Error when running "linux" containers over Podman

[q2dg]

I've commented this issue in #643 (comment) , but I reference it here for visibility.

[hellt]

Hi
Podman is still in the experimental phase. We are not (actively) supporting it and rely on the community to steer it forward if needed.

[steiler]

@q2dg you might want to take a look at the container I assume it stopped or crashed.
Take a look at the logs of the container.

[q2dg]

Well, they are empty...I suspect because they not even get to be started. I attach a screenshot

Thanks!!

[steiler]

you should enable -d (debugging log) and you see the state of stopped and exited. That explains why the network namespace link is no longer existing.
You would also need to share the topology file.

But I see you're using alpine containers, whcih do not have an entrypoint (or it is no long running process, unsure atm).
But you would need to provide a command that would run as the container process, which would remain running. A stupid easy option is the following:

name: alpinetest
topology:
  nodes:
    node1:
      kind: linux
      image: alpine:latest
      cmd: sleep infinity
    node2:
      kind: linux
      image: alpine:latest
      cmd: sleep infinity
  links:
    - endpoints: ["node1:eth1","node2:eth1"]

[q2dg]

Oooh..it was that!!! The missing entrypoint!! Now it works perfectly with (root-ful) Podman!
Thanks a lot!!! (and sorry for the inconvenience!)