FreeIPA server container

Building FreeIPA server image

This repository contains Dockerfiles and associated assets for building FreeIPA server container images from the official yum/dnf repositories.

There are multiple Dockerfiles available for images based on various operating systems. Use -f option to podman build or docker build to pick a specific operating system. For example, running

podman build -t freeipa-server -f Dockerfile.centos-8-stream .

will build image based on CentOS 8 Stream packages using podman, and with

docker build -t freeipa-server -f Dockerfile.fedora-rawhide .

FreeIPA image based on Fedora rawhide will be built with docker. Note that when using docker / moby-engine, the docker daemon needs to be running.

Running FreeIPA server container

The FreeIPA container runs systemd to manage all the necessary services within a single container. Running a systemd-based container may require special handling or parameters to be passed to the container runtime.

With podman, normal podman run is typically enough.

With docker on systems with cgroups v2, it may be necessary to use user namespace remapping for the container cgroup to be properly created and mounted within the container read-write as systemd expects it, with

{ "userns-remap": "default" }

in /etc/docker/daemon.json. Restart of the docker service is needed after this edit. This approach also isolates the root in the container from the root on the host, which is a good thing in general.

With docker on systems with cgroups v1, it may be necessary to invoke docker run with option -v /sys/fs/cgroup:/sys/fs/cgroup:ro.

On SELinux enabled systems, it may be also necessary to enable running systemd in containers by setting SELinux boolean container_manage_cgroup on the host with

setsebool -P container_manage_cgroup 1

The FreeIPA container will store all its configurations and data on volume mounted to /data directory in the container. If we create directory which will hold the server data on the host with

mkdir /var/lib/ipa-data

we can then create the FreeIPA container with podman using

podman run --name freeipa-server-container -ti \
    -h ipa.example.test --read-only \
    -v /var/lib/ipa-data:/data:Z freeipa-server [ opts ]

and with docker using

docker run --name freeipa-server-container -ti \
    -h ipa.example.test --read-only \
    -v /var/lib/ipa-data:/data:Z freeipa-server [ opts ]

In case cgroup v1 is used on the host, -v /sys/fs/cgroup:/sys/fs/cgroup:ro option may be necessary in the docker run case.

If you receive error like

IPv6 stack is enabled in the kernel but there is no interface that
has ::1 address assigned. Add ::1 address resolution to 'lo' interface.
You might need to enable IPv6 on the interface 'lo' in sysctl.conf.

you might also need to add option

    --sysctl net.ipv6.conf.all.disable_ipv6=0

When running DNS server (the --setup-dns argument to ipa-server-install) in a container with read-only root filesystem (the --read-only option to podman run or docker run), the setup code in the container won't be able to edit /etc/resolv.conf in the container to point it to itself. Add --dns=127.0.0.1 option to the podman run or docker run invocation to allow the FreeIPA server to reach its own DNS server.

To allow for unprivileged container operation, use the -h ... option to set the hostname for the FreeIPA server in the container. If it's not possible to set the hostname for the container, specify it with IPA_SERVER_HOSTNAME environment variable, for example with podman run -e IPA_SERVER_HOSTNAME=.... This might however not work with read-only containers. Do not use the ipa-server-install --hostname ... argument.

Upon the first invocation with empty directory mounted to /data, the container will run either ipa-server-install or ipa-replica-install to configure FreeIPA master or replica. For example

podman run -ti -h ipa.example.test --read-only \
    -v /var/lib/ipa-data:/data:Z \
    freeipa-server ipa-server-install -r EXAMPLE.TEST --no-ntp

will run interactive ipa-server-install and configure the FreeIPA master using the inputs provided. For unattended initial installation, use the -U argument to ipa-server-install and specify all the necessary inputs as argument on the command line, for example

docker run -ti -h ipa.example.test --read-only \
    -v /var/lib/ipa-data:/data:Z \
    -e PASSWORD=Secret123 \
    freeipa-server ipa-server-install -U -r EXAMPLE.TEST --no-ntp

The ipa-server-install command is the default and can be omitted.

Sometimes it is not convenient or possible to specify the arguments to ipa-server-install as arguments to podman run or docker run. In the case they can be specified either using environment variable IPA_SERVER_INSTALL_OPTS using the -e option, or they can be passed in using file ipa-server-install-options in the directory mounted to the container as /data. For example, when /var/lib/ipa-data/ipa-server-install-options contains

--realm=EXAMPLE.TEST
--ds-password=The-directory-server-password
--admin-password=The-admin-password

and podman run or docker run is executed with -v /var/lib/ipa-data:/data:Z, the content of ipa-server-install-options will be passed as arguments to ipa-server-install.

Since the ipa-server-install-options typically contains passwords, it is also possible to use podman secret create to store the whole content of that file, and the invoke podman run with options like

--secret source=options-with-credentials,target=/data/ipa-server-install-options

to expose the options in the container. The same holds for docker invocation.

If you want to instruct the container to create a replica, specify the ipa-replica-install command in the podman run or docker run parameters:

podman run -ti -h ipa.example.test --read-only \
   -v /var/lib/ipa-data:/data:Z \
   freeipa-server ipa-replica-install [ opts ]

Using ipa-replica-install-options also works and will invoke ipa-replica-install and pass it its content as argument, the same way ipa-server-install-options works for ipa-server-install.

Upon subsequent invocations when /data is found already populated with FreeIPA server configuration and data, the options are ignored and just the necessary services started in the container.

If you want to use the FreeIPA server not just from the host where it is running but from external machines as well, you might want to use the -p options to make the services accessible externally.

docker run -p 53:53/udp -p 53:53 \
    -p 80:80 -p 443:443 -p 389:389 -p 636:636 -p 88:88 -p 464:464 \
-p 88:88/udp -p 464:464/udp -p 123:123/udp ...

You will then likely want to also specify the --ip-address option to ipa-server-install with the IP address of the host, and also use the --add-host option to the docker run / podman run with the same IP address, especially when running the container as read only.

Alternatively, the IPA_SERVER_IP environment variable via the -e option to docker run / podman run can be used to define what IP address should the FreeIPA server put to DNS as its address. Using this mechanism will however not update the ipa-ca record.

If you have existing container with data volume, it should be safe to shut it down and run new one based on newer image, with the same data directory bind-mounted to /data. The container logic will detect that it is running with data produced by different image and attempt to upgrade the configuration and data. Of course, keeping backup of the data directory for cases when the upgrade process fails is recommended.

IPA-enrolled client in Docker

There are multiple *-client branches named after OS they are based on. Check out the branch you prefer and in the root of the repository, run:

docker build -t freeipa-client .

To run the client container, run it with correctly set DNS and hostname in the IPA domain, or you can link it to the freeipa-server container directly:

docker run --privileged --link freeipa-server-container:ipa \
    -e PASSWORD=Secret123 -ti freeipa-client

The first time this container runs, it invokes ipa-client-install with the given admin password.

Debugging

The container scripts provide some options for debugging:

• Enable shell script tracing in both the top-level init-data script and the ipa-server-configure-first script by setting the $DEBUG_TRACE environment variable.

• Disable container exit after script failure by setting the $DEBUG_NO_EXIT environment variable. After failure, the container will continue running, and can be entered for debugging with e.g. podman exec -it freeipa-server-container bash. This can also be achieved by specifying no-exit as the first word in the [opts] to the container.

• Force container exit after successfully configuring the FreeIPA server by specifying exit-on-finished as the first word in the [opts] to the container.

Example usage:

podman run [...] -e DEBUG_TRACE=1 -e DEBUG_NO_EXIT=1 freeipa-server

or

docker run [...] freeipa-server exit-on-finished -U -r EXAMPLE.TEST

You can also try to run

tests/run-partial-tests.sh Dockerfile

or

docker=podman tests/run-partial-tests.sh Dockerfile

which can uncover the general issues with running systemd in containers.

CI in GitHub Actions

To check the general health of the project, see https://github.com/freeipa/freeipa-container/actions where tests are run for various OS versions in the containers.

License

Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at

http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.