USB videos to create a usb key and then set up the VM. At the time, I am first to start the VM I am seeing the following error message. I did some google search and change "Enable PCIe ACS Override" to "yes" in "Settings / VM Manager", but also no success. Any idea what I am doing wrong? Error message upon start below: *** internal error: process exited while connecting to monitor: 2017-10-22T12:00:04.747339Z qemu-system-x86_64: -chardev pty,id=charserial0: char *** device redirected to /dev/pts/1 (label charserial0)*** 2017-10-22T12:00:04.791750Z qemu-system-x86_64: -device vfio-pci,host=00:1f.3,id=hostdev0,bus=pci.4,addr=0x1: vfio error: 0000:00:1f.3: group *** 13 is not viable*** Please ensure all devices within the iommu_group are bound to their vfio bus driver.

Let me revive an old thread. I am trying to figure out whether I can game on a RPi. I have Unraid and running a Win10VM. Have a Nvdia 1050, which should meet the requirements. However, my Unraid is running headless and I access the VMs via RDP. If I pass-through the Nvidia GPU to the VM, I cannot access the VM via RDP. Is there any work-around that will allow me to pass-through the Nvidia GPU and still use RDP to access the Win10 VM?

Thanks for your replies.

Oh. Easier than anticipated. I just changed the memory I have assigned and now things are working. Does this really work magically out-of-the-box with a Win10 VM? Feels too good to be true. I can just switch motherboard/gpu/all and everything just works?

I have a bad feeling that things are doomed, but let me reach out whether there is a solution to my issue. My mobo just died and I installed a new one (plus new ram, new gpu). Much to my surprise, Unraid just keeps working with the new mobo as if nothing had happened. This was good news. Unfortunately, my Win10 VM is not starting. I assume this may be a Win10 issue and not solvable? Please find the Unraid error log below. Any hope that I get my VM back? 2017-10-21 13:56:49.470+0000: starting up libvirt version: 3.5.0, qemu version: 2.10.0, hostname: Tower 2017-10-21 13:56:49.470+0000: Domain id=1 is tainted: high-privileges 2017-10-21 13:56:49.470+0000: Domain id=1 is tainted: host-cpu 2017-10-21T13:56:49.525348Z qemu-system-x86_64: -chardev pty,id=charserial0: char device redirected to /dev/pts/0 (label charserial0) 2017-10-21T13:56:49.529801Z qemu-system-x86_64: cannot set up guest memory 'pc.ram': Cannot allocate memory 2017-10-21 13:56:49.572+0000: shutting down, reason=failed 2017-10-21 13:57:32.661+0000: starting up libvirt version: 3.5.0, qemu version: 2.10.0, hostname: Tower 2017-10-21 13:57:32.661+0000: Domain id=2 is tainted: high-privileges 2017-10-21 13:57:32.661+0000: Domain id=2 is tainted: host-cpu 2017-10-21T13:57:32.732627Z qemu-system-x86_64: -chardev pty,id=charserial0: char device redirected to /dev/pts/0 (label charserial0) 2017-10-21T13:57:32.744414Z qemu-system-x86_64: cannot set up guest memory 'pc.ram': Cannot allocate memory 2017-10-21 13:57:32.780+0000: shutting down, reason=failed

I am running Win10 in a VM on Unraid. So far, I have not been prompted to install the Fall Creators update. Is this still related to the bug about using more than one core or just a matter of time that it will be pushed over the next few days?

Ok, thanks, looks like a new config does the trick.

Any thoughts whether this is a HD or a cabling issue? Diagnostic attached. Thanks! tower-diagnostics-20171020-2341.zip

Any insights whether this is now broken with the latest Clover as some on this forum have asserted?

Any thoughts on my Qs? Thanks in advance!

Great, thanks! Reads like the easiest way for me is to just try it out. It reads like this is a good an easy solution.

Thanks. Super helpful video. I have two questions: 1) For what kind of updates do I need to reduce the cores to 1? I have done so when installing the Win10 VM, but not sure whether this is required again for the "normal" daily Windows patches and updates? Or the upcoming Creators update? 2) Why do you suggest a small vdisk? Is it per se better to map network disks within the VM for file storage? I am a heavy user of Itunes. Is there any difference or preferred method in having the itunes library within the vdisk (VM) or on a mapped network drive? 3) I am about to start a Mac High Sierra VM (thanks for guidance in the other thread). When it comes to vdisk size, do I need to start a large 1TB vidks or is there also some way to do this with a small 70GB vdisk? It loks that Mac OS lacks the capability to auto-mount network disks at startup and also Photos library apparently cannot run on a non-Mac native file system. So, I better have it all inside the vdisk?

Thanks, this is great! Waiting for a reply whether things are broken with the latest Clover and then will give it a shot! This sounds really good!

Thanks, this is great! re 1 - what is UD? Unraid? If so, Unraid does not auto-mount anything, does it? re 2 - how to "assign" what VM it is assigned to? If I understand the livirt hotplug correctly, I no longer need to "mount" USB devices in the XML file for the specific VM. Do I still "assign" USB ports? Or is it "first come first use"? re 3 - very cool. That's a great thing and will be primary use case. Would still be good to have the option with libvirt, so questions 1 and 2 remain relevant.

From Sierra to High Sierra? And it just worked?