newbie here who's trying to understand Openstack networking topic.
My setup:
I have 3 physical machines with Openstack installed and running:
- controller (with Keystone, Glance, Placement, Neutron, Nova, Heat and Magnum installed)
- compute (with Nova installed)
- block storage (with Cinder installed)
These 3 machines are physically connected (through Ethernet cable) to a router inside the same LAN (with for example addresses 192.168.x.y) , so I don't have a management and a provider network as suggested in Openstack documentation, but I have only a single network.
All installations went good, I am able to create instances and all works correctly.
Now, I have installed Heat and Magnum on my controller to use K8s with Openstack, and I am having some issues during the cluster deployment and I think they're related also to my network configuration.
In fact, I am using the ml2 plugin and the Linux Bridge, but I wasn't able to create a VXLAN or a router, or a Floating IP (it doesn't exist the option to create it even in my Horizon dashboard) because they were disabled in my neutron.conf file (I have followed this guide during my installation, so the configurations for Neutron are pretty much the same (remember I have only one network).
Now, I was reading this from Openstack documentation, and I think that I must have been followed the "Network Option 2: Self-service network" (in the Configure Network options paragraph) for my Neutron installation.
So my questions are:
1) Is the "Network Option 2: Self-service network" the correct network configuration I have to follow in my specific case for my setup?
2) To modify my neutron.conf file with the new options, I have to delete the Neutron DB in Keystone and repopulate it at the end? What are the correct steps to follow?
3) To create my k8s cluster with Magnum do I need to create a VXLAN, with routers to have also external Floating IPs, right?