Openstack responds to all ARP requests and replies with the external bridge ip

asked 2015-08-19 18:45:20 -0500

tejasgandhi999 gravatar image

We have multinode Openstack Kilo setup. It was working normally till few days ago. While creating a bulk of VM's our neutron server died and we had to restart it. While restarting neutron server some of the agents died. We restarted them also. After doing all this we found that compute nodes were responding to all the ARP requests, even for the IP which were not on the network.

We have tried restarting all services but nothing works. Clearing the neurton agents and re adding the agents. But of no help. The compute nodes still responds to each ARP requests on br-ex.

Our setup has 1 controller node, 1 network node and 20 compute node. We have DVR setup. All the nodes are on centos 7.

Our External Network 172.27.0.0/16 Management Network 10.0.0.0/24 Instance Network 10.0.1.0/24

The configuration of Netron Server which is running on controller is neutron.conf http://paste.openstack.org/show/422292/ ml2_conf http://paste.openstack.org/show/422299/

On Network Node neutron.conf http://paste.openstack.org/show/422300/ ml2_conf http://paste.openstack.org/show/422301/ metadata_agent.ini http://paste.openstack.org/show/422302/ l3_agent http://paste.openstack.org/show/422303/ dhcp_agent http://paste.openstack.org/show/422304/

On Compute Node neutron.conf http://paste.openstack.org/show/422305/ ml2_conf http://paste.openstack.org/show/422306/ metadata_agent.ini http://paste.openstack.org/show/422307/ l3_agent http://paste.openstack.org/show/422308/

edit retag flag offensive close merge delete