Linux post-installation steps for Docker Engine¶
These optional post-installation procedures describe how to configure your Linux host machine to work better with Docker.
Manage Docker as a non-root user¶
The Docker daemon binds to a Unix socket, not a TCP port. By default it's the
root user that owns the Unix socket, and other users can only access it using
sudo. The Docker daemon always runs as the root user.
If you don't want to preface the docker command with sudo, create a Unix
group called docker and add users to it. When the Docker daemon starts, it
creates a Unix socket accessible by members of the docker group. On some Linux
distributions, the system automatically creates this group when installing
Docker Engine using a package manager. In that case, there is no need for you to
manually create the group.
Warning
The docker group grants root-level privileges to the user. For
details on how this impacts security in your system, see
Docker Daemon Attack Surface.
The Docker socket is root, wherever it goes
Anyone who can talk to /var/run/docker.sock can run docker run -v /:/host ... and own the machine, with no
password and no sudo log. That applies to:
- users in the
dockergroup (treat membership like passwordless sudo and keep the list short), - containers that mount the socket (autoheal, Dozzle, Homepage's Docker widget, deploy agents, Watchtower-style
updaters). Mounting it
:rodoes not help: read-only only stops the socket file being replaced, the API behind it is still fully writable.
Only give the socket to containers that need it, keep those images pinned and from projects you trust, and for tools that only read (dashboards, log viewers) put a filtering proxy such as docker-socket-proxy in between with only the API sections they need enabled.
Note
To run Docker without root privileges, see Run the Docker daemon as a non-root user (Rootless mode).
To create the docker group and add your user:
- Create the
dockergroup.
- Add your user to the
dockergroup.
- Log out and log back in so that your group membership is re-evaluated.
If you're running Linux in a virtual machine, it may be necessary to restart the virtual machine for changes to take effect.
You can also run the following command to activate the changes to groups:
- Verify that you can run
dockercommands withoutsudo.
This command downloads a test image and runs it in a container. When the container runs, it prints a message and exits.
If you initially ran Docker CLI commands using sudo before adding your user
to the docker group, you may see the following error:
WARNING: Error loading config file: /home/user/.docker/config.json -
stat /home/user/.docker/config.json: permission denied
This error indicates that the permission settings for the ~/.docker/
directory are incorrect, due to having used the sudo command earlier.
To fix this problem, either remove the ~/.docker/ directory (it's recreated
automatically, but any custom settings are lost), or change its ownership and
permissions using the following commands:
Configure Docker to start on boot with systemd¶
Many modern Linux distributions use systemd to manage which services start when the system boots. On Debian and Ubuntu, the Docker service starts on boot by default. To automatically start Docker and containerd on boot for other Linux distributions using systemd, run the following commands:
To stop this behavior, use disable instead.
You can use systemd unit files to configure the Docker service on startup, for example to add an HTTP proxy, set a different directory or partition for the Docker runtime files, or other customizations. For an example, see Configure the daemon to use a proxy.
The daemon.json I use¶
All my servers get the same /etc/docker/daemon.json:
{
"log-driver": "json-file",
"log-opts": {
"max-size": "10m",
"max-file": "3"
},
"dns": ["192.168.1.1"],
"shutdown-timeout": 60
}
| Key | Why |
|---|---|
log-opts max-size / max-file |
Caps each container's log at 3 × 10 MB. Without it, json-file logs grow until the disk fills. docker logs and log viewers like Dozzle keep working. |
dns |
Point containers' DNS at your router/resolver (replace 192.168.1.1). Without it, Docker copies the host's /etc/resolv.conf once, when each container starts: a container that starts during a boot-time DNS gap, or after a VPN/Tailscale client rewrites the host's resolv.conf, keeps the broken copy until it's restarted. |
shutdown-timeout |
On reboot or daemon stop, containers get 60 s instead of 15 s to exit cleanly. Databases (Postgres, MongoDB, SQLite apps) need it to flush instead of being killed mid-write. |
Apply it:
# Validate first: a broken daemon.json stops dockerd from starting at all
sudo dockerd --validate --config-file=/etc/docker/daemon.json
# dns / log-driver changes need a full restart (restarts every container)
sudo systemctl restart docker
# shutdown-timeout (and a few other keys) can be reloaded without restarting containers
sudo kill -SIGHUP "$(pidof dockerd)"
Log options only apply to new containers
Existing containers keep the logging config they were created with. docker compose up -d --force-recreate
(or the next deploy that recreates them) picks up the new defaults.
Lessons learned
The dns key came out of real incidents: containers that couldn't resolve names after a reboot
and stale DNS inside containers. If your
resolver returns IPv6 addresses but the host has no IPv6 route, read
AAAA records with no IPv6 route too.
Configure default logging driver¶
Docker provides logging drivers for
collecting and viewing log data from all containers running on a host. The
default logging driver, json-file, writes log data to JSON-formatted files on
the host filesystem. Over time, these log files expand in size, leading to
potential exhaustion of disk resources.
To avoid issues with overusing disk for log data, consider one of the following options:
- Configure the
json-filelogging driver to turn on log rotation. - Use an alternative logging driver such as the "local" logging driver that performs log rotation by default.
- Use a logging driver that sends logs to a remote logging aggregator.
To use the json-file driver as the default logging driver, set the log-driver and log-opts keys to appropriate values in the daemon.json file, which is located in /etc/docker/ on Linux hosts or C:\ProgramData\docker\config\ on Windows Server. If the file does not exist, create it first. For more information about configuring Docker using daemon.json, see daemon.json.
The following example sets the log driver to json-file and sets the max-size and max-file options to enable automatic log-rotation.
Note
log-opts configuration options in the daemon.json configuration file must be provided as strings. Boolean and numeric values (such as the value for max-file in the example above) must therefore be enclosed in quotes (").
Restart Docker for the changes to take effect for newly created containers. Existing containers don't use the new logging configuration automatically.
You can set the logging driver for a specific container by using the --log-driver flag to docker container create or docker run:
Next steps¶
- Compose conventions: pinned tags, healthchecks, autoheal, and validating before deploy.
- Take a look at the Docker workshop to learn how to build an image and run it as a containerized application.
Sources & further reading¶
- Linux post-installation steps
- Daemon configuration file (incl. which options are reloadable with SIGHUP)
- json-file logging driver
- Docker Engine security: daemon attack surface