Unable to start some VM with Debian 6

Hi,

on Cloudmin GPL running on CentOS 5 + Xen some new VM based on Debian 6 64 bit are unable to start:

xm create /xen/smtpsh2.cfg

Using config file “/xen/smtpsh2.cfg”.
Using <class ‘grub.GrubConf.GrubConfigFile’> to parse /boot/grub/menu.lst
Traceback (most recent call last):
File “/usr/bin/pygrub”, line 691, in ?
chosencfg = run_grub(file, entry, fs)
File “/usr/bin/pygrub”, line 545, in run_grub
g = Grub(file, fs)
File “/usr/bin/pygrub”, line 203, in init
self.read_config(file, fs)
File “/usr/bin/pygrub”, line 407, in read_config
self.cf.parse(buf)
File “/usr/lib64/python2.4/site-packages/grub/GrubConf.py”, line 287, in parse
self.add_image(GrubImage(title, img))
File “/usr/lib64/python2.4/site-packages/grub/GrubConf.py”, line 135, in init
_GrubImage.init(self, title, lines)
File “/usr/lib64/python2.4/site-packages/grub/GrubConf.py”, line 86, in init
self.reset(lines)
File “/usr/lib64/python2.4/site-packages/grub/GrubConf.py”, line 104, in reset
self._parse(lines)
File “/usr/lib64/python2.4/site-packages/grub/GrubConf.py”, line 98, in _parse
map(self.set_from_line, lines)
File “/usr/lib64/python2.4/site-packages/grub/GrubConf.py”, line 142, in set_from_line
setattr(self, self.commands[com], arg.strip())
File “/usr/lib64/python2.4/site-packages/grub/GrubConf.py”, line 107, in set_root
self._root = GrubDiskPart(val)
File “/usr/lib64/python2.4/site-packages/grub/GrubConf.py”, line 56, in init
self.disk = str
File “/usr/lib64/python2.4/site-packages/grub/GrubConf.py”, line 69, in set_disk
self._disk = int(val[2:])
ValueError: invalid literal for int(): ev/xvda1
No handlers could be found for logger “xend”
Error: Boot loader didn’t return any data!
Usage: xm create [options] [vars]

cat /xen/smtpsh2.cfg

memory = 512
maxmem = 4096
name = ‘smtpsh2’
vif = [ ‘ip=192.168.113.199,mac=00:16:3e:4B:62:43’ ]
address = ‘192.168.113.199’
netmask = ‘255.255.255.128’
disk = [‘phy:/dev/vol1tb/smtpsh2_img,sda1,w’,‘phy:/dev/vol1tb/smtpsh2_swap,sda2,w’]
vnc = 1
vnclisten = “0.0.0.0”
vncunused = 1
vncpasswd = “ciao”
vfb = [‘type=vnc,vncunused=1,vncpasswd=ciao,vnclisten=0.0.0.0’]
vcpus = 2
bootloader = “/usr/bin/pygrub”

On the same server, another similar installation can run fine with same settings. The only way to start this VM is change the default boot loader to Host System Kernel.

How can this problem be solved?
Thanks

Hi,

the solutions to the problems is to edit the /boot/grub/menu.lst of the VM (mount the device as loop) and edit the auto-config insert after the Debian kernel upgrade!