Enabling autologin and setting a custom hostname

Make sure that you have completed the steps described in the initial setup page before starting this tutorial.

Provisioning Fedora CoreOS

Fedora CoreOS does not have a separate install disk. Instead, every instance starts from a generic disk image which is customized on first boot via Ignition.

Ignition config files are written in JSON but are typically not user friendly. Configurations are thus written in a simpler format, the Butane config, that is then converted into an Ignition config. The tool responsible for converting Butane configs into Ignition configs is naturally called Butane.

First Ignition config via Butane

Let’s create a very simple Butane config that will perform the following actions:

  • Add a systemd dropin to override the default serial-getty@ttyS0.service.

  • The override will make the service automatically log the core user in to the serial console of the booted machine.

  • Set the system hostname by dropping a file at /etc/hostname,

  • Add a bash profile that tells systemd to not use a pager for output.

  • Raise kernel console logging level to hide audit messages from the console.

We can create a config file named autologin.bu now as:

variant: fcos
version: 1.4.0
systemd:
  units:
    - name: serial-getty@ttyS0.service
      dropins:
      - name: autologin-core.conf
        contents: |
          [Service]
          # Override Execstart in main unit
          ExecStart=
          # Add new Execstart with `-` prefix to ignore failure`
          ExecStart=-/usr/sbin/agetty --autologin core --noclear %I $TERM
storage:
  files:
    - path: /etc/hostname
      mode: 0644
      contents:
        inline: |
          tutorial
    - path: /etc/profile.d/systemd-pager.sh
      mode: 0644
      contents:
        inline: |
          # Tell systemd to not use a pager when printing information
          export SYSTEMD_PAGER=cat
    - path: /etc/sysctl.d/20-silence-audit.conf
      mode: 0644
      contents:
        inline: |
          # Raise console message logging level from DEBUG (7) to WARNING (4)
          # to hide audit messages from the interactive console
          kernel.printk=4

This configuration can then be converted into an Ignition config with Butane:

$ butane --pretty --strict autologin.bu --output autologin.ign

The resulting Ignition configuration produced by Butane as autologin.ign has the following content:

{
  "ignition": {
    "version": "3.1.0"
  },
  "storage": {
    "files": [
      {
        "path": "/etc/hostname",
        "contents": {
          "source": "data:,tutorial%0A"
        },
        "mode": 420
      },
      {
        "path": "/etc/profile.d/systemd-pager.sh",
        "contents": {
          "source": "data:,%23%20Tell%20systemd%20to%20not%20use%20a%20pager%20when%20printing%20information%0Aexport%20SYSTEMD_PAGER%3Dcat%0A"
        },
        "mode": 420
      },
      {
        "path": "/etc/sysctl.d/20-silence-audit.conf",
        "contents": {
          "source": "data:,%23%20Raise%20console%20message%20logging%20level%20from%20DEBUG%20(7)%20to%20WARNING%20(4)%0A%23%20to%20hide%20audit%20messages%20from%20the%20interactive%20console%0Akernel.printk%3D4%0A"
        },
        "mode": 420
      }
    ]
  },
  "systemd": {
    "units": [
      {
        "dropins": [
          {
            "contents": "[Service]\n# Override Execstart in main unit\nExecStart=\n# Add new Execstart with `-` prefix to ignore failure`\nExecStart=-/usr/sbin/agetty --autologin core --noclear %I $TERM\n",
            "name": "autologin-core.conf"
          }
        ],
        "name": "serial-getty@ttyS0.service"
      }
    ]
  }
}

Butane outputs valid Ignition configs. However, if you are tweaking the config after Butane, or manually creating Ignition configs, you will have to verify that the config format is valid with ignition-validate:

ignition-validate autologin.ign && echo 'Success!'

Booting Fedora CoreOS

Now that we have an Ignition config, we can boot a virtual machine with it. This tutorial uses the QEMU image with libvirt, but you should be able to use the same Ignition config on all the platforms supported by Fedora CoreOS.

We use virt-install to create a new Fedora CoreOS virtual machine with a specific config:

# Setup the correct SELinux label to allow access to the config
chcon --verbose --type svirt_home_t autologin.ign

# Start a Fedora CoreOS virtual machine
virt-install --name=fcos --vcpus=2 --ram=2048 --os-variant=fedora-coreos-stable \
    --import --network=bridge=virbr0 --graphics=none \
    --qemu-commandline="-fw_cfg name=opt/com.coreos/config,file=${PWD}/autologin.ign" \
    --disk=size=20,backing_store=${PWD}/fedora-coreos.qcow2

The virt-install command will start an instance named fcos from the fedora-coreos.qcow2 image using the autologin.ign Ignition config. It will auto-attach the serial console of the machine so you will be able to see the image bootup messages.

We use the backing_store option to virt-install --disk to quickly create a new disk image and avoid writing to the original image we have downloaded. This new disk image can be easily thrown away.

Depending on your version of virt-install, you may not be able to use --os-variant=fedora-coreos-stable and will get an error. In this case, you should pick an older Fedora variant (--os-variant=fedora31 for example). You can find the variants that are supported by you current version of virt-install with osinfo-query os | grep '^\s*fedora'.

Once the machine is booted up you should see a few prompts and then you should be automatically logged in and presented with a bash shell:

[  OK  ] Started rpm-ostree System Management Daemon.

Fedora CoreOS 32.20200715.3.0
Kernel 5.7.8-200.fc32.x86_64 on an x86_64 (ttyS0)

SSH host key: SHA256:XlbayjbgDKNoAAHQxsEL5Q7BdwLxxWSw4NXN9SALLmo (ED25519)
SSH host key: SHA256:3sx5jseteO4BvdOMWIi0J4koQL015mLonnD0UPTtnZk (ECDSA)
SSH host key: SHA256:K0fn5/TMJOoMs7Fu7RRkE7IBEf2t8OYCfVaVc+GJWGs (RSA)
ens2: 192.168.122.127 fe80::5054:ff:feb9:3d97
Ignition: user provided config was applied
No ssh authorized keys provided by Ignition or Afterburn
tutorial login: core (automatic login)

[core@tutorial ~]$

Let’s verify that our configuration has been correctly applied. As we were automatically logged in to the terminal, we can safely assume that the systemd dropin has been created:

[core@tutorial ~]$ systemctl cat serial-getty@ttyS0.service
# /usr/lib/systemd/system/serial-getty@.service
...

# /etc/systemd/system/serial-getty@ttyS0.service.d/autologin-core.conf
[Service]
# Override Execstart in main unit
ExecStart=
# Add new Execstart with `-` prefix to ignore failure`
ExecStart=-/usr/sbin/agetty --autologin core --noclear %I $TERM

We can also check that the hostname has correctly been set:

[core@tutorial ~]$ cat /etc/hostname
tutorial
[core@tutorial ~]$ hostnamectl
   Static hostname: tutorial
         Icon name: computer-vm
           Chassis: vm
        Machine ID: d06466128a1c4a6ab255d9581748755c
           Boot ID: 2a19abb9681e442cb1e10271350bfff3
    Virtualization: kvm
  Operating System: Fedora CoreOS 32.20200715.3.0
       CPE OS Name: cpe:/o:fedoraproject:fedora:32
            Kernel: Linux 5.7.8-200.fc32.x86_64
      Architecture: x86-64

Exploring Fedora CoreOS internals

Once we have access to the console of the machine we can browse around a bit to see some of the different pieces of the operating system. For example, even though this is an OSTree based system it was still composed via RPMs and we can inspect the system to see what it was composed of:

[core@tutorial ~]$ rpm -q ignition kernel moby-engine podman systemd rpm-ostree zincati
ignition-2.4.1-1.git5260a5b.fc32.x86_64
kernel-5.7.8-200.fc32.x86_64
moby-engine-19.03.11-1.ce.git42e35e6.fc32.x86_64
podman-1.9.3-1.fc32.x86_64
systemd-245.6-2.fc32.x86_64
rpm-ostree-2020.3-1.fc32.x86_64
zincati-0.0.12-2.fc32.x86_64

We can also inspect the current revision of Fedora CoreOS:

[core@tutorial ~]$ rpm-ostree status
State: idle
Deployments:
* ostree://fedora:fedora/x86_64/coreos/stable
                   Version: 32.20200715.3.0 (2020-07-27T11:36:29Z)
                    Commit: a3b08ee51b1d950afd9d0d73f32d5424ad52c7703a6b5830e0dc11c3a682d869
              GPGSignature: Valid signature by 97A1AE57C3A2372CCA3A4ABA6C13026D12C944D0

And check on zincati.service, which communicates with our update server and tells rpm-ostree when to do an update and to what version to update to:

[core@tutorial ~]$ systemctl status --full zincati.service
● zincati.service - Zincati Update Agent
     Loaded: loaded (/usr/lib/systemd/system/zincati.service; enabled; vendor preset: enabled)
     Active: active (running) since Thu 2020-08-06 14:50:36 UTC; 1h 41min ago
       Docs: https://github.com/coreos/zincati
   Main PID: 889 (zincati)
      Tasks: 2 (limit: 2288)
     Memory: 14.2M
     CGroup: /system.slice/zincati.service
             └─889 /usr/libexec/zincati agent -v

Aug 06 14:50:36 tutorial systemd[1]: Started Zincati Update Agent.
Aug 06 14:50:36 tutorial zincati[889]: [INFO ] starting update agent (zincati 0.0.12)
Aug 06 14:50:39 tutorial zincati[889]: [INFO ] Cincinnati service: https://updates.coreos.fedoraproject.org
Aug 06 14:50:39 tutorial zincati[889]: [INFO ] agent running on node 'dbe8968f75c34d9eb3d3c4c226aa8fdf', in update group 'default'
Aug 06 14:50:39 tutorial zincati[889]: [INFO ] initialization complete, auto-updates logic enabled

One other interesting thing to do is view the logs from Ignition in case there is anything interesting there we may want to investigate:

[core@tutorial ~]$ journalctl -t ignition
...

And finally, of course we can use the podman (or docker) command to inspect the current state of containers on the system:

[core@tutorial ~]$ podman version
[core@tutorial ~]$ podman info
podman commands can be run as root or as non-root user. docker commands need to be run as root via sudo unless the user has been added to the docker group.
Running containers via docker and podman at the same time can cause issues and result in unexpected behaviour. Refer to the FAQ Entry for more details.
The Docker daemon is not started by default but running any docker command will start it as it is socket activated via systemd.

Taking down the Virtual Machine

Let’s now get rid of that virtual machine so we can start again from scratch. First escape out of the serial console by pressing CTRL + ] and then type:

virsh destroy fcos
virsh undefine --remove-all-storage fcos

You may now proceed with the second tutorial.