Hi all.
I install the OpenvSwitch 2.3.1 on my system with linux 3.10.77.
My system has 4 physical interface eth0, eth1, eth2 and eth3.
I create 4 OVS bridge for each ethX interfaces and attch the ethX to its
OVS bridge.
The topology likes below.
ovs_eth0 ovs_eth1 ovs_eth2 ovs_eth3
| | | |
eth0 eht1 eht2 eht3
After attch the ethX interface, I do the following thing.
1. Clone the mac address of ethX to ovs_ethX.
2. Set the ip address of ehtX to 0.
3. Set the static ip to ovs_ethX.
It looks like below.
----
eth0 Link encap:Ethernet HWaddr 00:11:32:32:93:99
inet6 addr: fe80::211:32ff:fe32:9399/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:4022865 errors:0 dropped:0 overruns:0 frame:0
TX packets:11384 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:883577555 (842.6 MiB) TX bytes:2135843 (2.0 MiB)
ovs_eth0 Link encap:Ethernet HWaddr 00:11:32:32:93:99
inet addr:10.11.243.26 Bcast:10.11.243.255 Mask:255.255.255.0
inet6 addr: fe80::211:32ff:fe32:9399/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:4013419 errors:0 dropped:0 overruns:0 frame:0
TX packets:11389 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:882582231 (841.6 MiB) TX bytes:2136397 (2.0 MiB)
----
In my topology, I don't use any controller to control the flow, I use the
normal action in the flow table.
----
NXST_FLOW reply (xid=0x4):
cookie=0x0, duration=283795.407s, table=0, n_packets=4035185,
n_bytes=885870336, idle_age=0, hard_age=65534, priority=0 actions=NORMAL
-----
Atfet those setting, the network works.
Now. I write a script to do all thing I mentioned before and it will excute
after the system has boot done.
After reboot, sometimes it works, sometimes it doesn't work.
The ratio of goot do bad is 14000 to 2. I want to figure out why the
network doesn't work.
In that condtion. Only the ovs_eth0 can't reply packet but ovs_eth1,
ovs_eth2 and ovs_eth3 works.
I use the tcpdump to listen the packet. I found that the ovs_eth0 has
received the ARP request but didn't reply the ARP packet.
I try to check the network setting but I can't figure out.
ifconfig
------
eth0 Link encap:Ethernet HWaddr 00:11:32:32:93:99
inet6 addr: fe80::211:32ff:fe32:9399/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:4026717 errors:0 dropped:0 overruns:0 frame:0
TX packets:11384 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:884523841 (843.5 MiB) TX bytes:2135843 (2.0 MiB)
ovs_eth0 Link encap:Ethernet HWaddr 00:11:32:32:93:99
inet addr:10.11.243.26 Bcast:10.11.243.255 Mask:255.255.255.0
inet6 addr: fe80::211:32ff:fe32:9399/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:4017284 errors:0 dropped:0 overruns:0 frame:0
TX packets:11389 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:883539398 (842.6 MiB) TX bytes:2136397 (2.0 MiB)
------
ovs-vsctl show
--------
Bridge "ovs_eth0"
Port "ovs_eth0"
Interface "ovs_eth0"
type: internal
Port "eth0"
Interface "eth0"
--------
ovs-ofctl dump-flows ovs_eth0 (the n_packets counter will increase)
--------
NXST_FLOW reply (xid=0x4):
cookie=0x0, duration=284607.822s, table=0, n_packets=4039935,
n_bytes=887091701, idle_age=0, hard_age=65534, priority=0 actions=NORMAL
--------
I also try to flush the fdb and the entries has been add again later but it
doens't solve the problem.
First time I met this problem, I solve this by using ifconfig to assign the
same ip address again.
Does any one has the similar experience of OVS before?
Thanks.
Best Regards,
Hung-Wei Chiu
_______________________________________________
discuss mailing list
discuss@openvswitch.org
http://openvswitch.org/mailman/listinfo/discuss
