[ovs-discuss] dump flows in bridge mode
jpettit at nicira.com
Tue Nov 24 19:32:19 PST 2009
ovs-vswitchd sets up a default rule, which is completely wildcarded and has an action of "normal". This rules means that it acts as a learning switch. The rule looks something like this:
duration=4694s, table_id=1, priority=0, n_packets=0, n_bytes=0, actions=NORMAL
Are you seeing those packet and byte counters going up? Hopefully, you are.
The kernel datapath is a cache of flows that are hitting whatever the userspace flow table has configured. It times out entries that are idle for five seconds. If you want to see what those look like, you can run "ovs-dpctl show dp0". (The ovs-dpctl command is for interrogating the kernel datapath.)
If you configure ovs-vswitchd to connect to a controller and it sets up flows, you should see those in your "ovs-ofctl dump-flows". You can manually configure the ovs-vswitchd tables with "ovs-ofctl add-flow" and "ovs-ofctl del-flows".
Does this make sense?
On Nov 24, 2009, at 6:55 PM, Lee, Jeongkeun wrote:
> I'm running ovs as a replacement of a Linux bridge in a Xen dom0 server.
> When I run ping tests, 'ovs-ofctl show dp0' and dump-ports display substantial information (# of pkts are increasing), but dump-flows shows no flow entry.
> Does it mean that we don't have control over the flows unlike we do add/mod-flows on a OpenFlow switch?
> Thank you.
> Jeongkeun "JK" Lee
> Hewlett-Packard Labs
> Palo Alto, CA
> discuss mailing list
> discuss at openvswitch.org
More information about the discuss