Cant ping self-service network [closed]

asked 2017-03-31 19:03:05 -0600

rrottach gravatar image

updated 2017-04-07 11:07:30 -0600

I am installing newton on a 2 node setup, a controller and a compute node using neutron networking. I have it up and running and I can launch instances on both the provider and self service networks, I can ssh/ping the provider interface on the router but I cannot ssh/ping the self service interface on the router. I have verified that the bridges, tap interfaces, and the vxlan are all being created and added as networks, and routers are being created. I can ssh into the instances I create on the provider network and I can NAT the provider network to public facing ip's.

The problem is on the self service network, I created a cirros vm (vm1) and logged in via the horizon console and I can ping another vm (vm2) on the self service network but not the interface on the router or any external ip such as 8.8.8.8.

vm1 - 192.168.2.15

vm2 - 192.168.2.10

self service interface on router - 192.168.2.5

I have no errors in my logs that I can find so I have no place to start. I know this isnt a lot of information to start off with so any ideas or suggestions would be greatly appreciated.

UPDATE:

I think i have been looking at neutron networking wrong so I redid the networking as follows

Networks

Provider network - 10.5.0.0/24

Selfservice network - 10.0.0.0/16

Router Interfaces

Provider interface - 10.5.0.12

SelfService interface - 10.0.0.5

Instances

VM1 - 10.0.0.11 - (cirros image)

VM2 - 10.0.0.11 - (ubuntu 16.04 image)

Additional Ports

network:dhcp - 10.0.0.6

I created the cirros vm first and verified that I could ping vm1, vm2, selfservice network router interface, provider network router interface, network:dhcp, and the internet.

I could not connect to the vms through the selfservice network so I added a floating ip to vm1 (10.5.0.11) and I was able to connect. I repeated that with both vms and i can connect to both vms through the provider network.

Since I could not ping the selfservice router interface (10.0.0.5) or the vm's (10.0.0.9,10.0.0.11) from the controller I added a route in my untangle router and now I can ping the selfservice router interface (10.0.0.5) but I cannot ping the vms (10.0.0.9,10.0.0.11).

I think there is a problem between the selfservice router interface and the VM's, is this true or am I attempting to do something that openstack is not designed to do? I have added a screen shot to help explain this

C:\fakepath\Capture 2017-04-04 at 15.47.15.png

edit retag flag offensive reopen merge delete

Closed for the following reason the question is answered, right answer was accepted by rrottach
close date 2017-04-11 17:32:34.643564

Comments

1

Assuming you use a centralized router and Linuxbridge as mechanism driver, trace packets at various points across their path through the system and find out where they disappear.

Bernd Bausch gravatar imageBernd Bausch ( 2017-04-01 02:43:14 -0600 )edit
2

This gives you clues where the error in your configuration might be. The operation guide has a section on NW troubleshooting that you can also use for inspiration.

Bernd Bausch gravatar imageBernd Bausch ( 2017-04-01 02:46:28 -0600 )edit

I am going to close this question based on the troubleshooting links I think things are working as expected

rrottach gravatar imagerrottach ( 2017-04-11 17:30:19 -0600 )edit