Networking

In Podman, you will see differences in networking between rootless and rootfull containers. This guide will help to understand how networking can be used in Podman.

Podman pods

By definition, all containers in the same Podman pod share the same network namespace. Therefore, the containers will share the IP Address, MAC Addresses and port mappings. You can always communicate between containers in the same pod, using localhost.

Rootless Networking

When using Podman as a rootless user, the network is setup automatically. The container itself does not have an IP Address, because without root privileges, network association is not allowed. You will also see some other limitations.

Publishing Ports

Port publishing as rootless containers can be done for “high ports” only. All ports below 1024 are privileged and cannot be used for publishing.

Instead of:

$ podman run -dt -p 80:8080/tcp registry.fedoraproject.org/f29/httpd

you want to use:

$ podman run -dt -p 8080:8080/tcp registry.fedoraproject.org/f29/httpd

Note: You can also use podman -P to automatically publish and map ports.

Container <-> Host Communication

If you want to reach a rootless container from your localhost, you can use port publishing (as in the example above).

You can check the ports published and occupied:

$ podman port -l
8080/tcp -> 0.0.0.0:8080

Note: The -l is a convenience argument for latest container. You can also use the container’s ID or name instead of -l or the long argument --latest.

Container <-> Container Communication

Communicating between two rootless containers can be achieved in multiple ways. The easiest and most convenient way is to communicate via published ports and the underlying host.

Check, if a “listening” container is running:

$ podman ps

Check the published ports:

$ podman port <container_id>

Check the address of your host:

$ ip addr

Start a new container to contact your host + the published port:

$ podman run -it --rm fedora curl <Host_IP_Address>:<Published_Port>

Rootfull Networking

This section describes how networking can be used in rootfull containers.

Publishing Ports

Port publishing works the same way as rootless containers, but you will be able to use privileged ports, as long as they are free.

$ sudo podman run -dt -p 80:8080/tcp registry.fedoraproject.org/f29/httpd

Note: You can also use podman -P to automatically publish and map ports.

Container <-> Host Communication

Rootfull containers are reachable via their published ports.

You can check which ports are published:

$ sudo podman port -l
8080/tcp -> 0.0.0.0:80

And you should be able to reach the website from your local machine:

$ curl localhost

Container <-> Container Communication

Rootfull containers can communicate via their IP Address in the same network.

$ sudo podman inspect <container_id> | grep IPAddress
            "IPAddress": "10.88.0.83",

$ sudo podman run -it --rm fedora curl <Container_IP_Address>:<Container_Port>

Configuring Networking

The installation of Podman provides a default network configuration commonly installed in /etc/cni/net.d/ as 87-podman-bridge.conflist. The default network name is defined in /usr/share/containers/libpod.conf. If you want to change the default network, you should copy the libpod.conf to /etc/containers/libpod.conf and change the new file.

To create a new network, you can use the podman network create command, which will create a new file in /etc/cni/net.d/.

Using DNS in Container Networks

Podman provides a convenient way to allocate local DNS records to containers via the dnsname plugin. This can become handy, if you want to communicate between 2 or more containers.

The feature will be automatically enabled for newly created networks via podman network create. If you want to add this feature to the default network, you can either create a new network and make it default or add the needed lines to cat /etc/cni/net.d/87-podman-bridge.conflist - a reboot may be required.

{
  ...

  "plugins": [

    ...

    {
      "type": "dnsname",
      "domainName": "example.com"
    }
  ]
}