Ballooning not working with Fedora 41/42 VMs using kernel 6.14/6.15

Taomyn

Active Member
Aug 30, 2023
123
30
33
Luxembourg
I have two Fedora servers, one is still on v41 the other I upgraded to v42, but I cannot get either when using any 6.14 kernel to use memory ballooning. If I revert the v42 server kernel 6.13 everything is normal - I couldn't figure out how to install 6.13 on v42 but the issue doesn't block the server very often.

For example, my Fedora 42 server:
Code:
#Fedora 42
agent: 1
args: -vnc 0.0.0.0:04
balloon: 1024
bios: ovmf
boot: order=scsi0
cores: 1
cpu: host
efidisk0: cpool-01:vm-104-disk-0,efitype=4m,pre-enrolled-keys=1,size=1M
hotplug: disk,network,usb,memory,cpu
machine: q35,viommu=virtio
memory: 2048
meta: creation-qemu=8.0.2,ctime=1692967959
name: FRINK
net0: virtio=BC:24:11:21:8A:43,bridge=vmbr0
numa: 1
onboot: 1
ostype: l26
scsi0: cpool-01:vm-104-disk-1,discard=on,iothread=1,size=10G
scsihw: virtio-scsi-single
smbios1: uuid=74151127-8bdd-42d7-9c92-fd9a91f92de4
sockets: 1
tags: vm;linux
vga: virtio,clipboard=vnc
vmgenid: e02004df-51d4-4303-9f0c-82a6448feb8a

It starts up and looks like this:

1749020544703.png

If I then stress it out to force it to use more ram, nothing changes and it tries to use what was already there - it never tries to allocate any of the extra 1GB ram that should be available:

1749020819355.png

The exact same thing happens on the Fedora 41 server but it's worse as it's a far larger server running many containers etc, and basically comes to a screeching halt and I have to kill the VM. I've removed 6.14, keeping it on 6.13, and avoid updates to the kernel.

Any ideas? The host itself is not having memory issues, I have other VMs that include Windows Server and Debian 12, and they're all fine. It's just Fedora with the issue.

1749021221139.png
 
I'm still getting the same thing with kernel v6.15 which I thought had fixed the issue but didn't - so I foolishly upgraded the one server from Fedora v41 to v42 (so no kernel 6.13) and a few days later it started to have issues. In this has the ballooning is set for 4GB/10GB and it never goes above 6GB before it eventually either crashes or grinds to a halt. For now I have disabled ballooning on both.

The host and the two VMs are as up-to-date as I can make them according to update checks.

CPU(s) 32 x AMD Ryzen 9 5950X 16-Core Processor (1 Socket)
Kernel Version Linux 6.8.12-11-pve (2025-05-22T09:39Z)
Boot Mode EFI (Secure Boot)
Manager Version pve-manager/8.4.1/2a5fa54a8503f96d
Repository Status Proxmox VE updates Production-ready Enterprise repository enabled
 
I get what you're saying and thank-you for responding, I did try disabling hotplug at one time and it made no difference. Back when it was possible and I could boot both VMs with kernel v6.13, none of this was happening. I also have several Debian 12 and Windows 2025 servers running and none of them have problems.