0

So I recently tried to deploy a couple of instances and connect them to the external LAN that is also used by the compute nodes to talk to each other and the controller. I've opted for the flat provider networks as it seemed to be the easiest to implement, however I failed to do so and I'm struggling to understand what I'm missing or what I'm doing wrong.

I have followed the official doc and it did not allow my instances to connect to the external LAN that is connected to the provider network.

I found this doc by RedHat that shows how to configure your network to allow instances to connect through, but this doesn't work for me, as my management interface that I require to access the compute nodes via ssh is the same as I want to route the instances traffic on. The way the interfaces are configured in this doc would mean my interface no longer has an IP address and that would mean I would no longer be able to access the compute nodes. So this can't be what I need either.

What is the correct way to do this? What am I missing here?

McWaffel
  • 11
  • 1
  • 2
  • You say "It did not allow the instances to connect". What happens when you try? Is there an error message? Also add the configuration of that provider network and the command (if you use the CLI) that launches the instances. – berndbausch Feb 05 '21 at 14:47
  • There's no network connection between physical machines residing in the physical network and virtual machines hosted by the OpenStack compute nodes. Ping, ssh, scp etc.. all don't work. It's like they're physically disconnected. – McWaffel Feb 08 '21 at 08:36
  • It is your responsibility as the cloud operator to make that connection. This is done with configuration settings and commands. This is why I asked for your configuration. Add the commands by which you set up the network, and launched the instances. Woth thst information, I have a chance to help. – berndbausch Feb 08 '21 at 10:30

0 Answers0