[sfc-dev] Problem when setting up sfc103


Alioune
 

Hi Brady,

I'm using distribution-karaf-0.5.2-Boron-SR2 since the ODL version used in demo103 (stable/boron) is removed from nexus [1] and I  provision nodes set by step without changing IP addresses.


-  br-sfc bridges of classifiers are set to "Controller "tcp:192.168.1.5:6653" and I expected to see Controller "tcp:192.168.1.1:6633".

- I observe also that br-sfc bridges don't appear in node/topology after installing "odl-dlux-node"

Therefore I think that the root of this issue could be OVSDB (see logs [2) or the attached file).
I take these logs before provisioning the service chain (python setup_sfc.py)

Can OVSDB team tells what could cause the warnings in logs ?

Best regards,



[1] DIST_URL=https://nexus.opendaylight.org/content/repositories/opendaylight.snapshot/org/opendaylight/integration/distribution-karaf/0.5.0-SNAPSHOT/

[2] https://docs.google.com/document/d/1sS1Qxy81_sigXQQoRwtdWwN-t6NACpehAafUPPCwlK0/edit?usp=sharing

[3] vagrant@classifier1:~$ sudo ovs-vsctl show
36976c5e-4f92-4dc1-8116-b0eaa5ae43e1
    Manager "tcp:192.168.1.1:6640"
        is_connected: true
    Bridge br-sfc
        Controller "tcp:192.168.1.1:6653"
            is_connected: true
        Port br-sfc
            Interface br-sfc
                type: internal
        Port veth-br
            Interface veth-br
        Port "sff0-dpl"
            Interface "sff0-dpl"
                type: vxlan
                options: {dst_port="6633", exts=gpe, key=flow, "nshc1"=flow, "nshc2"=flow, "nshc3"=flow, "nshc4"=flow, nsi=flow, nsp=flow, remote_ip=flow}
    ovs_version: "2.5.90"

2017-06-29 20:12 GMT+02:00 Brady Johnson <bradyallenjohnson@...>:


Alioune,

I see an issue in the output of "ovs-vsctl show". Im expecting to see the Ip/Port of the controller for the br-sfc bridge, but its not there. Normally, you should see

    Controller "tcp:192.168.1.1:<cant remember the OpenFlow port>" 
        is_connected: true

which is the connection used by ODL to write the OpenFlow flows on the bridge.

That is directly related to the following WARN in the logs, which is the root of the rest of the problems that you'll see in the logs.

2017-06-29 18:50:37,776 | WARN  | on-dispatcher-40 | SfcOvsUtil                   | 307 - org.opendaylight.sfc.ovs - 0.3.2.Boron-SR2 | OVS Node for IP address 192.168.1.20 does not exist!

Have you modified the SFF configurations in the demos or change the IP addresses of any of the VMs?

Regards,

Brady



On Thu, Jun 29, 2017 at 7:08 PM Alioune BA <alioune.ba.am@...> wrote:
Hi All,

I'm trying to build the sfc103 demo by creating vagrant vBox for each node.

This demo was working for me at the end of last year but I get some errors when setting it up see [1] or the attached file.

It seems that this is due to SFF1 initialization but the br-sfc into SFF1 is connected to odl and can ping.

vagrant@sff1:~$ ifconfig eth1
eth1      Link encap:Ethernet  HWaddr 08:00:27:00:73:2a 
          inet addr:192.168.1.20  Bcast:192.168.1.255  Mask:255.255.255.0
          inet6 addr: fe80::a00:27ff:fe00:732a/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:340 errors:0 dropped:0 overruns:0 frame:0
          TX packets:675 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:31253 (31.2 KB)  TX bytes:81757 (81.7 KB)

vagrant@sff1:~$ sudo ovs-vsctl show
36976c5e-4f92-4dc1-8116-b0eaa5ae43e1
    Manager "tcp:192.168.1.1:6640"
        is_connected: true
    Bridge br-sfc
        Port br-sfc
            Interface br-sfc
                type: internal
    ovs_version: "2.5.90"

I used both distribution-karaf-0.5.0-Boron and distribution-karaf-0.5.2-Boron-SR2 but still get the same issue.

Did someone get this error before ?
Any suggestion to solve that.

Best Regards

[1] https://docs.google.com/document/d/1sS1Qxy81_sigXQQoRwtdWwN-t6NACpehAafUPPCwlK0/edit?usp=sharing