Proxmox VE 1.9 - VNC console and 100% CPU usage

[cbourque]

Hi all,

I have just installed my first Proxmox setup in order to migrate my current VMs from XEN to KVM and I'm experiencing a weird behavior:

I boot a VM with openSUSE's (11.3) rescue cd to remove old XEN settings and after some time (it always reproducible but the time it takes may vary from like 5 seconds to a couple of minutes) the VNC console hang and the CPU usage of the VM jumps to 100%?!?!?!?!

After that my only option is to reset the VM from the admin console and retry!

I tried to launch VNC from the host instead but got the same result...

Any idea?

Thanks

Christian

vega:~# pveversion -v
pve-manager: 1.9-24 (pve-manager/1.9/6542)
running kernel: 2.6.32-6-pve
proxmox-ve-2.6.32: 1.9-47
pve-kernel-2.6.32-6-pve: 2.6.32-48
qemu-server: 1.1-32
pve-firmware: 1.0-14
libpve-storage-perl: 1.0-19
vncterm: 0.9-2
vzctl: 3.0.29-2pve1
vzdump: 1.2-16
vzprocps: 2.0.11-2
vzquota: 3.0.11-1
pve-qemu-kvm: 0.15.0-1
ksm-control-daemon: 1.0-6

 

[jdw]

I had something very similar happen.

Make sure your existing machine is not touching a serial port. If the serial port was not configured on KVM, attempting to access it from the VM (e.g. for virtual serial console leftover from Xen), CPU usage will go to 100% and the VM will hang.

 

[cbourque]

Hi jdw,

Thanks for your answer but I don't think it's related to the VM itself since I get the same problem with the rescue cd only (no hard disk attached)!

 

[cbourque]

Ok I think I've found something: initialy the VM was configured with 2 Cores/Socket so I changed this value to 1 and now the problem is gone!

I also did a test by downgrading the host kernel to pve-kernel-2.6.32-4-pve and this time it worked with 2 Cores/Socket!

So there seems to be a bug with the current kernel (pve-kernel-2.6.32-6-pve)!