Container unable to communicate with host on CircleCI

[fgiorgetti]

First of all, I believe this is not a bug, but some environment/session issue when running Podman
on CircleCI (as I am unable to reproduce it anywhere else).

So I am opening this discussion just to get some further assistance/insights on what else I could
look for, in order to figure out the root cause for this network issue.

Here is what I am doing. First, I create a podman network (rootless), connect a container to it.
At this point the container is able to communicate with the host machine.

Later I create a second container (connected to the same podman network) and both containers
are able to communicate with the host machine, as expected.

Then I remove this second container and I notice that the first container is no longer able to reach
the host machine.

When I restart the old container, I see an error message, but the connectivity is restored:

ERRO[0010] failed to move the rootless netns slirp4netns process to the systemd user.slice: dbus: invalid bus address (no transport)

I have a circleci pipeline that reproduces the issue and it keeps systemd journal and some extra information,
but I could not find any relevant clue, other than the error reported above.

Here are the links to the repo containing the .circleci/config.yml [1] and to the pipeline execution [2].

[1] - https://github.com/fgiorgetti/circleci-sandbox/blob/main/.circleci/config.yml
[2] - https://app.circleci.com/pipelines/github/fgiorgetti/circleci-sandbox/3/workflows/f6447b70-4d2e-47dc-97ed-697198df30c0/jobs/3

Thanks in advance.