[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: [libvirt] [RFC Incomplete Patch] Libvirt + Openvswitch



Libvirt has a function virNetDevBridgeRemovePort() which can
remove port from the Linux Bridge, but it seems that no one calls it.

Wanted to confirm if port removal happens automatically for Linux
Bridges if VM goes down? The difference between OVS and 
Linux Bridge is that OVS will need a hook that removes all ports on
VM shutdown event (and maybe also for some other events?).

Thanks,
Ansis


On Tue, Jan 31, 2012 at 9:26 PM, Laine Stump <laine laine org> wrote:
On 01/31/2012 06:57 AM, Daniel P. Berrange wrote:
On Fri, Jan 27, 2012 at 02:58:58AM -0800, Dan Wendlandt wrote:
Hello all,

I know of many people who want to spin up VMs using libvirt + kvm/qemu and
attach those VMs to an openvswitch bridge (see: http://www.openvswitch.org).
  However, the only way I know of to get this working is a kludge that uses
to tap devices along with<interface type="ethernet">  while running
ovs-vsctl outside of libvirt.  Even worse, doing this on RHEL/Fedora seems
to require privilege tweaks (e.g., running qemu as root, not dropping
capabilities), which may not be acceptable for production deployments
(see:
http://fedoraproject.org/wiki/How_to_debug_Virtualization_problems#Errors_using_.3Cinterface_type.3D.27ethernet.27.2F.3E).

So I would like to start taking steps toward better libvirt/openvswitch
integration.  My initial step has the fairly limit goal of enabling
kvm/qemu VM NICs to attach to an openvswitch bridge in much the same way VM
NIC can already attached to the linux bridge.  For example, specifying:

<interface type="openvswitch">
<source bridge="br0"/>
<mac address="ca:fe:de;ad:be:ef"/>
</interface>
IMHO we should not be introducing a new "type" for OpenVSwitch. Contrary
to common understanding, type='bridge' is not referring explicitly to
Linux software bridging. Rather it refers to the concept of bridging the
guest to the LAN at the network level, of which Linux software briding
is one possible impl. OpenVSwitch is another possible impl. Other hypervisors
have different impls too of course.

If OpenVSwitch is available in the kernel, is there really any reason
to *not* use it ?  ie, could we just have

    <interface type="bridge">
      <source bridge="br0"/>
      <mac address="ca:fe:de;ad:be:ef"/>
    </interface>

and if we see that 'br0' is using OpenVSwitch, then libvirt can know
to just do the right thing.

Of course! <facepalm/> If the code can tell that br0 is an openvswitch bridge rather than a linux host bridge, we don't even have to add type='openvswitch' to the <source> element!

I definitely like this the best, too.



That way every application that uses
libvirt today will automatically be able to take advantage of the
benefits OpenVSwitch brings without further work

Except that an interfaceid is needed (or is that optional?)


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]