What exactly would cause noVNC to fall over and die on reboot of a VM and require being fully stopped before it started working again (restart does not resolve though VM does indeed reboot as noted in logs when checked after stop/start)?

 

A fresh install or image of a Windows 11 or 10 system, virtio drivers installed, I hit restart it DOES the restart, but noVNC just shows a black screen or sometimes the last message before the reboot (restarting, stopping services, etc).

Outside of changing the drive location and ram, I left ALL the rest of the settings as unraid defaults.

I have tried with OLDER virtio drivers, I have tried with different vm storage locations (NVMe/HDD, on array, on unassigned device), changed from raw to qcow2 for the storage, there are settings I have not touched such as BIOS and the like, again, I chose the Windows 11 option after hitting Add VM, and just changed disk location and ram size.

 

This is now occurring on the latest unraid when it was not from my prior install in 2019, at some point probably past 6.9 but I dont recall for sure.

It used to be a MIXTURE of noVNC unable to connect AND the VM would sit at 100% CPU, now it seems to just be noVNC is broken.

 

I nuked the system to resolve that and some other issues I thought was my install, alas, the problem remains

 

I have looked through many threads here, and it seems to be a common issue that's occurring now, it seemed to exclude passthrough GPU folks but I just found one as well so thats not it. Some mentioned virtiofs but I have not touched that, again, add vm > windows 11 > ram size > disk location > create > install windows/drivers > restart (sometimes works the first time, then keels over after settling down)

 

The good thing is that at least it responds to shutdown controls from the dashboard/vm section unlike before when some would stick at 100%

Edited May 8, 20242 yr by Hammerfest

I think I may have found the issue, well, at least a pointer!

 

When I check with a 3rd party remote connect software, I noticed the resolution is non existent (like no option to chose even), and when i checked the device manager the display adapter is listed as in error and stopped "because of an error" under the red had device

 

So something in unraid/qemu has changed which is nuking display capability's

 

EDIT: some more info: this is with no pass-through GPU, using the driver disk loaded, driver says "Red Hat QXL Controller" and is from 2020, I know I created new VM's for lab testing in late 2020 early 2021 and I updated the driver disk so its not gonna be the drivers (alone at least), something outside that has changed

Edited May 7, 20242 yr by Hammerfest