#Open vSwitch Lab kaskdaksdkdaks
Get started with Open vSwitch, flows and OpenFlow controllers.
##Pre-reqs
Linux system with OVS installed.
##Setup
Use this script to add a bridge with ports for testing. The interfaces are moved into IP namespaces to isolate them from the main namespace and from each other.
#!/bin/bash
if [ $# -ne 2 ]; then
echo "usage: $0 port_num ovs_br"
exit 1
fi
set -xe
port=p$1
ns=ns$1
br=$2
mac=00:00:00:00:00:0$1
ip=10.0.0.${1}/24
ovs-vsctl --may-exist add-br $br
ovs-vsctl add-port $br $port
ovs-vsctl set Interface $port type=internal
ip netns add $ns
ip link set $port netns $ns
ip netns exec $ns ip link set $port address $mac
ip netns exec $ns ip address add $ip dev $port
ip netns exec $ns sysctl -w net.ipv6.conf.${port}.disable_ipv6=1
ip netns exec $ns ip link set $port up
Call script like:
# sh add-port.sh 1 br0
# sh add-port.sh 2 br0
Then the bridge should look like:
# ovs-vsctl show
e3784497-dc8f-432d-9a2c-923148962c73
Bridge "br0"
Port "p2"
Interface "p2"
type: internal
Port "br0"
Interface "br0"
type: internal
Port "p1"
Interface "p1"
type: internal
##Test 1 - the NORMAL flow
####Test connectivity:
# ip netns exec ns1 ping -c1 10.0.0.2
####Show flows
# ovs-ofctl dump-flows br0
NXST_FLOW reply (xid=0x4):
cookie=0x0, duration=560.596s, table=0, n_packets=10, n_bytes=828, idle_age=551, priority=0 actions=NORMAL
This flow gets created by default when you create a bridge.. The NORMAL flow causes the bridge to behave like a simple MAC learning switch. It applies to all ports because no in_port was specified and that is like a wildcard for all ports
####Show the mac table
# ovs-appctl fdb/show br0
port VLAN MAC Age
1 0 00:00:00:00:00:01 9
2 0 00:00:00:00:00:02 9
####Delete all flows
# ovs-ofctl del-flows br0
and ping again - it should fail this time.
####Re-add the NORMAL flow
# ovs-ofctl add-flow br0 actions=NORMAL
##Test 2 - forwarding by port numbers
####Delete all flows
# ovs-ofctl del-flows br0
####Find ofport numbers (OpenFlow port numbers)
# ovs-ofctl show br0
####Add the flows
# ovs-ofctl add-flow br0 in_port=1,actions=output:2
# ovs-ofctl add-flow br0 in_port=2,actions=output:1
####Test
# ip netns exec ns1 ping -c1 10.0.0.2
####Test 3 - forward by destination mac address
####Delete all flows
# ovs-ofctl del-flows br0
####Add the flows
# ovs-ofctl add-flow br0 dl_dst=00:00:00:00:00:01,actions=output:1
# ovs-ofctl add-flow br0 dl_dst=00:00:00:00:00:02,actions=output:2
# ovs-ofctl add-flow br0 dl_dst=ff:ff:ff:ff:ff:ff,actions=flood
####Test
# ip netns exec ns1 ping -c1 10.0.0.2
Use dump-flows and watch the packet counters. Wait several minutes for the ARP cache in the namespaces to expire or delete those entries, and ping again. That should cause the broadcast/flood flow to happen on next ping.
##OVS with an external OpenFlow controller
###Setup
Install the Ryu controller from here
####Start Ryu with the sample L2 learning module
$ ryu-manager ryu/ryu/app/simple_switch.py
loading app ryu/ryu/app/simple_switch.py
loading app ryu.controller.ofp_handler
instantiating app ryu/ryu/app/simple_switch.py of SimpleSwitch
instantiating app ryu.controller.ofp_handler of OFPHandler
####Point the test bridge at it
# ovs-vsctl set-controller br0 tcp:127.0.0.1:6633
####Check there are no flows yet
# ovs-ofctl dump-flows br0
NXST_FLOW reply (xid=0x4):
####Send some data
# ip netns exec ns1 ping -c1 10.0.0.2
####Check that the controller added new flows
# ovs-ofctl dump-flows br0
NXST_FLOW reply (xid=0x4):
cookie=0x0, duration=5.808s, table=0, n_packets=1, n_bytes=42, idle_age=0, in_port=2,dl_dst=00:00:00:00:00:01 actions=output:1
cookie=0x0, duration=0.808s, table=0, n_packets=0, n_bytes=0, idle_age=0, in_port=1,dl_dst=00:00:00:00:00:02 actions=output:2
####Check controller console for new log messages
packet in 270705776096578 00:00:00:00:00:01 00:00:00:00:00:02 1
packet in 270705776096578 00:00:00:00:00:02 00:00:00:00:00:01 2
packet in 270705776096578 00:00:00:00:00:01 00:00:00:00:00:02 1
####Study and change the sample code
See here