Hi Jan,

On Feb 23, 2011, at 16:23 , Jan Kundrát wrote:

> On 02/23/11 15:51, Troy Dawson wrote:
>> Just because the client kernel is running 2.6.18-238.1.1.el5xen does not
>> mean it is trying to be a xen host.  It means that it is running
>> paravirtualized.  If your xen machine was setup to be a paravirtualized
>> client, then it *has* to continue to run the xen kernel.  You can't just
>> switch from the one to the other (as far as I know).
> 
> Hi Troy,
> the domU has always been a fully virtualized one, as requested by the
> builder=hvm stanza given in the configuration file.

I guess that explains why noone else is seeing this problem. Why would I want to run an SL5 Xen VM under an SL5 Xen  hypervisor as an HVM instead of a paravirt VM? I agree it should work, though, and I know that it did in the past.
 
> The disk image
> contains everything, from the bootloader and partitions to the kernel,
> and the in-the-image-installed Grub is invoked and asks me what kernel
> to boot.
> 
> When I was speaking about kernel changes, I meant that I have installed
> various versions of the kernel RPM inside the domU, one of them being
> kernel-2.6.18-194.32.1.el5, other kernel-2.6.18-238.1.1.el5 and yet
> another being kernel-xen-2.6.18-238.1.1.el5.
> 
> Now, no matter what kernel and Xen versions I choose to run in the dom0,
> physical host, I haven't managed to boot the domU using
> kernel-2.6.18-238.1.1.el5. I'm always using full virtualization, this
> has remained fixed during all tests.
> 
> If I pick any of kernel-2.6.18-194.32.1.el5 or
> kernel-xen-2.6.18-238.1.1.el5 at the Grub's prompt displayed inside the
> vncviewer which I use to access the domU's console, it boots fine. Note
> that the kernel-xen package actually boots using the following lines:
> 
> title Scientific Linux SL (2.6.18-238.1.1.el5xen)
>        root (hd0,0)
>        kernel /boot/xen.gz-2.6.18-238.1.1.el5
>        module /boot/vmlinuz-2.6.18-238.1.1.el5xen ro root=LABEL=/
>        module /boot/initrd-2.6.18-238.1.1.el5xen.img
> 
> so that kernel is actually running on top of Xen which itself runs in
> the fully virtualized machine, which runs inside Xen on a physical machine.

Interesting, I thought it was impossible to run Xen under Xen and that this kind of "recursive virtualization" is an exclusive feature of KVM.

>> When you are running it on Gentoo, you probably set it up to not be
>> paravirtualized, so it happily ran the regular kernel.
> 
> In fact, the virt-manager run it via kvm, so without any traces of Xen
> at all. That particular physical machine has never had Xen on it.

KVM is readily available on SL5, so this may be a way to solve your actual problem.

Best regards,
	Stephan

>> If you are wondering, I did test the scenario you have.  I currently
>> have a xen host running 2.6.18-238.1.1.el5.  Some of it's clients
>> are/were running the older kernel, some 2.6.18-238.1.1.el5.  All of them
>> are working fine.
> 
> I've just updated kernel on another domU instance to the -238, again a
> fully virtualized one, and the symptoms are the same, ie. it won't boot
> and gets stuck on the serial thing. I suspect that both domUs have been
> installed via the same (or at least very similar) kickstart file via PXE.
> 
> I guess I can clean up the image and provide it for testing, if you
> think it could help debugging this issue. The same applies for the
> kickstart file.
> 
> Thank you for your help so far, I'm really lost at what I'm doing wrong
> here.
> 
> With kind regards,
> Jan
> 

-- 
Stephan Wiesand
DESY -DV-
Platanenenallee 6
15738 Zeuthen, Germany