Hello,
I’m trying to deploy Xen in a Gentoo machine
(Macbook Core 2 Duo), but I’m having a lot of questions and issues about
installation and operation of Xen, mainly with backend drivers.
The info is:
-
Machine: Macbook Core 2
Duo, it’s hardware is well known, but if needed I’ll post later my
lspci. It has T7200 Core 2 Duo with Intel
VT, SATA disk with Intel ICH7
controller, and Intel Gigabit Ethernet.
-
Gentoo: Installed from
2007.0, with 2.6.20.1 for normal kernel, and 2.6.20-r6 for xenfied kernel. With
normal kernel, all is working: Gnome 2.20 with 3D acceleration, sound, disks, keyboard,
touchpad, network (only with Ethernet, I’m not trying Atheros Wifi at the
moment). I’m not using initrd, my kernel have only the controllers I
need, and very few modules. So, SCSI support, SATA drivers for Intel ICH7 and
ext2/ext3, Intel GMA950 are inside the kernel, not compiled as modules.
-
Xen: I emerged latest
available version, 3.1.2, with xen-tools. No problem with compile and
installation (emerge didn’t complaint). To generate the xen dom0 kernel,
I used my working kernel adding usual dom0 options, like privileged domain, Virtual
PCI, backend drivers.
My goal is to get domUs with graphical support
(planned to use X/Gnome in that virtual machines). At first state, I’m
planning to store VM in files, maybe later I’ll migrate them to separate
partitions (my macbook have only one hard disk).
Well, my questions / issues:
-
After installation and
basic configuration, I added xend to default rc, and then rebooted. I always
get some error when I start xend (or its started at boot), mainly with blktap:
if I generate a dom0 kernel with only backend drivers (as told in Gentoo Xen
How To), xend says blktap0 is already mounted. But if I use a dom0 kernel with
backend drivers in kernel and fronted drivers as modules, xend complaints about
not available slot to mount
blktap0. Is there any way
to check status of /dev/blktap0?
-
When I try to start VM (images
downloaded from web), or create it with domi, I get non-successful boots,
always failed detecting boot partition inside domU (the known kernel panic
about not finding correct boot partition). I tried with ‘file:/var/lib/xen/fedora.img’
as sda1, or ‘tap:aio:/var/lib/xen/fedora.img’ as xvda or xvda1, but
with no luck. I suspect from the above error with blktap0, but I’m not
sure if my dom0 is giving right filesystem to domU. Any method / utility to
check that?
-
My last movement was to
install Virtual Manager (which is in the portage tree), and use virt-install
tool to create the VM. When I try that, with Fedora 7 as domU, in the boot
sequence I get timeouts for vkb0 and vb0 (virtual keyboard and net, I suppose),
and the well known kernel panic abot hotplug scripts not working (but I get
that with vkbd0), and I didn’t found much information about that. What
could I check? Dom0 hotplug scripts are working well, as far I know…Maybe
I need to add some options to my kernel?
I’m getting stuck at that point, I’m not
sure about my kernel dom0 config, maybe someone can post a working dom0 kernel
config with gentoo?
Thanks for your patience reading this, english is not
my mother tongue ;-)
