On Fri, 26 Jan 2007, Marc Teichgraeber wrote:
* How would a HVM-config file look like? It would be nice to know to
understand the difference.
here is a hvm config file I use to do installs using a systemimager setup
(it boots off the iso, but it reads the systemimager config off the floppy
image)
kernel = "/usr/lib/xen/boot/hvmloader"
device_model = '/usr/lib64/xen/bin/qemu-dm'
builder='hvm'
memory = 512
shadow_memory = 8
vcpus=1
vif = [ 'type=ioemu, bridge=xenbr0' ]
nographic=1
sdl=0
vnc=0
serial='pty'
name = "VM1-64"
fda='/var/tmp/localcfg.img'
disk = ['file:/var/tmp/systemimager.iso,hdb:cdrom,w',
'phy:/dev/sda5,hda,w']
boot='d'
I do everything off xm console here; thus the nographic=1. if you are
using a graphical install, you will want to enable vnc or sdl.
Also note; even if you are using nographic like this, *you still need to
install the libvncserver libraries or else your hvm domU won't load* -
this tripped me up the other day.
Of course, this only works on the new CPUs that support VT.
* Why can I not do this 'boot from install media' stuff in not HVM-configs?
I dont understand this, i mean, the prophet has to climb the mountain
someway, respectively the guest-operating system hast to enter the host.
Momentarily I rsync one instance to another and make it unique afterwards or
rsync from a "real" System into sparse files on my guest an make them
bootable the.
Xen in paravirtualized mode reads the kernel from the kernel file in the
Dom0 that you specify on the kernel= config line. (note kernel= line in
my HVM config points at the 'hvmloader' program) this kernel needs to be
specially modified to run within Xen if you are in paravirtualized mode;
With qemu (or with HVM; if you notice, HVM uses a whole lot of qemu
code) you can use an unmodified kernel; and qemu (and hvm) supplies
enough bios emulation to read the bootloader/kernel off your install
media.
_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users
|
Home