OpenBSD client uses 100% CPU and freezes

[brudy]

Hi There

I run 40 virtual machines in my cluster. All of them behave fine except the one openBSD based appliance.

This appliance uses openBSD 6.1. I have contacted the vendor, but he cannot find anything suspicious on the appliance itself (www.seppmail.ch).

I can trigger the situation by opening the console and hitting enter multiple times. The CPU of the virtual machine goes immediately to 100%.

It is only the KVM process, which starts to use more CPU. The actual CPU of the VM remains 100%/99% idle.

Looking at the CPU usage on Proxmox, the average usage is 5-6%, while on the VM itself it is 100% idle most of the time. On all the other machines CPU usage correlates, but not on this one.

Changing the console does not have any effect.

Sometimes the CPU just goes to 100% and stays without any external intervention. In this case a qm reset is the only solution to recover.

Is there a way to turn on debugging for this machine to find more detailed information?

I got:

# pveversion --verbose
proxmox-ve: 5.1-30 (running kernel: 4.13.8-3-pve)
pve-manager: 5.1-38 (running version: 5.1-38/1e9bc777)
pve-kernel-4.13.4-1-pve: 4.13.4-26
pve-kernel-4.13.8-3-pve: 4.13.8-30
libpve-http-server-perl: 2.0-7
lvm2: 2.02.168-pve6
corosync: 2.4.2-pve3
libqb0: 1.0.1-1
pve-cluster: 5.0-19
qemu-server: 5.0-17
pve-firmware: 2.0-3
libpve-common-perl: 5.0-22
libpve-guest-common-perl: 2.0-13
libpve-access-control: 5.0-7
libpve-storage-perl: 5.0-17
pve-libspice-server1: 0.12.8-3
vncterm: 1.5-3
pve-docs: 5.1-12
pve-qemu-kvm: 2.9.1-3
pve-container: 2.0-17
pve-firewall: 3.0-5
pve-ha-manager: 2.0-4
ksm-control-daemon: 1.2-2
glusterfs-client: 3.8.8-1
lxc-pve: 2.1.1-2
lxcfs: 2.0.8-1
criu: 2.11.1-1~bpo90
novnc-pve: 0.6-4
smartmontools: 6.5+svn4324-1
zfsutils-linux: 0.7.3-pve1~bpo9
openvswitch-switch: 2.7.0-2

 

[manu]

OpenBSD VMs are known to hang from time to time on Qemu/KVM.
This was discussed on the openbsd mailing lists here: http://openbsd-archive.7691.n7.nabb...eezes-and-lockup-Proxmox-PVE5-0-td322999.html
People tried to mitigate the issue with a serial terminal ( see https://pve.proxmox.com/wiki/Serial_Terminal) .

Now a shameless plug for a cousin Proxmox product:
there is the Proxmox Mail Gateway which runs very well on Qemu/KVM.

 

[brudy]

Thank you, I moved now the configuration to serial without vga.

 

[remy_dev]

I have the same problem, it's pretty strange,

I can't repeat the mistake, I often have a different result, but in a general way i didn't have the probleme when i stop using VGA
However... despite this trick it usually crashes when I stop the system with the command :shutdown -p now

 

[fabian]

might be related to https://lkml.kernel.org/r/ee528f72-4131-0700-60bf-3d503908bd6d@gameservers.com ?

 

[remy_dev]

That sounds like my problem.

For information :
After 4 days of trouble-free operation, the virtual machine uses 100% of the processor and crashes. There's php-fpm and nginx on it.
In parallel I launched another openbsd without any additional service. It hasn't crashed yet, I keep testing to see if the machine will eventually crash.

 

[remy_dev]

I reported the problem to openbsd.

I get an answer:
I should try this

echo options kvm-intel preemption_timer=N >>/etc/modprobe.d/kvm-intel.conf


Is this option fair? I can't find information in the wiki.
What impact does it have?

Maybe I should use the parameter only on a type of processor I don't use yet (amd) to avoid impacting other machines?

 

[shelterx]

Old thread but I figured out that high vm idle usage on the host is caused by having usb drivers enabled in OpenBSD.
You most likely won't need them in a proxmox environment anyway so on the OpenBSD guest you can disable them by adding a file called

/etc/bsd.re-config

then add the following content to the file:

disable usb
disable ahci
disable uhci
disable xhci