I took an old server which had two NIC ports and installed Centos 6.5. I then installed openstack all in one server from the following:
http://openstack.redhat.com/Quickstart
I decided to use the havana version because icehouse just came out and I wanted to wait until after the bugs were fixed.
After doing a bunch of yum installs I had it working. I was actually able to create a few VMs using the Cirros Linux image and created a simple network. I was able to ping from one VM to the other.
However the documentation was not clear on how to connect your VMs to the outside world through a physical network.
After searching around for a few days, I found this:
http://openstack.redhat.com/Neutron_with_existing_external_network
I decided to use eth0 as my management ip to access Openstack and use eth1 as the port for the VMs to connect to the physical network.
I also learned that OpenStack uses OpenVswitch and Neutron for network connectivity. You have to learn a little bit about OpenVswitch on a server.
This led me to make the following changes.
under /etc/sysconfig/network-scripts
I duplicated (to backup) and changed the following files.
[root@centos-6-5-openstack network-scripts]# more ifcfg-eth1
DEVICE=eth1
TYPE=OVSPort
UUID=100083c1-6174-4a59-b3d0-09081eef106a
ONBOOT=yes
DEVICETYPE=ovs
HWADDR=00:30:48:F9:B9:C9
OVS_BRIDGE=br-ex
NAME="System eth1"
[root@centos-6-5-openstack network-scripts]# more ifcfg-br-ex
DEVICE=br-ex
DEVICETYPE=ovs
TYPE=OVSBridge
BOOTPROTO=static
IPADDR=192.168.250.11
NETMASK=255.255.255.0
DNS=198.6.1.1
ONBOOT=yes
Then I tied the eth1 interface to the bridge "br-ex"
"br-ex" was automatically created on my server, but if it's not there, you should create a bridge (ova-vsctl add-br <NAME>) and associate the physical interface (eth1) to that bridge.
[root@centos-6-5-openstack network-scripts]# port br-ex eth1
[root@centos-6-5-openstack network-scripts]# ovs-vsctl show
f6d22683-9529-48b6-b9d0-5e49cb720d44
Bridge br-ex
Port br-ex
Interface br-ex
type: internal
Port "qg-d267267b-56"
Interface "qg-d267267b-56"
type: internal
Port phy-br-ex
Interface phy-br-ex
Port "eth1"
Interface "eth1"
Bridge br-int
I also had to make a change to a file:
Add to the /etc/neutron/plugin.ini file these lines:
network_vlan_ranges = physnet1
bridge_mappings = physnet1:br-ex
I had to make sure that eth1 is a port for bridge br-ex not br-int. For some reason it was not doing this, so I removed the port from br-int using ovs-vsctl del-port command
Red hat wants you to restart nework services after making the changes. However when I issued that command, in ifconfig I would see that eth1 still had an ip address and br-ex was not updated. So I rebooted the server.
I also used the following to troubleshoot
tcpdump -nei br-ex
tcpdump -nei eth1
After reboot all was well.
root@centos-6-5-openstack network-scripts]# ifconfig
br-ex Link encap:Ethernet HWaddr 00:30:48:F9:B9:C9
inet addr:192.168.250.11 Bcast:192.168.250.255 Mask:255.255.255.0
inet6 addr: fe80::5822:16ff:fe25:91a/64 Scope:Link
UP BROADCAST RUNNING MTU:1500 Metric:1
RX packets:100 errors:0 dropped:0 overruns:0 frame:0
TX packets:20 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:4834 (4.7 KiB) TX bytes:1336 (1.3 KiB)
br-int Link encap:Ethernet HWaddr BE:9D:E4:A4:B0:44
inet6 addr: fe80::230:48ff:fef9:b9c9/64 Scope:Link
UP BROADCAST RUNNING MTU:1500 Metric:1
RX packets:213 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:16262 (15.8 KiB) TX bytes:468 (468.0 b)
eth0 Link encap:Ethernet HWaddr 00:30:48:F9:B9:C8
inet addr:10.161.32.230 Bcast:10.161.39.255 Mask:255.255.248.0
inet6 addr: fe80::230:48ff:fef9:b9c8/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:14400 errors:0 dropped:0 overruns:0 frame:0
TX packets:11354 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:1686756 (1.6 MiB) TX bytes:7627901 (7.2 MiB)
Interrupt:18 Memory:d8020000-d8040000
eth1 Link encap:Ethernet HWaddr 00:30:48:F9:B9:C9
inet6 addr: fe80::230:48ff:fef9:b9c9/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:905 errors:0 dropped:0 overruns:0 frame:0
TX packets:978 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:82360 (80.4 KiB) TX bytes:87771 (85.7 KiB)
Interrupt:19 Memory:d8060000-d8080000
int-br-ex Link encap:Ethernet HWaddr 5A:EC:7F:16:11:35
inet6 addr: fe80::58ec:7fff:fe16:1135/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:100 errors:0 dropped:0 overruns:0 frame:0
TX packets:190 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:4752 (4.6 KiB) TX bytes:14550 (14.2 KiB)
From my physical gateway switch/router 192.168.250.1 I tried pinging the br-ex interface 192.168.250.11
Next I created the tenant subnet, vm and the openstack router.
First I had to source my admin credentials
source /root/keystonerc_admin
Next I followed the instructions on creating a public network. I added a floating ip.
http://openstack.redhat.com/Floating_IP_range
By default packstack creates a demo public network, with a subnet that I was not using. So I cleared the gw of this first.
# neutron router-gateway-clear router1
Then deleted the subnet
# neutron subnet-delete public
Next I recreated a new subnet
neutron subnet-create --name public --enable_dhcp=False --allocation-pool=start=192.168.250.10,end=192.168.250.20 --gateway=192.168.250.1 public 192.168.250.0/24
Then created a router using Horizon.
Then I tied the router to the public subnet.
neutron router-gateway-set test-rtr public
[root@centos-6-5-openstack network-scripts(keystone_admin)]# neutron subnet-show public
+------------------+------------------------------------------------------+
| Field | Value |
+------------------+------------------------------------------------------+
| allocation_pools | {"start": "192.168.250.10", "end": "192.168.250.20"} |
| cidr | 192.168.250.0/24 |
| dns_nameservers | |
| enable_dhcp | False |
| gateway_ip | 192.168.250.1 |
| host_routes | |
| id | 8942189b-8fbf-44e9-adbb-d62dd0d27015 |
| ip_version | 4 |
| name | public |
| network_id | 1f438ff9-43cf-4eb4-8b92-a385dc1dff8d |
| tenant_id | 3a10de8a82444118865a6398b336ee68 |
+------------------+------------------------------------------------------+
From Horizon, my setup looks simple:
One thing that threw me off during the whole process was trying to figure out if the openstack Gateway IP was working or not.
I noticed that the status kept saying down.
[root@centos-6-5-openstack network-scripts(keystone_admin)]# neutron port-show router-gw | grep status
| status | DOWN
After googling it, I found out that this was a bug and a display issue. So you can't tell if the gateway ip is actually working or not because the status is broken.
From my VM I am now able to access my physical switch
