Ask Your Question

Openstack VM Unable to Get DHCP IP, Request but No Replay [closed]

asked 2013-07-14 03:48:02 -0600

skyrainman gravatar image


I am having an issue with my VM unable to get an IP from the DHCP server. My setup is as follows:

1xController w/Quantum, namespaces false 1xCompute 1xCinder

After running "tcpdump -n -i eth2" I can see traffic being sent from my compute node, however, I do not see the packets being received by my controller/quantum node. However, I can see all other traffic on both nodes, and I can ping from one node to the other on the same interface if I add an IP to the bridge.

Below I have given the output of ovs-vsctl show. Also, when I add a floating ip to my VM, I do not see the IP come up on the controller/quantum node. If I add the IP manually, I can't ping the DHCP server.

My NIC setup is as follows:

eth0 - management eth1 - public eth2 - data/vm

I am using the openvswitch plugin with vlans. The only error I can see on both servers is:

013-07-14T08:10:04Z|00247|bridge|ERR|bridge br-prv: failed to set bridge Ethernet address: Device or resource busy 2013-07-14T08:10:07Z|00248|netdev_linux|ERR|ioctl(SIOCSIFHWADDR) on br-int device failed: Device or resource busy 2013-07-14T08:10:07Z|00249|bridge|ERR|bridge br-int: failed to set bridge Ethernet address: Device or resource busy 2013-07-14T08:10:07Z|00250|netdev_linux|ERR|ioctl(SIOCSIFHWADDR) on br-prv device failed: Device or resource busy 2013-07-14T08:10:07Z|00251|bridge|ERR|bridge br-prv: failed to set bridge Ethernet address: Device or resource busy

Here is my output from ovs-vsctl show on controller:

[root@controller-01 ~(keystone_admin)]# ovs-vsctl show d44ed3f4-2e5e-4d5f-8a41-e6b1664402fb Bridge br-prv Port "eth2" Interface "eth2" Port br-prv Interface br-prv type: internal Port phy-br-prv Interface phy-br-prv Bridge br-ex Port "eth1" Interface "eth1" Port phy-br-ex Interface phy-br-ex Port br-ex Interface br-ex type: internal Bridge br-int Port int-br-prv Interface int-br-prv Port "tapc9ec4cea-28" tag: 1 Interface "tapc9ec4cea-28" type: internal Port int-br-ex Interface int-br-ex Port br-int Interface br-int type: internal ovs_version: "1.10.0"

Here is my output from ovs-vsctl show on compute: [root@compute-01 ~(keystone_admin)]# ovs-vsctl show e84211f2-9d94-4851-9357-ef73c050e240 Bridge br-int Port "qvo4b11e62b-a4" tag: 1 Interface "qvo4b11e62b-a4" Port int-br-prv Interface int-br-prv Port br-int Interface br-int type: internal Bridge br-prv Port phy-br-prv Interface phy-br-prv Port br-prv Interface br-prv type: internal Port "eth2" Interface "eth2" ovs_version: "1.10.0"

Any help would be appreciated.

edit retag flag offensive reopen merge delete

Closed for the following reason the question is answered, right answer was accepted by skyrainman
close date 2013-07-15 11:22:34.995812


Perhaps you need to configure your physical switch to carry tagged packets. Try connecting you eth2s directly with a cable as a test.

darragh-oreilly gravatar imagedarragh-oreilly ( 2013-07-14 11:51:41 -0600 )edit

I tried connected them together and I am able to get a DHCP address. I'm running Cisco IOS and my port settings are: switchport trunk encapsulation dot1q, switchport mode trunk. I see it tagged the port on the bridge with Vlan 1 and on my switch vlan 1 is shutdown and is native, your thoughts?

skyrainman gravatar imageskyrainman ( 2013-07-14 13:31:14 -0600 )edit

Do 'quantum net-show net-id' and look for provider:segmentation_id - that should be the global vlan for this quantum net. The vlans on br-int are only local, and are translated: 'ovs-ofctl dump-flows br-prv'. How have you configured network_vlan_ranges and bridge_mappings in ovs_quantum_plugin.ini?

darragh-oreilly gravatar imagedarragh-oreilly ( 2013-07-14 17:43:14 -0600 )edit

Resolved! That was the problem, I granted access to VLAN 1 instead of 900. After reading, I understand that VLAN translation is being used. Thank you much!

skyrainman gravatar imageskyrainman ( 2013-07-14 18:26:01 -0600 )edit

@darragh-oreilly, thanks again! One thing to note was that once the VLAN issue was fixed, I got the error: Stderr: '\ndnsmasq: failed to create listening socket for Address already in use\n' - to resolve, I stopped dnsmasq and restarted quantum-dhcp-agent, that fixed it.

skyrainman gravatar imageskyrainman ( 2013-07-14 18:36:18 -0600 )edit

1 answer

Sort by ยป oldest newest most voted

answered 2013-07-15 07:41:37 -0600

darragh-oreilly gravatar image

The problem was that the physical switch was dropping packets that were tagged with 802.1Q headers by OVS on the hosts. The physical switch ports in question were properly configured as type trunk, but their allowed vlans did not include the range specified by network_vlan_ranges in ovs_quantum_plugin.ini.

edit flag offensive delete link more

Get to know Ask OpenStack

Resources for moderators

Question Tools

1 follower


Asked: 2013-07-14 03:48:02 -0600

Seen: 1,231 times

Last updated: Jul 15 '13