Xen in Hardy

[Mark Goldfinch]

Hi Leandro,

Leandro Pereira de Lima e Silva wrote:
> That's what I did:
> - I was trying to create a virtual machine with hardy on it (hardy on 
> the host too) using xen-create-image.
> - xen-tools isn't coming with hooks for installing hardy (I've already 
> filled a bug report about that), so I tried symlinking hardy.d to 
> edgy.d just like the symlinks from feisty.d and gutsy.d.
> - Runned: xen-create-image --force --fs ext3 --image-dev 
> /dev/group/volume_root --memory 512 --passwd --swap-dev 
> /dev/group/volume_swap --dist hardy --mirror 
> http://br.archive.ubuntu.com/ubuntu --gateway gateway_ip_addr --ip 
> ip_addr --netmask netmask_addr --hostname hostname_addr, where 
> "group/volume_root", "group/volume_swap", "gateway_ip_addr", 
> "ip_addr", "netmask_addr", "hostname_addr" are actually other things 
> that I've changed to get some privacy.
> - Tried to boot, but initramfs can't find the root.
> - Fixed a few things on the configuration file generated by 
> xen-create-image (tried different combinations, this is how it ended):
>   - Add "root        = '/dev/sda2 ro'" after the memory line.
>   - Fixed the disk section that just contained "',sda1,w'," inside of 
> it, at the end it became:
>       disk        = [ 'phy:group/volume_swap,sda1,w', 
> 'phy:group/volume_root,sda2,w' ]
> - Add "xenblk" to /etc/initramfs-tools/modules and recreated the 
> initrd file. (the same as the above item, what seemd to work after a 
> few tentatives)
> - Runned xm create -c hostname_addr
> - Machine starts to boot and hangs at:
>  * Running local boot scripts 
> (/etc/rc.local)                            [ OK ]
>
> - Switched to gutsy (in xen-create-image), thinking that maybe it 
> hangs because of some new hook needed for hardy.
> - Got the same problem that I had with hardy.
>
> The LVM group and the volumes existed previously, the machine is an 
> Athlon 64 X2 with 3.5gb of ram. I'm using a custom linux kernel 
> version 2.6.24.3 <http://2.6.24.3> based on the available source for 
> 2.6.24-12-xen. Tried also using the official kernel (2.6.24-12-xen) 
> with the original initrd but failed on the initrd problem that I 
> described above, tried it with a new initrd and it got stuck in the 
> same "Running local boot scripts [ OK ]" as above.
>
> I'm using xen-tools version 3.8-4ubuntu2 and xen-hypervisor-3.2 
> version 3.2.0-0ubuntu9.
>
> I tried to cover every aspect that may be relevant to the setup, 
> please tell me if I've missed anything.

Does the machine actually hang or are you missing just missing your 
login prompt?

If its the latter then it could be that Xen is running its /dev/console 
on a device which there isn't a getty running for.

Have a look at the configuration file for your Xen image, does it 
contain a like like either:

extra = "console=hvc0"

or

extra = "xencons=xvc console=xvc0"

If so, then you need to update upstart by creating a new file within 
/etc/event.d with the following contents:

========
# xvc0 - getty
#
# This service maintains a getty on xvc0 from the point the system is
# started until it is shut down again.

start on runlevel 2
start on runlevel 3
start on runlevel 4
start on runlevel 5

stop on runlevel 0
stop on runlevel 1
stop on runlevel 6

respawn
script
        if [ ! -c /dev/xvc0 ]; then
                mknod --mode=600 /dev/xvc0 c 204 191;
        fi

        exec /sbin/getty 38400 xvc0
end script
========

Then you can run "sudo start xvc0" to start your system console, or just 
reboot. If you feel like doing so you can remove all your tty[1-6] files 
from /etc/event.d as there's no virtual consoles on DomU hosts.

Hopefully this should then get you a working system console..

Thanks,
Mark.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.ubuntu.com/archives/ubuntu-server/attachments/20080403/91ec7784/attachment.html>