OpenDaylight cluster not working

asked 2017-01-10 09:07:17 -0700

anonymous user

Anonymous

updated 2017-01-10 12:05:37 -0700

Hi guys,

I’m having a few issues getting clustering to work in release Beryllium 0.4.4 and I wondered if anyone would be able to help. I’ve installed the following features:

feature:install odl-restconf odl-l2switch-switch odl-mdsal-apidocs odl-dlux-all odl-mdsal-clustering odl-openflowplugin-flow-services http bundle:install -s mvn:org.jolokia/jolokia-osgi/1.1.5

I have run the following script on each virtual machine that hosts OpenDaylight, “./configure-cluster-ipdetect.sh 192.168.56.103 192.168.56.104”, with .103 and .104 being the two instances of OpenDaylight.

On a third VM, I then load Mininet with “sudo mn --controller remote,ip=192.168.56.103,port=6633 --topo linear,10”. At this stage a pingall completes successfully, but from what I can see the cluster isn’t actually being utilised. Therefore I connect each switch to the second controller using “sudo ovs-vsctl set-controller s1 tcp:192.168.56.103:6633 tcp:192.168.56.104:6633”.

I believe HA clustering should now be fully functional, however, a pingall fails completely – none of the hosts are able to ping each other. I’ve used the GET command to verify that the cluster has been formed and using Wireshark I can see that each instance of OpenDaylight is communicating with each other.

Running the “sudo ovs-vsctl list controller” command, I can see that the “role” for each controller connection is set to “other”, when this should be “master” or “slave”. A bit of research suggested I need to be running OpenFlow version 1.3 for this to function, so I removed the comment in /etc/custom.properties for the “ovsdb.of.version=1.3” entry and this time loaded Mininet with the command “sudo mn --controller remote,ip=192.168.56.103,port=6633 --topo linear,10 --switch=ovsk,protocols=OpenFlow13”.

Now each controller connection successfully registers as a “master” or “slave”, but I am still not able to ping between hosts within Mininet.

Does anyone have any suggestions as to what the problem may be? I also get the same issue when running a three node cluster.

Thank you for your time

EDIT: Checking the logs, I can see the following error repeated several time:

2017-01-10 19:54:20,972 | ERROR | OFRpc - 3 | RPCBroker | 165 - org.opendaylight.controller.sal-remoterpc-connector - 1.3.4.Beryllium-SR4 | executeRpc for (urn:opendow:service?revision=2013-08-19)add-flow failed with root cause: java.lang.NullPointerException. For exception details, enable Debug logging.

edit retag flag offensive close merge delete