This is an archived copy of the Xen.org mailing list, which we have preserved to ensure that existing links to archives are not broken. The live archive, which contains the latest emails, can be found at http://lists.xen.org/
Home Products Support Community News


Re: [Xen-devel] Fw: [Xen-API] status of xen api

xen-devel-bounces@xxxxxxxxxxxxxxxxxxx wrote on 10/31/2007 08:18:55 AM:

> On Tue, Oct 30, 2007 at 08:41:24PM -0400, Stefan Berger wrote:
> > > * I can't get the VDIs for a VBD (nor the VDIs for a SR). How am I
> > >   supposed to get the backend of a VM's vbd?
> > >
> > >   If I call the VBD.get_VDI, I get an empty value for the ref
> > >   (communication attached).
> >
> > Is this a domain that happens to only have a ramdisk?
> No, it's a domain with a single block (and net) device.
> > If you create a domain with access to a disk and use xm in xen-api mode
> > then does this still happen?  To do this you have to activate the 'server'
> > node in /etc/xen/xm-config.xml and add (xen-api-server ((
> > none ))) to /etc/xen/xend-config.sxp.
> Thanks for the tip about xen/xend-config.sxp, I'll be able, at least, to
> capture and analyze the traffic. ;)
> And you're correct. With xm in xen-api mode, the newly created VM has
> the correct VDI entry.
> > >
> > > * my network object has no PIFs, is this normal?
> >
> > Not sure about this, though I'd say that the network should have a list of
> > PIFs with at least one PIF in it, which represents 'peth0'.
> Hm, I personally don't use peth0. I use a bridge with IP address and
> attach the real eth0 to the bridge with my distro's usual configuration
> scripts.

> So then, without peth0, this is expected?

Maybe this has changed over time, but at least now peth0 is the physical interface and eth0 is the bridge. So I would expect that now a network object contains a PIF with name 'peth0' and if I remember correctly the last time I check it did.

Xen-devel mailing list
<Prev in Thread] Current Thread [Next in Thread>
  • Re: [Xen-devel] Fw: [Xen-API] status of xen api, Stefan Berger <=