Ask Your Question
2

Havana - Dual Node using using Neutron GRE

asked 2014-04-18 13:22:16 -0600

Nader gravatar image

updated 2014-04-18 13:26:48 -0600

Hi there,

I'm working on implementing POC based on Havana bits. Main goal for now is to get GRE to work :) Please find below details of the installation environment and issues we are facing. Any advise and support will be highly appreciated.

Thanks,

INSTALLATION ENVIRONMENT: - Two physical machines running CentOS 6.5 based install of Linux (2.6.32-431.el6.x86_64) - Each Server has 2 network interfaces each configured with static IP addresses. For this installation eth2 will be "management traffic" and eth3 will be "Data traffic". The following IP addresses are used: -- control + Neutron (combined) host eth2 172.16.1.71, eth3 172.16.10.71 -- compute host eth2 172.16.1.72, eth3 172.16.1.72 - Iptables is disabled. Also we have set selinux to permissive.

Basically we followed the steps on the wiki page to get things deployed and OVS configured. http://wiki.stackinsider.org/index.ph...

From Horizon interface, we configured Tenant Network and Started a VM. But unfortunately it is failing to get an IP from DHCP.

Checked the networking and OVS configuration, it looked good to me. Please find below the details

CONTROLLER & NEUTRON NODE (Combined)

ifconfig

br-ex Link encap:Ethernet HWaddr E8:39:35:C5:55:EC inet addr:172.16.10.71 Bcast:172.16.10.255 Mask:255.255.255.0 inet6 addr: fe80::ea39:35ff:fec5:55ec/64 Scope:Link UP BROADCAST RUNNING MTU:1500 Metric:1 RX packets:72867 errors:0 dropped:0 overruns:0 frame:0 TX packets:14 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:0 RX bytes:9041998 (8.6 MiB) TX bytes:944 (944.0 b)

br-int Link encap:Ethernet HWaddr 4A:74:7A:97:C6:4C inet6 addr: fe80::ac84:79ff:fe42:16b7/64 Scope:Link UP BROADCAST RUNNING MTU:1500 Metric:1 RX packets:243 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:12874 (12.5 KiB) TX bytes:468 (468.0 b)

br-tun Link encap:Ethernet HWaddr EE:BF:98:D4:00:4F inet6 addr: fe80::9087:70ff:fe14:c7b1/64 Scope:Link UP BROADCAST RUNNING MTU:1500 Metric:1 RX packets:0 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:0 (0.0 b) TX bytes:468 (468.0 b)

eth2 Link encap:Ethernet HWaddr E8:39:35:C5:55:E8 inet addr:172.16.1.71 Bcast:172.16.1.255 Mask:255.255.255.0 inet6 addr: fe80::ea39:35ff:fec5:55e8/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:477700 errors:0 dropped:0 overruns:0 frame:0 TX packets:331140 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:199887899 (190.6 MiB) TX bytes:68040243 (64.8 MiB)

eth3 Link encap:Ethernet HWaddr E8:39:35 ... (more)

edit retag flag offensive close merge delete
add a comment

1 answer

Sort by ยป oldest newest most voted
1

answered 2014-04-18 15:23:02 -0600

dbaxps gravatar image

updated 2014-04-19 01:46:54 -0600

Core idea of setup you are interested in had been provided by http://kashyapc.com/2013/11/
However, there is some distance from POC to production running,if you
would like you may go through stuff described in my answer. It would
give you a snapshot of 3 months development in step by step details.
I am very sorry, but having ready for production , stable running
"Two Nodes F20 Clusters Neutron&GRE with only one Ehernet adapter for each node"
go back to CentOS 6.5 doesn't seem to me as a good idea. I sincerely apologize
for first redaction of my feedback

 I can provide a link based on updated  setup for Two  Fedora 20 VMs running on non-default libvirt subnet   suggested by  Kashyap Chamarthy in November 2013. Setup down here is for Two  Node Fedora 20 physical boxes with  native Havana packages installed.  I have in meantime several  Two Node Clusters been built as descibed bellow running on daily basis in QA environment for about 2.5  months. So, I am quite sure that provided HowTo is correct. Have to mention that there is no any answer file, every setup has been done manually following ,in general, original Kashyap's instructions, however setup supposed to run in meantime has several important changes ignoring those would cause all procedure to crash due to changes in Fedora 20 happened since 11/2013. All F20 updates have happened sinse 01/23 where applied to each one of working clusters.

"Setting up Two Physical-Node OpenStack RDO Havana + Neutron GRE + Gluster backend for Cinder" on Fedora 20 boxes with both Controller and Compute nodes each one having one Ethernet adapter .

edit flag offensive delete link more

Comments

Feel free to view any changes been done to configuration since 01/23 up to the most recent Link to Lxer's newswire.
So, implementing POC based on Havana bits had been, actually, done by Kashyap in November 2013.

dbaxps gravatar imagedbaxps ( 2014-04-18 16:05:00 -0600 )edit
add a comment

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Get to know Ask OpenStack

Resources for moderators

Question Tools

1 follower

subscribe to rss feed

Stats

Asked: 2014-04-18 13:22:16 -0600

Seen: 242 times

Last updated: Apr 19 '14

Related questions

Asymetric ARP & DHCP issues on GRE tenant net...

Disabling security groups in Horizon

Instances not reachable from outside

OVS Issue ! - Challenge

DNSMASQ-DHCP not releasing IPs after VM deletion

specify OVS path

Instance failing to spawn, logs show `NovaException: Unexpected vif_type=binding_failed`

Link is down in openvswitch port

packstack ovs

Should the br-int in up status?