SxC97

I had a data loss scare due to the constantly dropping drives. I assumed that this was due to either faulty cables or a lack of ram in the system when VMs and docker services were running. I ordered some replacement cables and some more ram for the system and things appear to be rock solid now. I've also tried to get the Windows VM in question working again but now, when I try to boot the vm, even with nothing passed through, I get a blue screen that says that the OS is corrupted and that I can't even boot into it. This is alright though since I kept all important documents on the array and not in the actual VM. I created a new VM called "Windows 10 - New" and went through the entire set up process through VNC. But when I try to pass through the GPU I get the following error in the logs: 2022-03-29T22:56:19.887355Z qemu-system-x86_64: -device vfio-pci,host=0000:2e:00.0,id=hostdev0,x-vga=on,bus=pci.4,addr=0x0,romfile=/mnt/disk1/isos/RX5700XT.rom: Failed to mmap 0000:2e:00.0 BAR 0. Performance may be slow I think this is due to the aforementioned ROM dumping issue. I've tried to look up tutorials online but I haven't made any progress. I first tried the method outlined by SpaceInvaderOne here. At the step where it says to press the power button on your computer once to complete the dumping process, I press the button but nothing happens. I then found this tutorial online to dump the vbios from the command line but the resulting file was much smaller than the one that I had downloaded from TechPowerUp. 512K vs 60K. and it still didn't work. Do you have any tips for dumping my vbios? Or a tutorial that I can follow? Edit: I managed to fix the 2022-03-29T22:56:19.887355Z qemu-system-x86_64: -device vfio-pci,host=0000:2e:00.0,id=hostdev0,x-vga=on,bus=pci.4,addr=0x0,romfile=/mnt/disk1/isos/RX5700XT.rom: Failed to mmap 0000:2e:00.0 BAR 0. Performance may be slow error by following the commands posted on this thread: echo 0 > /sys/class/vtconsole/vtcon0/bind echo 0 > /sys/class/vtconsole/vtcon1/bind echo efi-framebuffer.0 > /sys/bus/platform/drivers/efi-framebuffer/unbind Unraid says that the VM is running but it still just shows a black screen on my monitor. galaxy-diagnostics-20220329-1805.zip

Ok there have been some updates. First of all, thank you for helping me out with this, I really appreciate it! Secondly, I'm sorry for not mentioning which VM was acting up, I could have sworn I mentioned it was the Windows 10 one but it looks like I just forgot 😅 1. I changed the multifunction to 'on' and it did in fact fix the gpu pass through issue. I could have sworn I set PCIe ACS override to 'Both' in the VM settings, but I recently added the Nvidia GPU to the system and it looks like it might have screwed with some settings? 2. I don't have another computer to dump the .rom file from for the 5700xt so I just downloaded it from online and it has been a little finicky. I booted the Windows 10 VM and tried to download the gpu-z utility but I kept getting an error message: Error 0xc00007b “Application was unable to start correctly” This message appears when I try to launch any application on the VM. I checked online and it looks like this can happen when you have a corrupted file system. I tried to download Restoro (which every article online recommended) but I cant run it because it gives me the same error message. I then tried to just restore windows from the settings but it didn't work due to some unspecified error. I also tried using VNC instead of having a GPU passed through but no dice. Is there any hope for this VM, or should I just make a new one? 3. I bound the gpu to VFIO, another thing that I'm pretty sure I did before but apparently got reset when I added the second GPU. I also tried to set the Nvidia GPU as the boot VGA device but I couldn't find the option on my motherboard bios, an MSI B450 Gaming Max Plus. I've tried to outline the steps I took in detail here just in case someone else has a similar issue and stumbles on this thread in the future. Also sorry it took so long to reply, occasionally when the system locks up due to a misbehaving VM, Unraid drops one of my drives and I then have to remove it from the array, start the array, then stop the array, re-add the disk, start the array again and wait for a parity check to finish which can take ~24 hours. Apparently this is a known issue in Unraid.

My bad, I had deleted the vm template to get the server working again. I've re-created the template and downloaded the diagnostics again. I generated these diagnostics file while the vm was not running though, hopefully this still provides some insight into whats happening. If I need to get the diagnostics files while the vm is running, is there a way to do it from ssh? Since the webUI stops working while the vm is running? galaxy-diagnostics-20220306-1320.zip

I've been having some issues with my gaming vm recently. I'll outline the steps that I've taken so far. I booted up my gaming VM (with a GPU passed through) a few days ago and the webUI stopped responding and there was no video output. I eventually hard-shutdown the server by holding down the power button and rebooted it after a few seconds. The webUI was operational again but after a few seconds it again locked up. I assumed this was because I had Unraid configured to automatically start the array and start the gaming vm when the system booted up. I still had SSH access though so I deleted the vm template from /etc/libvirt/qemu and rebooted again. I now had access to the webUI so I quickly booted into safe mode and disabled the array auto-start. I then downloaded the diagnostics and found that the ssd that the vm domains were stored on had some corrupted blocks and it recommended me to run xfs_repair, so I went into maintenance mode, and ran the xfs_repair tool from the GUI with the -v option. I now had a lost+found folder filled with files that were just a long string of numbers and I had no idea how or where to restore them to. I figured I would just use the system as normal and see if those files were even necessary. I restarted the VM service from the settings tab and created a new VM with most of the same options (including the original vdisk file) except I didn't pass it a gpu or any usb devices. After VNCing into the windows VM, it told me that the file system was corrupted and that it would take a while to fix. After that finished, I rebooted the VM and found that I was back into the VM like normal. I then tried to pass through a GPU and found that once again the webUI started crashing and there was still no video output from the VM. I again had to delete the VM template from /etc/libvirt/qemu and hard reboot to fix it. I also noticed that after the webUI stopped responding, the virsh list command just hung and didn't return anything. Perhaps the VM is crashing the libvert or qemu process? All of this just because I wanted to play some Elden Ring Anyways, I then downloaded the diagnostics and decided to post here because I'm all out of ideas at this point. galaxy-diagnostics-20220305-2002.zip

Haha, too true! I've only had Unraid for ~6 months so I'm still getting used to basic things like navigating the UI. While I'm here, I figured I'd post my setup as I hadn't seen anything like this posted before. I have an Unraid box running several VMs and a laptop in clamshell mode on my desk and I want to be able to quickly switch my input/output devices between the laptop and any one of the VMs. My solution is to pass one of the 4 ports on the Renesas to one of the VMs each. I then connect each of the 4 ports and the laptop to a USB switch. You can then connect anything into the USB switch and, just by changing the output on the switch, you can connect your USB peripherals to any of the VMs or the laptop. My switch had 4 inputs for peripherals and 4 outputs for different computers. The 4 inputs was not enough for all of the things that I wanted to plug in so I bought a 16 port USB hub to plug into one of the input ports so I could switch more than 4 usb devices between the VMs. An important thing to keep in mind: if you are also planning on adding things that need USB power, make sure you buy a hub that is powered from a wall outlet, not phantom power from the connected devices! I also bought a hub that had on/off switches for each of the ports, so you can quickly disable devices without having to unplug them, but this is not a necessary feature. I have a lamp that attaches to my monitor and it's easier to turn it on and off by disabling the usb port it's plugged into rather than fiddling with the buttons on the lamp itself. Now I can hot plug things like my Amp & DAC stack, webcam, flash drives, etc... to any VM or my laptop and switch between them in a second! This even make things like moving files between the laptop and VMs very easy using the flash drive as temporary storage. An important thing to keep in mind, if you want to connect things like USB HDDs or SSDs to multiple VMs, don't connect them to the USB hub! I saw several people on Amazon mention that going through several USB hubs and controllers is not great for hard drives, you could end up losing or corrupting data! You are better off passing the drive directly to the VM, going through the fewest number of USB controllers. I feel like I might have done a poor job explaining this set up so I've made a simple diagram to explain things visually. I hope this port helps other people who want to use VMs in a multi-OS, desktop setup!

Ok, so I bought this card, plugged it in and was wondering why it didn't work as expected. Turns out..I'm an idiot. You have to go to Tools -> System Devices and select the checkboxes next to the usb controllers. This is how mine looks: IOMMU group 29: [1912:0015] 27:00.0 USB controller: Renesas Technology Corp. uPD720202 USB 3.0 Host Controller (rev 02) This controller is bound to vfio, connected USB devices are not visible. IOMMU group 30: [1912:0015] 28:00.0 USB controller: Renesas Technology Corp. uPD720202 USB 3.0 Host Controller (rev 02) This controller is bound to vfio, connected USB devices are not visible. IOMMU group 31: [1912:0015] 29:00.0 USB controller: Renesas Technology Corp. uPD720202 USB 3.0 Host Controller (rev 02) This controller is bound to vfio, connected USB devices are not visible. IOMMU group 32: [1912:0015] 2a:00.0 USB controller: Renesas Technology Corp. uPD720202 USB 3.0 Host Controller (rev 02) This controller is bound to vfio, connected USB devices are not visible. Then you have to click the `Bind Selected to VFIO at Boot` button. Reboot and the controllers should show up in the VM's edit page. I also couldn't boot my VM after putting the card in but that was easily fixed by creating a new VM pointing to the same old .img file.

Would you mind checking if the controller works in a MacOS VM? Also, what cables did you need to connect the card to the rest of your computer? I’m assuming you needed a power cable and a data cable?