[Pkg-libvirt-maintainers] virt-manager not yet in testing

[Guido Günther]

On Thu, Feb 21, 2008 at 10:47:12PM +0100, Henning Sprang wrote:
> Hmm - you probably mean you used it with KVM or qemu as of yet?!
That's KVM yes.
> 
> > In that case we had several paravirt
> >  instances already. What we did was starting the instances with xm,
> >  then running
> >
> >  virsh dumpxml domain-id > bla.xml
> >  virsh define bla.xml
> 
> Sounds interesting - thanks for the hint!
Note there's a bug in the current libvirt code concerning the bootlader
XML tag on paravirt guests, you'll need something like:

diff --git a/src/xend_internal.c b/src/xend_internal.c
index fd38a61..5428cd7 100644
--- a/src/xend_internal.c
+++ b/src/xend_internal.c
@@ -1438,9 +1438,6 @@ xend_parse_sexp_desc(virConnectPtr conn, struct sexpr *root,
         if (tmp != NULL) {
             bootloader = 1;
             virBufferVSprintf(&buf, "  <bootloader>%s</bootloader>\n", tmp);
-        } else if (sexpr_has(root, "domain/bootloader")) {
-            bootloader = 1;
-            virBufferVSprintf(&buf, "  <bootloader/>\n");
         }
         tmp = sexpr_node(root, "domain/bootloader_args");
         if (tmp != NULL && bootloader) {

to make this work with paravirt guests. The fix isn't perfect and
therefore not in libvirt CVS. I didn't get around to clean this up yet.

> >  Afterwards things can be mainained via virt-manager nicely.  But you're
> >  right virt-manager/virt-install need Debian support for xen paravirt
> >  guests. I might need to write something along these lines in the next
> >  weeks shouldn't we switch to KVM entirely.
> 
> I cannot promise to get too much involved, as I admin Xen from the
> console. But I will test whatever you do :)
Then virsh/virtinst/virt-viewer are your friends. Bootstapping Debian
with these is quiet nice. I use these tools far more than virt-manager
itself.

[..snip..]
> So, looks like I can report to the Debian folks at FOSDEM now, that we
> have the same tools available as the commercial Linux guys :)
Yes please do, I'll be around too.
Cheers,
 -- Guido