There was a mistake. The problem is on a new
Ubuntu-dom0/Ubuntu-domU system (not in CentOS). Replace CentOS with Ubuntu.
God, I must be feeling sleepy then. Bu the problem remains. I suspect there is
a clash between
/usr/local/lib and /usr/lib for python packages and the
site-packages/dist-packages saga. I've deleted all site-packages/xen thing with
PYTHON_PREFIX_ARG= (empty) ensured that packages went to right place which is I
Any help will be invaluable for me.
From: Bruce Edge [bruce.edge@xxxxxxxxx]
Sent: Monday, June 21, 2010 9:53 PM
To: Debasish Bose
Subject: Re: [Xen-users] Xen 4.0.1 rc3 pre install problem
This may or may not address your problem but it worked for Ubuntu systems with
a similar requirement, build once and distribute.
Apply the attached patch (not mine) to xen 4.0.1.rc3.pre and then build using:
export KERNELS="" ; export PYTHON_PREFIX_ARG="" ; make dist
This generates a dist dir that puts everything in the right place for Ubuntu
systems. I know the sysconfig/default changes apply to debian as well, the ones
I'm not sure about are the python lib location move.
This patch works for 4.0.0 and 4.0.1.rc3.pre, but not 4.1.
On Mon, Jun 21, 2010 at 9:05 AM, Debasish Bose
I've successfully ran dom0/domU (both Debian) in my local laptop
using 188.8.131.52 dom0 with xen-4.0.1-rc3-pre. Great. Now I want to install it
into couple of test-machines without compiling from source. So I did following..
a. make dist-xen
b. make dist-tools PYTHON_PREFIX_ARG="--install-layout=deb"
c. make dist-stubdom
This should pre-pare my dist/ directory with binaries. With the help of
top-level /.install.sh I should be able to install it any other machine (same
ARCH=x86_64/amd64). I've tar-zipped, copied and inflated. The target machine is
a CentOS 64-bit machine whereas the build machine (my laptop) is a
Debian/Ubuntu 10.04 64 bit one. After installation, my xend started,
bridge-magic happened, `brctl -show` is fine, xm info / xm list is fine, ls
/dev/xen is fine (evtchn,gntdev) too. But when I'm trying
to run a domU (using a working domU-xmexample-config) it has some problem with
tap driver config and complains
"ValueError: need more than 1 value to unpack"
File "/usr/lib/python2.6/dist-packages/xen/xend/XendConfig.py", line 342, in
File "/usr/lib/python2.6/dist-packages/xen/xend/XendConfig.py", line 857, in
cfg = self._parse_sxp(sxp_cfg)
File "/usr/lib/python2.6/dist-packages/xen/xend/XendConfig.py", line 720, in
self.device_add(dev_type, cfg_sxp = config, target = cfg)
File "/usr/lib/python2.6/dist-packages/xen/xend/XendConfig.py", line 1448, in
ret_uuid = self.device_duplicate_check(dev_type, dev_info, target, config)
File "/usr/lib/python2.6/dist-packages/xen/xend/XendConfig.py", line 1237, in
blkdev_file = blkdev_uname_to_file(dev_uname)
File "/usr/lib/python2.6/dist-packages/xen/util/blkif.py", line 95, in
File "/usr/lib/python2.6/dist-packages/xen/util/blkif.py", line 90, in
(taptype, fn) = fn.split(":", 2)[1:3]
ValueError: need more than 1 value to unpack
in xmexample config,
kernel = "/vm/tmp/vmlinuz-184.108.40.206-domU-deb"
ramdisk = "/vm/tmp/initrd.img-220.127.116.11-domU-deb"
memory = 256
name = "debdomU"
vif = [ 'mac=6A:60:41:4A:60:41,ip=172.16.30.201,bridge=eth1' ]
disk = [ "tap:aio:/vm/tmp/core.img,xvda1,w" ,
extra = "3 console=hvc0"
vmlinuz/initrd/core.img/swap.img are made on Ubuntu and just copied over.
The same config worked earlier on a ubuntu-dom0/ubuntu-domU setup. Is this
something to do with CentOS-dom0/Ubuntu-domU? Is copying binaries (pyc) built
in Ubuntu wrong? Then how even xend/xendomains got started? If xend started
with same python tool chain, what's wrong with domU booting up?
Please help me guys. Should I re-compile?
Xen-users mailing list
Xen-users mailing list