You are not logged in.
This problem is so incredible that I thought twice before even posting it, as in “nobody will ever believe this is really happening”. But here we go, it’s time to ask.
The hardware setup is described here, but it doesn’t really matter, I guess. The machine is a server + massive desktop + a VM host, all at once.
Now the problem is that whenever a KVM + qemu VM is reset (which happens often for some of the VMs, often forcibly, as part of an experiment that repeatedly tests something on VMs), the sound volume on the default sound device is reset to 75%. A comfortable volume level is somewhere around 20%.
I guess it’s clear why this is a problem: You sit at your mega-desktop, you watch YouTube or whatnot, and all of a sudden a VM gets automatically killed and restarted and … your speakers go BLAH!!!
I have no clue why this is happening, but it started with the switch to wireplumber. The old Pipewire setup had no such issues and the even older “just call it Pulseaudio and don’t build 27 levels of abstractions on it” setup didn’t have such a problem either.
Why is this happening? Why 75%? (Why not 74%?) Why is it tritggered by VM resets? Where is it configured? In case the VM resets cause something weird on USB buses or some such — is there a way to (at least) tell wireplumber what the default volume should be? Next level: Why not pick the last known volume on a similar device as the default?
The device (not sure if it matters) is a Razer Nommo Pro connected via USB.
I’ve stopped all VM activity on the machine and the volume has been stable for ~a day now. This really seems to be somehow caused by by the combination of VM resets + wireplumber.
What’s going on? Any ideas?
Offline
So this is somehow linked to (the equivalent of) virsh’s destroy command. Other VM operations leave the volume on the host alone.
Offline