Re: [Xen-users] various minor problems

["Fajar A. Nugraha" <fajar@xxxxxxxxx>]

On Fri, Jun 25, 2010 at 11:43 PM, Doublehp <doublehp-xen@xxxxxxxxxxxx> wrote:
> 1: is it possible to use ioemu:sda instead of ioemu:hda ?

I assume this is a HVM domU?
On new version of Xen anyway, yes, it's possible (I've used hda, sda,
xvda). But in most cases it won't matter.

First, I'm not sure if you should put ioemu: in the first place. It
does harm in the past (prevents GPLPV from working), and the domU
works just fine without it.

Second, on non xen-aware HVM domU (like Windows, or Linux without PV
drivers), it would still detect the same thing (first disk on primary
IDE controller).

For PV domU, you should never put ioemu. As for which disk (hda, sda,
xvda), older kernel version (like 2.6.18-xen) used to respect them and
present it as-is to domU. Newer kernels (like pv_ops) will silently
change it to xvda no matter what you put in domU config file.

So why do you want to put sda instead of hda?

>
> 2: is ioemu keyword optionnal ? conf seems to work without

>From what I understand, it started out to mean something, using
emulated device (i.e. using qemu) instead of PV device. It has been
somewhat useless for a long time (and could even cause harm, like I
mentioned earlier).

Perhaps others can give a better explanation about this. The short
version is I don't see a reason why you should still put ioemu.

>
> 3: why d i need to put vfb = [ "type=vnc,vncunused=1,keymap=fr" ] line ?
> seperate vnc = 1 + vncunused = 1 look identical to me, but are ignored.
> have the conf parser changed ?

Historically, the virtual frame buffer (vfb) which is accessible via
vnc/sdl is only available for HVM domU. vfb for domU was added later,
using a different syntax (vfb line). So there are two different
syntax, one for HVM, and one for PV.

If you ask "why not make the syntax the same since the functionality
is the same", then you have to ask the developers.

>
> 4: during dom0 startup:
>> * Starting Xen control daemon ...
>> * Error: either "local" is duplicate, or "192.168.0.80/32" is a
> garbage.       [ ok ]
>
> before I do <<many things with my conf>>, it used to complain about
> inet6. So, it seems the problem is not really IPv6 support, but
> something else. This may be Gentoo specific issue, or even kernel specific.
>
> 192.168.0.80 is an alias for eth0; after Xend, all my IPs are removed,
> and I have to re-declare all of them manually.

It might be related to the fact that xen's default network bridge
script does some interface renaming and ip address/route transfer. In
your case it doesn't work so well. I suggest you create bridges
manually using your OS config files, and comment out network-script
line on xend-config.sxp. Note that this approach requires some basic
knowledge on how Linux bridging works.

-- 
Fajar

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users