Errors booting vm XEN

Hi All.

I have this problem when I tried to re-load a guest vm using XEN 4.0 on OpenSuSE 11.3 (always happens the second time startup or reboot vm) :

**Error starting domain: POST operation failed: xend_post: error from xen daemon: (xend.err ‘Device 0 (vif) could not be connected. Hotplug scripts not working.’)

Traceback (most recent call last):
File “/usr/share/virt-manager/virtManager/engine.py”, line 824, in run_domain
vm.startup()
File “/usr/share/virt-manager/virtManager/domain.py”, line 1286, in startup
self._backend.create()
File “/usr/lib64/python2.6/site-packages/libvirt.py”, line 333, in create
if ret == -1: raise libvirtError (‘virDomainCreate() failed’, dom=self)
libvirtError: POST operation failed: xend_post: error from xen daemon: (xend.err ‘Device 0 (vif) could not be connected. Hotplug scripts not working.’). **

Is strange, the XEN is working but this is about the xen HotPlug scripts and network, since after shutdown vm, the vifX.0 (X is any number) is still up and active(shown by ip add).

Is like the xen hotplug scripts aren’t shutdown the vif, ant the second time it appears lik bussy. When restarted the Host, the vif dissapears and I can start the vm without errors, but again, the second try gives the same error.

It is affecting both type vm, paravirtualized and full virtualized.

Any ideas?

Regards

Hi there,

as far as i remember (maybe wrong!) does xen try to send some commands over an vif to the guest system.

So i think you found a bug :slight_smile: I would fill in a bug report to xen, or maybe to opensuse too.

There is a bug entry about this. You can’t start any new vm after you shutdown vm. It’s already resolved, but they’re waiting with release of new kernel to gather some more fixes. Now only workaround is restart whole machine.

ok thanks for the replies. I will wait until this bug is resolved.

I’m also seeing this – after shutdown I get this error and have to reboot. Any idea if there’s a faster turnaround than rebooting while we wait for a fix? I’ve been experimenting with domU installs and this bug makes cycling through tests painful. I tried restarting xend and that did nothing. I’d like to hear when the fix is available. Maybe someone can update this thread?

same problem here :frowning:

This has worked for me (using opensuse x86_64 for laptop)

Install xen-kmp-desktop and kernel-xen-base , reboot and thats all for me. Maybe you can try it :slight_smile:

vael - just tried that and I get the same thing. I’m also running openSUSE 11.3 x86_64 but on a server.

Did you maybe do something else in there? Seems like there should be a way to reset things without rebooting at least. This seems like some issue with a Xen network resource not getting cleaned up. I’m wondering if there’s some way to reset all Xen networking or something. But ideally a patch … any idea when that’s scheduled to hit the openSUSE updates?

vael - since you’re an openSUSE guy, have you run any paravirtualized guests? I’ve gotten the hang of installing fully virtualized guests. When I try to install paravirtualized Linux guests (I was trying CentOS that way) I’m having trouble getting the booted guest to be able to get the stage2 image. It seems in the paravirtualized path networking isn’t there yet when it needs those, whereas with a fully virtualized guest it gives me network setup capabilities if I say to load from the web. Haven’t found many docs on this …

Actually I can now power off and on once I’m past the guest install. During the install, if things fail, I still get the error and have to reboot before I can try again. So there must be something about the networking set-up during install that’s different than it is for running guests. If I get better at creating guests I suppose this won’t be a real issue. :sunglasses:

Just full virtualization

I am running a personal laptop, so I can reboot and do crazy things without problem here.
I Got a full virtualized Debian, but I cant install a second virtual machine, so, this is a mass :frowning:

Do you have the memory to implement a second VM?

I tried to install a new VM using full virtualization. It used to work well in my earlier version of Suse. Now, with a fresh install of OpenSuse I cannot start a new VM. I tried all tricks like changing the /var/lib/xen partition to ext4 instead of ext3 … .It does not work. Help please. Should I wait for the new version of kernel for the fix.

a lot of memory

Looks like this got fixed with this release of kernel.