Open Stack

Hi,

I am in the process of setting up a cluster of 4 CentOS 6.4
servers, with Neutron and VLAN. The current goal is to create
under the same tenant 3 subnets in different VLANs. Also, I've used
the following guides to do this:

http://openstack.redhat.com/Neutron_with_OVS_and_VLANs

http://docs.openstack.org/grizzly/openstack-network/admin/content/under_the_hood_openvswitch.html

I have tried alternative setups:

A. All services (including compute) on one host + 3 additional compute
hosts.
B. All services (except compute) on one host + 3 additional compute hosts.

In case A, the guest VM is by default launched on the same host that
runs the rest of the services -- then the VM is reachable.

In case B, the VM is started on one of the compute hosts -- and then the
guest
VMs are not reachable (though running on the compute nodes and viewable
through the instance console) e.g.:

ip netns exec qrouter-b702c4e9-cb3a-4156-8412-f902363075f2 ping 10.0.103.5

PING 10.0.103.5 (10.0.103.5) 56(84) bytes of data.
>From 10.0.103.1 icmp_seq=2 Destination Host Unreachable
>From 10.0.103.1 icmp_seq=3 Destination Host Unreachable
>From 10.0.103.1 icmp_seq=4 Destination Host Unreachable


I've double checked the security group and rules (to allow ssh and icmp).
Logs don't show anything interesting.

Questions:
* Do the compute hosts need any additional services set up (except the
quantum-openvswitch-agent and openswitch)?

* Could you suggest any troubleshooting tips/techniques?

cheers,
/Nico
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack/attachments/20130903/ee5c42fc/attachment.html>