High dom0 cpu usage by [xen-balloon]

After mistakenly switching off the computer without closing most of the vms today, I am experiencing a high cpu usage by dom0 (50% approximately, 75°).
In particular, the task manager shows how is [xen-balloon] to use all this cpu.

Today I had installed bluez in fedora to manage the bluetooth and updated the kernel (5.4.150-1), but uninstall the first and use the previous kernel (5.4.143-1) do not improve the situation.
Any help will be appreciated!

I didn’t find any other topics about this, sorry in this case

There are few other reports of this in https://github.com/QubesOS/updates-status/issues/2645 already. I’ve forwarded them to xen-devel already, and I’m not the only one: xen-balloon - search results


Thanks for your reply! I will follow the reports and hope we could fix this soon and post here some good news

I experience the same using the newest vm-kernel, in memory-constrained VMs (like sys-firewall).

qubes-4.0 / current-testing

(so I’ve changed the default in-vm-kernel to the previous version).

i had this issue with 4.0 testing. upgrading to 4.1 fixed it.

if you can’t fix/upgrade right away: this bug only triggered when i started enough app-vms so that their combined maxmem exceeded 4 gb, which is not much, but when you trim everything down to a minimum, you may keep a working system without overheating your cpu!

I’ve changed the default kernel in dom0 so every vm have changed its default to previous version, but the problem persist also, also changing only the default in-vm-kernel and let the newest in dom0

Unfortunately I already use 4.1. I think this could be a good workaround, also because the problem is that dom0 uses almost all the ram at its disposal (3050 in my case), but despite I have changed the maximum ram in grub.cfg with a lesser value, it keep using the same amount of ram.

As I use qube in my main laptop I’m worried about having an average temperature of 75°-76° all the time

Recent kernel from current-testing repo (5.4.151, 5.10.71) should have it fixed.


I updated it yesterday and everything seem to work fine now, thanks!