No instance IP address for second and subsequent instances [closed]

asked 2014-01-14 19:47:21 -0500

tomwaterhouse gravatar image

I am new to OpenStack and advanced networking, so may be missing something obvious.

I've setup OpenStack Havana on CentOS 6.5 using RDO. I have configured it as a multi-node deployment on Cisco UCS blades. I am using Neutron networking.

The short version of the story is that when I create a single instance in OpenStack that instance is able to get an IP address. When I create a subsequent instances each instance does NOT get an IP address. OpenStack does allocate the address but the VM instance cannot reach the DHCP server.

Example - I create the following network: neutron net-create private-network neutron subnet-create private-network 10.0.0.0/24 --name private-subnet

Then I create two instances via the UI: test1 - 10.0.0.2 test2 - 10.0.0.4

When getting information for the network from the OpenStack UI I see that 10.0.0.3 is allocated for DHCP.

Using 'ip netns exec <network namespace=""> tcpdump -i any' I can see DHCP traffic for 10.0.0.2. I don't see any for 10.0.0.4. </network>

Using 'ip netns exec <network namesapece=""> ping 10.0.0.2' is successful but not for 10.0.0.4.</network>

I have also done the above but with network 192.168.1.0 and get the same results. My goal was to create a second network and see if I could replicate the issue.

Any suggestions would be appreciated, and I would be more than happy to generate any logs/output necessary to troubleshoot this.

Regards,

Tom

edit retag flag offensive reopen merge delete

Closed for the following reason question is not relevant or outdated by larsks
close date 2014-04-25 11:08:44.405294

Comments

I have experienced this. This is intermittent issue. Can you try restarting your services couple of times and see how it goes.

dheeru gravatar imagedheeru ( 2014-01-14 21:31:11 -0500 )edit