I have made some progress in that I was able to connect to my cirros image from the public network, but only from the host on which openstack is installed and on which the instance is running.
At the end of Lars's video, mentioned below, he assigns a gateway ip address to the public (192.168.20.0/24) network to the br-ex device, and then adds a rule that I translated into this command:
iptables -t nat -I POSTROUTING 1 -s 192.168.20.0/24 -j MASQUERADE
but this breaks the connectivity, so I removed that so that I could still ssh into the cirros instance from my physical host.
Currently, I'm able to log in from the openstack physical host, but not from the rest of my 192.168.0.0 network.
My networking is a little bit rusty, so I'm not sure what the next step is to allow me to log into the instances on the 192.168.20.0/24 network from existing hosts on the 192.168.0.0 network.
BTW, is there a script that will provide a dump of the configurations like the one for which you provided a URL below?
On 5/21/14, 4:24 AM, Kashyap Chamarthy wrote:
On Tue, May 20, 2014 at 01:17:21PM -0400, Eric Berg wrote:I've done a fresh install of RDO using packstack on a single host like this: packstack --allinone --provision-all-in-one-ovs-bridge=n And then followed the instructions here: http://openstack.redhat.com/Neutron_with_existing_external_network I've also generally followed Lars's approach from this video with the same lack of connectivity: https://www.youtube.com/watch?v=DGf-ny25OAw My public network is 192.168.20.0/24. But I'm not able to ping or ssh from my 1902.168.0.0 network, the host running OpenStack is at 192.168.0.37. My instance is up and running with a 10.0.0.2 IP and 192.168.20.4 floating IP. I can ping 192.168.20.3, but not 192.168.20.4. I can use the net namespace approach to log into my cirros instance, but can't get to 192.168.20.0/24 hosts.That at-sounds you've got most of it right. You're not able to SSH via floating IPs. Couple of things: - You might want to check if your iptables rules are correct. i.e. when you run something like this, you should see SNAT/DNAT rules: $ ip netns exec qrouter-2c7ba7dc-0101-417a-b76d-1cae17ae654e iptables -t nat -L -nv | grep NAT 0 0 DNAT all -- * * 0.0.0.0/0 126.96.36.199 to:188.8.131.52 0 0 DNAT all -- * * 0.0.0.0/0 184.108.40.206 to:220.127.116.11 26 1704 ACCEPT all -- !qg-fb9ff0ad-56 !qg-fb9ff0ad-56 0.0.0.0/0 0.0.0.0/0 ! ctstate DNAT 0 0 DNAT all -- * * 0.0.0.0/0 18.104.22.168 to:22.214.171.124 5 324 DNAT all -- * * 0.0.0.0/0 126.96.36.199 to:188.8.131.52 0 0 SNAT all -- * * 184.108.40.206 0.0.0.0/0 to:220.127.116.11 0 0 SNAT all -- * * 18.104.22.168 0.0.0.0/0 to:22.214.171.124 0 0 SNAT all -- * * 126.96.36.199/24 0.0.0.0/0 to:188.8.131.52 - Ensure you have security group rules for SSH are set correctly (you can enumerate them by doing '$ neutron security-group-rule-list') I recently did a 2-node IceHouse install (but this is manual setup), here are my configurations of Nova/Neutron and iptables rules (scroll down to bottom).This is my first OpenStack install. I'm a little confused at how a stock installation (based on packstack) could somehow not include the ability to access the VMs from the network on which the OS compute host is running. Any help troubleshooting this would be greatly appreciated.
-- Eric Berg Sr. Software Engineer Rubenstein Technology Group 55 Broad Street, 14th Floor New York, NY 10004-2501 (212) 518-6400 (212) 518-6467 fax eberg rubensteintech com www.rubensteintech.com