SimonF

What did you try post diagnostics and i can review novnc

In this file has the PCI devices to be bound to vfio. root@computenode:~# cat /boot/config/vfio-pci.cfg BIND=0000:03:00.0|8086:56a0 0000:05:00.0|8086:56a 0000:05:01.0|8086:56a0 root@computenode:~# On my system 05:01.0 does not exist so my VMs do not auto start. So I would need to remove this. Also the vendor:product is in correct. Take a note of your devices in tools->system devices. Make a change to the checked devices save and then set how you want this will update the file in /boot. So I unchecked group 20 saved, added in group 20 and now my file looks like this root@computenode:~# cat /boot/config/vfio-pci.cfg BIND=0000:03:00.0|8086:56a0root@computenode:~# Changes do not take effect until a reboot.

Ok thanks not sure why this is not kept as should be based on this line. But will look into it. https://github.com/limetech/webgui/blob/081a87381e5314e19619008b2226396cb466ea7b/emhttp/plugins/dynamix.vm.manager/include/libvirt.php#L886

Shares that are prefer wi!l be move to the pool. Appdata is normaly kept on the cache. If it was set to yes only files notmin use would be moved to the array and the array drive would be kept spun up. No issue running single drive you may need to look at doing backups more often. i.e. using appdata plugin.

Is that the qemu parms bit? Will look to add persistance to gui for it.

If you have only a single dev in the cache pool with files on it if it was to fail they would be lost hence the triangle. if you added a 2nd drive to the pool it would be protected. 5g is likely to be the file system overhead.

XML parts are only required if gpu has large memory i.e. 24g. Would you post the full xml for reference?

Also have you tried the spice option or connecting via a vnc client?

You may need to set video and audio parts to be multifumction. <hostdev mode='subsystem' type='pci' managed='yes'> <driver name='vfio'/> <source> <address domain='0x0000' bus='0x01' slot='0x00' function='0x0'/> </source> <alias name='hostdev0'/> <address type='pci' domain='0x0000' bus='0x03' slot='0x00' function='0x0' multifunction='on'/> </hostdev> <hostdev mode='subsystem' type='pci' managed='yes'> <driver name='vfio'/> <source> <address domain='0x0000' bus='0x01' slot='0x00' function='0x1'/> </source> <alias name='hostdev1'/> <address type='pci' domain='0x0000' bus='0x03' slot='0x00' function='0x1'/> </hostdev> Similar to this ^

Sorry I dont know @hugenbdd would get able to answer

Have a look at mover turning plugin

That is just changes to XML to automatically change rather than manual. It will not fix the issue.

Try adding this to the xml. <qemu:commandline> <qemu:arg value='-fw_cfg'/> <qemu:arg value='opt/ovmf/X-PciMmio64Mb,string=65536'/> </qemu:commandline>

Have a look at this https://docs.unraid.net/unraid-os/manual/vm-support#enable-msi-for-interrupts-to-fix-hdmi-audio-support

That is expected as the gpu is exclusivily allocated to the vm. it is no longer visible to the host. Why do you have iommu group 12 bound to vfio? post diagnostics,

Can you provide screen print of system devices in tools? GPU check boxes must not be checked if using nvidia driver.

Have you bound the gpu to vfio if yes the correct driver will not be bound to the card.

Can you post diagnostics. Also if you change to spice does that work?

Which driver version are you using? You may need 470.182.03 for your card

No, report on git hub moving to QEMU 7.2 seems to fix the issue. Hopefully 6.13 will have QEMU updated from 7.1

Yes

Looks like you dont have Vt-d turned on in the bios. Click on the middle icon it will show if IOMMU is enabled. Which motherboard do you have?

Are you able to provide screen print of system devices in tools does it show IOMMU group? both hvm and iommu need to be enabled

That is on the host msata to usb on the hub is passthru to my test unraid vm. Might be qemu wont support full 20g

I have a msi Z690 which the USB C works shows as 20G even thou the hub I have connected is on 5.