PDA

View Full Version : Unable to ping to virtual router from external network



kenzietandun
14-Jun-2016, 09:29
Hello,

I'm trying out SUSE Openstack Cloud 5. Everything has been working fine and I've managed to launch a cirros instance.

The problem is, I can't ping the virtual network (floating ip) from my local network.

Everything is done in VirtualBox. This is my configuration:

192.168.124.1 ------ Host machine --- LAN
|
192.168.124.10 ---- Admin Node
|
192.168.124.81 ---- Compute Node
|
192.168.124.82 ---- Control Node --- 192.168.126.2
| -------------------VLAN300
192.168.126.1 ------ External Router
|
LAN

Compute node as Nova server. Control node runs all other services (Database, RabbitMQ, Keystone, Glance, Cinder, Neutron, Horizon).

Neutron is setup to use Linuxbridge.

Pinging from External Router to Control Node works (192.168.126.1 --> 192.168.126.2) which makes me think that the VLAN is working properly.

Pinging from virtual router to Control Node works (192.168.126.129 -> 192.168.126.2).

Pinging from my LAN to Control Node works (192.168.1.0/24 -> 192.168.126.2).

I tried creating another router with a floating ip assigned, pinging between 2 floating ips works.

Pinging from External Router to Virtual Router however, doesn't work (192.168.126.1 ----X---> 192.168.126.129).

Here's my /etc/crowbar/network.json http://pastebin.com/ixxF2rVZ

If more information is needed please let me know.

Please help!

Thanks

Kenzie

smflood
14-Jun-2016, 10:18
On 14/06/16 09:34, kenzietandun wrote:

>
> I'm trying out SUSE Openstack Cloud 5. Everything has been working fine
> and I've managed to launch a cirros instance.

Is there any particular reason why you're using SUSE OpenStack Cloud 5
(based on SUSE Linux Enterprise Server 11 SP3) and not OpenStack Cloud 6
(SLES 12 SP1)?

HTH.
--
Simon
SUSE Knowledge Partner

------------------------------------------------------------------------
If you find this post helpful and are logged into the web interface,
please show your appreciation and click on the star below. Thanks.
------------------------------------------------------------------------

kenzietandun
14-Jun-2016, 11:27
On 14/06/16 09:34, kenzietandun wrote:

>
> I'm trying out SUSE Openstack Cloud 5. Everything has been working fine
> and I've managed to launch a cirros instance.

Is there any particular reason why you're using SUSE OpenStack Cloud 5
(based on SUSE Linux Enterprise Server 11 SP3) and not OpenStack Cloud 6
(SLES 12 SP1)?

HTH.
--
Simon
SUSE Knowledge Partner

------------------------------------------------------------------------
If you find this post helpful and are logged into the web interface,
please show your appreciation and click on the star below. Thanks.
------------------------------------------------------------------------

Hi smflood, thanks for the quick reply.

I searched for an iso that has everything included to deploy SUSE Openstack and found this https://susestudio.com/a/Mrr6vv/suse-openstack-cloud-5-admin

I tried Openstack Cloud 6 couple of times before this and had the same problem, so I think it's probably a configuration mistake from my end.

My configuration from Cloud 6 is the same as the one in Cloud 5, except I used openvswitch instead of LinuxBridge.

I'm reinstalling my Openstack Cloud 6 now, I'll reply again in ~45 minutes when it's done.

Thanks

Kenzie

kenzietandun
14-Jun-2016, 12:08
Alright, Openstack Cloud 6 is up and running.

The problem still persists, Public's admin page (192.168.126.2) is pingable, Virtual router (192.168.126.129) is not. Ugh.

http://i.imgur.com/I3wmnFS.png

eblock
16-Jun-2016, 15:31
Hi,
it's been a while since I worked last with SUSE Cloud, I can't check pings etc at the moment. But the docs say

All SUSE OpenStack Cloud nodes need to be physical machines. Although the Administration Server and the Control Node can be virtualized in test environments, this is not supported for production systems.
If you have more details about your network setup, maybe someone with more experience could help.
I also spotted a minor difference between your network.json and the defaults, but I doubt that it's responsible for the described issue.

Regards,
Eugen