Rishumacher

Hi, After giving up on this issue for a couple of years, I've decided to try once more before abandoning hope of using Unraid as a VM server. This thread details my previous attempt, but in summary I had tried to passthrough a GTX 1060 on my Unraid server into a Linux vm and tried to login to it as a daily desktop driver without success. I was told that it may be hard to dump vBIOS in Linux if the GTX 1060 is the primary graphics card. So I recently added an AMD RX 550 as a secondary graphics card. I created an Ubuntu VM and tried to passthrough the AMD card as per the official guide but all I can see is a blank screen on a monitor directly connected to the server. I tried all troubleshooting steps mentioned on the guide like setting the iGPU as the main GPU. I even tried setting the AMD GPU as a secondary GPU as per this thread but still get the same result. My goal is to be able to experiment with different Linux flavours (Ununtu, Fedora, Linux Mint, etc) as my daily desktop environment for web browsing and video streaming on a 4K monitor. I have a spare HDMI port on my monitor that I can directly hook up to the Unraid server below my desk, but I'd like to use my wireless mouse and keyboard to switch between my office laptop and the Unraid server, rather than attach wired inputs to the server. Any guidance or support you can provide would be immensely appreciated and could really help me navigate through this long standing bug bear. zion-diagnostics-20240920-2154.zip

Hi @ghost82, Thanks for pin pointing the error. Unfortunately I've correcting the ROM issue multiple times with no success. I've downloaded the ROM for my graphics card techpowerup and removed the Nvidia header based on instructions in SpaceinvaderOne's tutorial but get the same result (VM starts, blank screen, IP cannot be found on router). I've tried to dump vbios using Spaceinvader One's user script, but I'm getting rom: input/output error like many users have reported. Moreover, I didn't realise that attaching GPU to vfio will make in unavailable to Docker and I need my Nvidia GPU to run deep learning docker containers. So, I'll abandon this approach and instead want to setup an Ubuntu VM using igpu that can be used as a daily desktop and not have choppy performance like the default VNC card. Would you be able to guide me in doing that? Appreciate your help.

Hi @ghost82, Thanks for your help. Based on your suggestions I think I've made some headway. Here's what I've done. 1) Created a new Ubuntu VM using VNC as the Graphics Card and Network bridge set to br0 (so that I can connect to the VM using an MS RDP client for MAC). 2) The VM started alright and I used VNC remote to install Ubuntu, create a user and then install xrdp using this guide. 3) With VNC still as the graphics card, I'm able to connect using Microsoft RDP client, but the graphics is extremely sluggish. 4) I then changed the graphics card and audio to GTX 1060 based on your guide above (the bus changed from 0x04 to 0x05 for the new VM). 5) With this new config, I got a blank screen when connecting via MS RDP. 6) I amended the XML to add a GPU ROM manually (I extracted the GPU ROM a while ago and it may be incorrect) using this guide. 7) Now the VM starts but I can't connect to it using MS RDP. Even my router page does not detect a new client added and its corresponding ip address. Can you suggest how to proceed from here? Also, for every new VM that I create, is it the normal process to start with VNC -> setup user and xrdp -> then change the graphics card? Thank You. zion-diagnostics-20220730-1748.zip

Hi ghost82, Thanks for your response. I'm not sure I follow your instructions completely. 1. Sorry, I uploaded an incorrect diagnostics file that I had downloaded from this forum!!! Attaching the right diagnostic file now. 2. I don't have a matrox g200 graphics card, just Intel integrated graphics and GTX 1060. I created the Ubuntu VM with Intel graphics as the primary and GTX 1060 as the secondary. It has successfully started but how do I connect to it? When I right click on the VM, I don't see an option to "remote" to it. 3. I have ACS Override and "Bind to VFIO at Boot" selected for the Nvidia VGA and audio devices. However, I'm unable to dump the gpu vbios using SpaceInvader's script as I'm getting an error. Appreciate your help. zion-diagnostics-20220724-2356.zip

Hi Forum users, I've recently installed Unraid OS on an old PC and I love the value Unraid provides from old hardware. As a NAS server it's working perfectly and using Docker containers is also super easy. One thing that I'm unable to figure out is using my Unraid server to play around with multiple flavors of Linux. I'd like to use an Ubuntu VM as my general use PC for web surfing and programming workstation. I want to pass through my GTX 1060 and use Nvidia drivers to get a more smooth UI experience (and also for building simple deep learning apps) than VNC drivers provide. I've setup my audio and video devices with ACS override so that the Graphics and Sound card are both Nvidia and use the same bus and slot but different "function" nos. However, I'm unable to figure out a way to connect to the VM using some remote access client. I'm using a Macbook to connect to Ubuntu on my Unraid server. I've come across some step by step guides to setup GPU passthrough for a Windows VM, but couldn't find any beginner's guide for using Linux with GPU passthrough. If I create an Ubuntu VM using the VNC drivers, I'm able to connect using the default web based client but the performance is very choppy and I can't get a 4K output or audio working. Any help will be greatly appreciated, I'm attaching my system diagnostics. tower-diagnostics-20220509-1130.zip