SimonF

You need to include the full path /dev/disk/by-id/....... Using /dev/nvmen1 should be fine but that could change if hardware is changed.

If you plan to format in the VM use one without part as that is the first partion on the disk. Re pci passthru is your Bios up todate.

Bind to VFIO so you can attach as a PCI device. or you can put the drive path in the disk location. paths are here. ls /dev/disk/by-id/

Yes you can create hooks scripts. Examples below. oot@computenode:~# ls /etc/libvirt/hooks/qemu.d/ ArcReset Notify* USB_Manager* VMTest root@computenode:~# cat /etc/libvirt/hooks/qemu.d/VMTest #!/bin/bash VM_NAME="$1" ACTION="$2" if [[ "$ACTION" == "prepare" ]]; then export LIBGL_ALWAYS_SOFTWARE=0 export DRI_PRIME=1 export VDPAU_DRIVER=nvidia export LIBVA_DRIVER_NAME=nvidia export __GLX_VENDOR_LIBRARY_NAME=nvidia export DISPLAY=:0 # Log the environment to verify it's applied env | grep -E 'LIBGL|DRI_PRIME|VDPAU|LIBVA|GLX' >> /tmp/qemu-hook.log fi root@computenode:~# cat /etc/libvirt/hooks/qemu.d/Notify #!/usr/bin/env php <?php if ($argv[2] == 'started' || $argv[2] == 'stopped'){ shell_exec("/usr/local/emhttp/webGui/scripts/notify -i normal -s \"VM {$argv[1]}\" -d \"".ucfirst($argv[2])."\" >/dev/null 2>&1 & disown") ; } ?> root@computenode:~# cat /etc/libvirt/hooks/qemu.d/USB_Manager #!/usr/bin/env php <?php #begin USB_MANAGER if ($argv[2] == 'prepare' || $argv[2] == 'stopped'){ shell_exec("/usr/local/emhttp/plugins/usb_manager/scripts/rc.usb_manager vm_action '{$argv[1]}' {$argv[2]} {$argv[3]} {$argv[4]} >/dev/null 2>&1 & disown") ; } #end USB_MANAGER ?> root@computenode:~#

Did you reboot after adding entries to system drivers?

You need to have the following. Mar 31 18:38:00 Tower kernel: i915 0000:00:02.0: [drm] fb0: i915drmfb frame buffer device Mar 31 18:38:00 Tower kernel: i915 0000:04:00.0: Your graphics device 4908 is not properly supported by i915 in this Mar 31 18:38:00 Tower kernel: kernel version. To force driver probe anyway, use i915.force_probe=4908 Mar 31 18:38:00 Tower kernel: module parameter or CONFIG_DRM_I915_FORCE_PROBE=4908 configuration option, Mar 31 18:38:00 Tower kernel: or (recommended) check for kernel updates. Mar 31 18:38:00 Tower kernel: xe 0000:04:00.0: Your graphics device 4908 is not officially supported Mar 31 18:38:00 Tower kernel: by xe driver in this kernel version. To force Xe probe, Mar 31 18:38:00 Tower kernel: use xe.force_probe='4908' and i915.force_probe='!4908' Mar 31 18:38:00 Tower kernel: module parameters or CONFIG_DRM_XE_FORCE_PROBE='4908' and Mar 31 18:38:00 Tower kernel: CONFIG_DRM_I915_FORCE_PROBE='!4908' configuration options. Examples in system drivers

install Network UPS Tools (NUT) by desertwitch from Community Applications

If you are running nut you can add 2nd as a slave.

Will look tomorrow as late here can you provide intel_gpu_top -J -n 1 when transcoding. If you run intel_gpu_top to you see any usage?

You may be able to change date with apctest on CMD line. I don't use APC so don't know if it works. https://opensource.com/article/21/12/linux-apcupsd

A310 needs change to intel top timers, Latest version of GPU stat and top fix this so check you are on the latest.

Correct

Yes install the tail scale client for the VM os.

Try 2025.03.29 to see if that fixes your issue.

Release 2025.03.29 - Chg Use new -n intel_gpu_top option for more reliable data gather. This removes the top timers requirements. - N.B must be running the latest intel_gpu_top from ich777 i.e 2025.03.15 or later.

One of the reason for them being plugins it allows out of cycle updates and fixes. If was rolled into the core then would be max 3/4 updates a year.

Do you have the latest version of Intel_gpu_top plugin. intel_gpu_top -J -n 1 should provide a clean json. I am in the process of changing to use this new method. New Intel.php is a few post up.

If easier I can make a package available to install.

Repo is here https://github.com/SimonFair/gpustat-unraid

Looking for some testers. I have worked with ich777 to put my modified code into intel_gpu_top to allow it to run n times. intel_gpu_top -J -n 1 This allows me to not need timers and I have raised this upto the intel_gpu_top team to include. You will need to be running this version of Top. Release Notes 2025.03.15 Bumped 'intel_gpu_top' version. Based on Intel-GPU-Tools v2.0 with patch from SimonF Make copy of existing file i.e. cp /usr/local/emhttp/plugins/gpustat/lib/Intel.php /usr/local/emhttp/plugins/gpustat/lib/Intel.old Install this file in /usr/local/emhttp/plugins/gpustat/lib It is working on both igpu and dgpu on my test system. Let me know feedback and I will package over the weekend to release. Intel.php

Framebuffer disabled in step 3, other bit means booting without binding to vfio and the driver resets the card.

AMDGPU and lspci -k will show it in use at boot.

Release 2025.03.28 - Add Fan for Intel DGPU. - Fix Handbrake Icon for Intel. - Add Metrics for A310s and others that report via clients. - Add Memory for Processes(Intel) - Add Compute Bar(Intel) needs to be enabled in settings.

Post diagnostics

Thanks will release an update tomorrow.