Ask Your Question

OpenStack PackStack All-in-one Neutron Network with 2 cables

asked 2014-03-07 10:28:10 -0500

JonF gravatar image

updated 2014-03-13 14:50:03 -0500

I have been trying to set up a single, all-in-one installation but can't get floating IP's working. I tried following the instructions at , along with many other resources. I tried with one ethernet cable and now I'm at two so that I don't lose connectivity whilst restarting the network service - which still breaks when the subnet mask is set to 0 on the bridge.

I've created link:this little presentation to document some of the specifics. I can get a floating IP address assigned but it can't get to the gateway. Can anyone help please? I will create a tut if I can get this figured out.

Here are the config files - Can someone spot anything?



[root@OSAi1-1 ~(keystone_admin)]# neutron router-list
| id                                   | name    | external_gateway_info                                                       |
| 302eb8ac-39ed-471e-beda-1e532bd444e8 | router1 | {"network_id": "e077c4ae-2106-4261-9f7f-a5b6335a1b50", "enable_snat": true} |
[root@OSAi1-1 ~(keystone_admin)]# neutron sub-list
Unknown command [u'sub-list']
[root@OSAi1-1 ~(keystone_admin)]# neutron subnet-list
| id                                   | name          | cidr            | allocation_pools                                 |
| d75a8232-bde8-4600-a264-1c47aaf47224 | public_subnet |    | {"start": "", "end": ""}     |
| dfdb51f5-80fe-4fc9-86d4-31e3d259796a | public_subnet | | {"start": "", "end": ""} |
[root@OSAi1-1 ~(keystone_admin)]# neutron net-list
| id                                   | name    | subnets                                              |
| a74e9505-3532-4520-a816-2a4b2964dc06 | Private | dfdb51f5-80fe-4fc9-86d4-31e3d259796a |
| e077c4ae-2106-4261-9f7f-a5b6335a1b50 | Public  | d75a8232-bde8-4600-a264-1c47aaf47224    |

br-ex     Link encap:Ethernet  HWaddr D0:67:E5:FA:BF:50  
          inet addr:  Bcast:  Mask:
          inet6 addr: fe80::7c72:59ff:fe7a:4b2d/64 Scope:Link
          UP BROADCAST RUNNING  MTU:1500  Metric:1
          RX packets:367774 errors:0 dropped:0 overruns:0 frame:0
          TX packets:7569 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0 
          RX bytes:49915635 (47.6 MiB)  TX bytes:1333273 (1.2 MiB)

br-int    Link encap:Ethernet  HWaddr FA:2A:72:1D:D5:4B  
          inet6 addr: fe80::b86f:56ff:fe83:3208/64 Scope:Link
          UP BROADCAST RUNNING  MTU:1500  Metric:1
          RX packets:22 errors:0 dropped:0 overruns:0 frame:0
          TX packets:6 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0 
          RX bytes:2322 (2.2 KiB)  TX bytes:468 (468.0 b)

em1       Link encap:Ethernet  HWaddr D0:67:E5:FA:BF:4E  
          inet addr:  Bcast:  Mask:
          RX packets:272062 errors:0 dropped:0 overruns:0 frame:0
          TX packets:185 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000 
          RX bytes:22744042 (21.6 MiB)  TX bytes:15636 (15.2 KiB)

em2       Link encap:Ethernet  HWaddr D0:67:E5 ...
edit retag flag offensive close merge delete


this seems to be a commonly asked question; check on this site for some suggestions, maybe edit your question to add more details so others can give specific answers.

smaffulli gravatar imagesmaffulli ( 2014-03-07 14:35:34 -0500 )edit

I think I need someone to I have updated the presentation with some steps that I have followed, but still can't get this working. I am not a networking person so it could be very easy to resolve. Beer anyone? I am willing to contribute to a good cause.

JonF gravatar imageJonF ( 2014-03-12 16:37:27 -0500 )edit

I'm trying to set up an identical all-in-one deployment with two interfaces just like you. In the guide you have a link to, you show both interfaces coming off the server as plugged into the same router. Whats your physical router config like? VLANS?

ten_ten_steve gravatar imageten_ten_steve ( 2014-03-13 13:55:51 -0500 )edit

No VLANs. It's a SonicWall that I don't have rights to. There is a DHCP server on there which I've been told can cause issues, but I can't change the 2 ports I'm using. If I knew what to ask, I may be able to.

JonF gravatar imageJonF ( 2014-03-13 14:24:52 -0500 )edit

Hi. When I reboot the network gets hosed. I can ping the IP address of the "management interface" (em1) from the machine, but not from the network. If I run "service network restart" I can ssh to the management interface IP address but then the VMs can't get to the outside world. Any suggestions?

JonF gravatar imageJonF ( 2014-03-18 10:56:28 -0500 )edit

2 answers

Sort by ยป oldest newest most voted

answered 2014-03-13 14:24:57 -0500

ten_ten_steve gravatar image

updated 2014-03-13 15:05:48 -0500

I noticed that on page 13 of your deployment log/slides, you have your cirros test instance connected directly to your public subnet. This is probably not what you want (and it goes against the typical OpenStack networking model).

When you ran neutron subnet-list, your query returns 2 subnets named "public_subnet" but having different address ranges. I think what happened is, when you created your test instance and connected it to a network, you accidentally connected it directly to your, where you wanted to connect it to the private subnet, then assign a floating IP from your public subnet.

Also, it looks like you have the public and private network reversed in your neutron config. Essentially, you want to have your external subnet on the left (blue line). I think this goes back to fixing your network & subnet naming scheme.

I recommend terminating your current instance, and launching a new one while connecting it to your private subnet (e.g. I would also recommend re-naming your subnet as "private_subnet" this will help differentiate the two.

Some further guidance can be found here Check out the part that says:

Using the neutron cli, I deleted everything that was deletable (nets, subnets, ports, routers, etc) In particular router1 was showing in cli but not in horizon, and I had to delete it using l3-agent-router-remove.

Following this video, I did the following steps:

neutron router-create router1
neutron net-create private
neutron subnet-create private --name private_subnet
neutron router-interface-add router1 private_subnet
neutron net-create public --router:external=True
neutron subnet-create public --name public_subnet --enable_dhcp=False --allocation-pool    start=,end= --gateway=
neutron router-gateway-set router1 public
edit flag offensive delete link more


I'm not sure I understand. Slide 13 has Cirros connected to 172 (now renamed to private_subnet [see slide 14]). The network layout did not change with that update. 172 is the default packstack (internal?) network and I'd like to get a 10. floating address assigned.

JonF gravatar imageJonF ( 2014-03-13 14:44:50 -0500 )edit

Your question is almost identical. So I have reinstall again? I have done that like 5 times. If it means getting it working I will. What does the permissive bit do?

JonF gravatar imageJonF ( 2014-03-13 14:48:21 -0500 )edit

Will do! I have many times, so this shouldn't take long. :) Thank you very much.

JonF gravatar imageJonF ( 2014-03-13 15:15:04 -0500 )edit

how'd it turn out?

ten_ten_steve gravatar imageten_ten_steve ( 2014-03-13 16:03:08 -0500 )edit

answered 2014-06-11 19:57:06 -0500

JoeOpenstack gravatar image

updated 2014-06-11 20:00:36 -0500

I am having the same identical problem. I am a networking guy and the networking is not a problem. In fact, I can connect the Openstack external network to the physical network, but only after boot, not during boot. I just discovered that I am encountering errors in /var/log/messages

Jun 11 17:36:52 joeopenstack4 NetworkManager[965]: ifcfg-rh: parsing /etc/sysconfig/network-scripts/ifcfg-eth0 ... Jun 11 17:36:52 joeopenstack4 NetworkManager[965]: ifcfg-rh: error: Unknown connection type 'OVSPort' Jun 11 17:36:52 joeopenstack4 NetworkManager[965]: ifcfg-rh: parsing /etc/sysconfig/network-scripts/ifcfg-br-ex ... Jun 11 17:36:52 joeopenstack4 NetworkManager[965]: ifcfg-rh: error: Unknown connection type 'OVSBridge'

I don't think that it has anything to do with having two ethernet cables or 2 NICs. I am using RDO Openstack as well.

edit flag offensive delete link more

Get to know Ask OpenStack

Resources for moderators

Question Tools


Asked: 2014-03-07 10:28:10 -0500

Seen: 1,449 times

Last updated: Jun 11 '14