Hi, Todd. I've had limited exposure to KVM so far, but I'll do my best to help
you out. I actually started to set up a KVM environment last week, but some
other things became higher priority. I'm hoping to have it completed this week.
The ovs-dpctl program communicates directly with the kernel datapath. The
vswitchd.conf file is used to configure ovs-vswitchd, which runs in userspace.
What you're currently doing is adding interfaces to the kernel's datapath, but
ovs-vswitchd doesn't know about them. That's why you're seeing those messages
about packets coming on unknown interfaces. Instead, try adding those
interfaces directly to vswitchd.conf (and don't run the add-if command, since
ovs-vswitchd will take care of adding them to the kernel). Your config file
will look something like this:
bridge.internalbr.port=dummy0
bridge.internalbr.port=tap0
bridge.internalbr.port=tap1
You can modify the file whenever you want and HUP the ovs-vswitchd process to
have it reload the configuration.
Let me know if that gets you any further along.
--Justin
On Feb 7, 2010, at 9:04 PM, Todd Deshane wrote:
> Hi All,
>
> I am trying to test some basic functionality with kvm guests.
>
> I downloaded the openvswitch-0.99.1 unstable tarball release.
>
> I created a dummy0 interface (I've tested on both Ubuntu 9.10 and Fedora 12)
>
> My ovs-vswitchd.conf is simply:
> bridge.internalbr.port=dummy0
>
> I start two kvm guests with:
>
> kvm -cdrom ~/Desktop/iso/ubuntu-9.10-desktop-i386.iso -net
> nic,model=e1000 -net
> tap,script=internal-ifup,downscript=internal-ifdown -m 1024
>
> the internal-ifup script:
> -------------------------------------------------------------
> #!/bin/sh
>
> switch=internalbr
>
> /usr/local/bin/ovs-dpctl del-if $switch $1
> /sbin/ifconfig $1 0.0.0.0 down
> -------------------------------------------------------------
>
>
> the internal-ifdown script
> -------------------------------------------------------------
> #!/bin/sh
> switch='internalbr'
> '''sudo''' /sbin/ifconfig $1 0.0.0.0 up
> '''sudo''' /usr/local/bin/ovs-dpctl add-if ${switch} $1
> exit 0
> -------------------------------------------------------------
>
> ovs-dpctl show (adds the tap interfaces as expected)
> dp0:
> flows: cur:0, soft-max:512, hard-max:262144
> ports: cur:4, max:1024
> groups: max:16
> lookups: frags:0, hit:49, missed:122, lost:0
> queues: max-miss:100, max-action:100
> port 0: internalbr (internal)
> port 1: dummy0
> port 2: tap0
> port 3: tap1
>
>
> I hard-coded the IP addresses in the guests, but they can't ping each other.
>
> The output of running: ovs-vswitchd ovs-vswitchd.conf
> Feb 07 23:46:17|00001|cfg|INFO|using "ovs-vswitchd.conf" as
> configuration file, ".ovs-vswitchd.conf.~lock~" as lock file
> Feb 07 23:46:18|00002|ofproto|INFO|using datapath ID 002320dc5ff5
> Feb 07 23:46:18|00003|bridge|INFO|created bridge internalbr on dp0
> Feb 07 23:46:18|00004|bridge|INFO|created port dummy0 on bridge internalbr
> Feb 07 23:46:18|00005|ofproto|INFO|datapath ID changed to 8efae2549dde
> Feb 07 23:49:24|00006|bridge|WARN|bridge internalbr: received packet
> on unknown interface 2
> Feb 07 23:49:24|00007|bridge|WARN|bridge internalbr: received packet
> on unknown interface 2
> Feb 07 23:49:24|00008|bridge|WARN|bridge internalbr: received packet
> on unknown interface 2
> Feb 07 23:49:25|00009|bridge|WARN|bridge internalbr: received packet
> on unknown interface 2
> Feb 07 23:49:29|00010|bridge|WARN|bridge internalbr: received packet
> on unknown interface 2
> Feb 07 23:50:51|00011|bridge|WARN|Dropped 10 log messages in last 79
> seconds due to excessive rate
> Feb 07 23:50:51|00012|bridge|WARN|bridge internalbr: received packet
> on unknown interface 3
> Feb 07 23:51:34|00013|bridge|WARN|Dropped 5 log messages in last 43
> seconds due to excessive rate
> Feb 07 23:51:34|00014|bridge|WARN|bridge internalbr: received packet
> on unknown interface 2
>
> I am assuming that there are other debugging techniques that would be
> helpful, but was wondering if I am doing something that is obviously
> wrong or if I am running into some simple known problem.
>
> I can and will try to debug further.
>
> I did also tried using eth0 as switch (bridge) and it similarly
> doesn't work on Ubuntu and on Fedora I get an "eth0 device in use"
> error. I do have some specific applications where I will need to use
> both an internal bridge (such as dummy0) and an external bridge (such
> as eth0).
>
> Any ideas or suggestion on thing that I could try?
>
> Thanks,
> Todd
>
> --
> Todd Deshane
> http://todddeshane.net
> http://runningxen.com
>
> _______________________________________________
> discuss mailing list
> discuss@openvswitch.org
> http://openvswitch.org/mailman/listinfo/discuss_openvswitch.org
_______________________________________________
discuss mailing list
discuss@openvswitch.org
http://openvswitch.org/mailman/listinfo/discuss_openvswitch.org
