Skip to content

Latest commit

 

History

History
678 lines (517 loc) · 31.5 KB

README.md

File metadata and controls

678 lines (517 loc) · 31.5 KB

Docker container for Firefox

Release Docker Image Size Docker Pulls Docker Stars Build Status Donate

This project implements a Docker container for Firefox.

The GUI of the application is accessed through a modern web browser (no installation or configuration needed on the client side) or via any VNC client.


Firefox logoFirefox

Mozilla Firefox is a free and open-source web browser developed by Mozilla Foundation and its subsidiary, Mozilla Corporation.


Table of Content

Quick Start

NOTE: The Docker command provided in this quick start is given as an example and parameters should be adjusted to your need.

Launch the Firefox docker container with the following command:

docker run -d \
    --name=firefox \
    -p 5800:5800 \
    -v /docker/appdata/firefox:/config:rw \
    jlesage/firefox

Where:

  • /docker/appdata/firefox: This is where the application stores its configuration, states, log and any files needing persistency.

Browse to http://your-host-ip:5800 to access the Firefox GUI.

Usage

docker run [-d] \
    --name=firefox \
    [-e <VARIABLE_NAME>=<VALUE>]... \
    [-v <HOST_DIR>:<CONTAINER_DIR>[:PERMISSIONS]]... \
    [-p <HOST_PORT>:<CONTAINER_PORT>]... \
    jlesage/firefox
Parameter Description
-d Run the container in the background. If not set, the container runs in the foreground.
-e Pass an environment variable to the container. See the Environment Variables section for more details.
-v Set a volume mapping (allows to share a folder/file between the host and the container). See the Data Volumes section for more details.
-p Set a network port mapping (exposes an internal container port to the host). See the Ports section for more details.

Environment Variables

To customize some properties of the container, the following environment variables can be passed via the -e parameter (one for each variable). Value of this parameter has the format <VARIABLE_NAME>=<VALUE>.

Variable Description Default
USER_ID ID of the user the application runs as. See User/Group IDs to better understand when this should be set. 1000
GROUP_ID ID of the group the application runs as. See User/Group IDs to better understand when this should be set. 1000
SUP_GROUP_IDS Comma-separated list of supplementary group IDs of the application. (no value)
UMASK Mask that controls how permissions are set for newly created files and folders. The value of the mask is in octal notation. By default, the default umask value is 0022, meaning that newly created files and folders are readable by everyone, but only writable by the owner. See the online umask calculator at http://wintelguy.com/umask-calc.pl. 0022
LANG Set the locale, which defines the application's language, if supported. Format of the locale is language[_territory][.codeset], where language is an ISO 639 language code, territory is an ISO 3166 country code and codeset is a character set, like UTF-8. For example, Australian English using the UTF-8 encoding is en_AU.UTF-8. en_US.UTF-8
TZ TimeZone used by the container. Timezone can also be set by mapping /etc/localtime between the host and the container. Etc/UTC
KEEP_APP_RUNNING When set to 1, the application will be automatically restarted when it crashes or terminates. 0
APP_NICENESS Priority at which the application should run. A niceness value of -20 is the highest priority and 19 is the lowest priority. The default niceness value is 0. NOTE: A negative niceness (priority increase) requires additional permissions. In this case, the container should be run with the docker option --cap-add=SYS_NICE. 0
INSTALL_PACKAGES Space-separated list of packages to install during the startup of the container. List of available packages can be found at https://mirrors.alpinelinux.org. ATTENTION: Container functionality can be affected when installing a package that overrides existing container files (e.g. binaries). (no value)
PACKAGES_MIRROR Mirror of the repository to use when installing packages. List of mirrors is available at https://mirrors.alpinelinux.org. (no value)
CONTAINER_DEBUG Set to 1 to enable debug logging. 0
DISPLAY_WIDTH Width (in pixels) of the application's window. 1920
DISPLAY_HEIGHT Height (in pixels) of the application's window. 1080
DARK_MODE When set to 1, dark mode is enabled for the application. 0
WEB_AUDIO When set to 1, audio support is enabled, meaning that any audio produced by the application is played through the browser. Note that audio is not supported for VNC clients. 0
WEB_AUTHENTICATION When set to 1, the application' GUI is protected via a login page when accessed via a web browser. Access is allowed only when providing valid credentials. NOTE: This feature requires secure connection (SECURE_CONNECTION environment variable) to be enabled. 0
WEB_AUTHENTICATION_USERNAME Optional username to configure for the web authentication. This is a quick and easy way to configure credentials for a single user. To configure credentials in a more secure way, or to add more users, see the Web Authentication section. (no value)
WEB_AUTHENTICATION_PASSWORD Optional password to configure for the web authentication. This is a quick and easy way to configure credentials for a single user. To configure credentials in a more secure way, or to add more users, see the Web Authentication section. (no value)
SECURE_CONNECTION When set to 1, an encrypted connection is used to access the application's GUI (either via a web browser or VNC client). See the Security section for more details. 0
SECURE_CONNECTION_VNC_METHOD Method used to perform the secure VNC connection. Possible values are SSL or TLS. See the Security section for more details. SSL
SECURE_CONNECTION_CERTS_CHECK_INTERVAL Interval, in seconds, at which the system verifies if web or VNC certificates have changed. When a change is detected, the affected services are automatically restarted. A value of 0 disables the check. 60
WEB_LISTENING_PORT Port used by the web server to serve the UI of the application. This port is used internally by the container and it is usually not required to be changed. By default, a container is created with the default bridge network, meaning that, to be accessible, each internal container port must be mapped to an external port (using the -p or --publish argument). However, if the container is created with another network type, changing the port used by the container might be useful to prevent conflict with other services/containers. NOTE: a value of -1 disables listening, meaning that the application's UI won't be accessible over HTTP/HTTPs. 5800
VNC_LISTENING_PORT Port used by the VNC server to serve the UI of the application. This port is used internally by the container and it is usually not required to be changed. By default, a container is created with the default bridge network, meaning that, to be accessible, each internal container port must be mapped to an external port (using the -p or --publish argument). However, if the container is created with another network type, changing the port used by the container might be useful to prevent conflict with other services/containers. NOTE: a value of -1 disables listening, meaning that the application's UI won't be accessible over VNC. 5900
VNC_PASSWORD Password needed to connect to the application's GUI. See the VNC Password section for more details. (no value)
ENABLE_CJK_FONT When set to 1, open-source computer font WenQuanYi Zen Hei is installed. This font contains a large range of Chinese/Japanese/Korean characters. 0
FF_OPEN_URL The URL to open when Firefox starts. (no value)
FF_KIOSK Set to 1 to enable kiosk mode. This mode launches Firefox in a very restricted and limited mode best suitable for public areas or customer-facing displays. 0
FF_CUSTOM_ARGS Custom argument(s) to pass when launching Firefox. 0

Deployment Considerations

Many tools used to manage Docker containers extract environment variables defined by the Docker image and use them to create/deploy the container. For example, this is done by:

  • The Docker application on Synology NAS
  • The Container Station on QNAP NAS
  • Portainer
  • etc.

While this can be useful for the user to adjust the value of environment variables to fit its needs, it can also be confusing and dangerous to keep all of them.

A good practice is to set/keep only the variables that are needed for the container to behave as desired in a specific setup. If the value of variable is kept to its default value, it means that it can be removed. Keep in mind that all variables are optional, meaning that none of them is required for the container to start.

Removing environment variables that are not needed provides some advantages:

  • Prevents keeping variables that are no longer used by the container. Over time, with image updates, some variables might be removed.
  • Allows the Docker image to change/fix a default value. Again, with image updates, the default value of a variable might be changed to fix an issue, or to better support a new feature.
  • Prevents changes to a variable that might affect the correct function of the container. Some undocumented variables, like PATH or ENV, are required to be exposed, but are not meant to be changed by users. However, container management tools still show these variables to users.
  • There is a bug with the Container Station on QNAP and the Docker application on Synology, where an environment variable without value might not be allowed. This behavior is wrong: it's absolutely fine to have a variable without value. In fact, this container does have variables without value by default. Thus, removing unneeded variables is a good way to prevent deployment issue on these devices.

Data Volumes

The following table describes data volumes used by the container. The mappings are set via the -v parameter. Each mapping is specified with the following format: <HOST_DIR>:<CONTAINER_DIR>[:PERMISSIONS].

Container path Permissions Description
/config rw This is where the application stores its configuration, states, log and any files needing persistency.

Ports

Here is the list of ports used by the container.

When using the default bridge network, ports can be mapped to the host via the -p parameter (one per port mapping). Each mapping is defined with the following format: <HOST_PORT>:<CONTAINER_PORT>. The port number used inside the container might not be changeable, but you are free to use any port on the host side.

See the Docker Container Networking documentation for more details.

Port Protocol Mapping to host Description
5800 TCP Optional Port to access the application's GUI via the web interface. Mapping to the host is optional if access through the web interface is not wanted. For a container not using the default bridge network, the port can be changed with the WEB_LISTENING_PORT environment variable.
5900 TCP Optional Port to access the application's GUI via the VNC protocol. Mapping to the host is optional if access through the VNC protocol is not wanted. For a container not using the default bridge network, the port can be changed with the VNC_LISTENING_PORT environment variable.

Changing Parameters of a Running Container

As can be seen, environment variables, volume and port mappings are all specified while creating the container.

The following steps describe the method used to add, remove or update parameter(s) of an existing container. The general idea is to destroy and re-create the container:

  1. Stop the container (if it is running):
docker stop firefox
  1. Remove the container:
docker rm firefox
  1. Create/start the container using the docker run command, by adjusting parameters as needed.

NOTE: Since all application's data is saved under the /config container folder, destroying and re-creating a container is not a problem: nothing is lost and the application comes back with the same state (as long as the mapping of the /config folder remains the same).

Docker Compose File

Here is an example of a docker-compose.yml file that can be used with Docker Compose.

Make sure to adjust according to your needs. Note that only mandatory network ports are part of the example.

version: '3'
services:
  firefox:
    image: jlesage/firefox
    ports:
      - "5800:5800"
    volumes:
      - "/docker/appdata/firefox:/config:rw"

Docker Image Versioning

Each release of a Docker image is versioned. Prior to october 2022, the semantic versioning was used as the versioning scheme.

Since then, versioning scheme changed to calendar versioning. The format used is YY.MM.SEQUENCE, where:

  • YY is the zero-padded year (relative to year 2000).
  • MM is the zero-padded month.
  • SEQUENCE is the incremental release number within the month (first release is 1, second is 2, etc).

Docker Image Update

Because features are added, issues are fixed, or simply because a new version of the containerized application is integrated, the Docker image is regularly updated. Different methods can be used to update the Docker image.

The system used to run the container may have a built-in way to update containers. If so, this could be your primary way to update Docker images.

An other way is to have the image be automatically updated with Watchtower. Watchtower is a container-based solution for automating Docker image updates. This is a "set and forget" type of solution: once a new image is available, Watchtower will seamlessly perform the necessary steps to update the container.

Finally, the Docker image can be manually updated with these steps:

  1. Fetch the latest image:
docker pull jlesage/firefox
  1. Stop the container:
docker stop firefox
  1. Remove the container:
docker rm firefox
  1. Create and start the container using the docker run command, with the the same parameters that were used when it was deployed initially.

Synology

For owners of a Synology NAS, the following steps can be used to update a container image.

  1. Open the Docker application.
  2. Click on Registry in the left pane.
  3. In the search bar, type the name of the container (jlesage/firefox).
  4. Select the image, click Download and then choose the latest tag.
  5. Wait for the download to complete. A notification will appear once done.
  6. Click on Container in the left pane.
  7. Select your Firefox container.
  8. Stop it by clicking Action->Stop.
  9. Clear the container by clicking Action->Reset (or Action->Clear if you don't have the latest Docker application). This removes the container while keeping its configuration.
  10. Start the container again by clicking Action->Start. NOTE: The container may temporarily disappear from the list while it is re-created.

unRAID

For unRAID, a container image can be updated by following these steps:

  1. Select the Docker tab.
  2. Click the Check for Updates button at the bottom of the page.
  3. Click the update ready link of the container to be updated.

User/Group IDs

When using data volumes (-v flags), permissions issues can occur between the host and the container. For example, the user within the container may not exist on the host. This could prevent the host from properly accessing files and folders on the shared volume.

To avoid any problem, you can specify the user the application should run as.

This is done by passing the user ID and group ID to the container via the USER_ID and GROUP_ID environment variables.

To find the right IDs to use, issue the following command on the host, with the user owning the data volume on the host:

id <username>

Which gives an output like this one:

uid=1000(myuser) gid=1000(myuser) groups=1000(myuser),4(adm),24(cdrom),27(sudo),46(plugdev),113(lpadmin)

The value of uid (user ID) and gid (group ID) are the ones that you should be given the container.

Accessing the GUI

Assuming that container's ports are mapped to the same host's ports, the graphical interface of the application can be accessed via:

  • A web browser:
http://<HOST IP ADDR>:5800
  • Any VNC client:
<HOST IP ADDR>:5900

Security

By default, access to the application's GUI is done over an unencrypted connection (HTTP or VNC).

Secure connection can be enabled via the SECURE_CONNECTION environment variable. See the Environment Variables section for more details on how to set an environment variable.

When enabled, application's GUI is performed over an HTTPs connection when accessed with a browser. All HTTP accesses are automatically redirected to HTTPs.

When using a VNC client, the VNC connection is performed over SSL. Note that few VNC clients support this method. SSVNC is one of them.

SSVNC

SSVNC is a VNC viewer that adds encryption security to VNC connections.

While the Linux version of SSVNC works well, the Windows version has some issues. At the time of writing, the latest version 1.0.30 is not functional, as a connection fails with the following error:

ReadExact: Socket error while reading

However, for your convenience, an unofficial and working version is provided here:

https://github.com/jlesage/docker-baseimage-gui/raw/master/tools/ssvnc_windows_only-1.0.30-r1.zip

The only difference with the official package is that the bundled version of stunnel has been upgraded to version 5.49, which fixes the connection problems.

Certificates

Here are the certificate files needed by the container. By default, when they are missing, self-signed certificates are generated and used. All files have PEM encoded, x509 certificates.

Container Path Purpose Content
/config/certs/vnc-server.pem VNC connection encryption. VNC server's private key and certificate, bundled with any root and intermediate certificates.
/config/certs/web-privkey.pem HTTPs connection encryption. Web server's private key.
/config/certs/web-fullchain.pem HTTPs connection encryption. Web server's certificate, bundled with any root and intermediate certificates.

NOTE: To prevent any certificate validity warnings/errors from the browser or VNC client, make sure to supply your own valid certificates.

NOTE: Certificate files are monitored and relevant daemons are automatically restarted when changes are detected.

VNC Password

To restrict access to your application, a password can be specified. This can be done via two methods:

  • By using the VNC_PASSWORD environment variable.
  • By creating a .vncpass_clear file at the root of the /config volume. This file should contain the password in clear-text. During the container startup, content of the file is obfuscated and moved to .vncpass.

The level of security provided by the VNC password depends on two things:

  • The type of communication channel (encrypted/unencrypted).
  • How secure the access to the host is.

When using a VNC password, it is highly desirable to enable the secure connection to prevent sending the password in clear over an unencrypted channel.

ATTENTION: Password is limited to 8 characters. This limitation comes from the Remote Framebuffer Protocol RFC (see section 7.2.2). Any characters beyond the limit are ignored.

Web Authentication

Access to the application's GUI via a web browser can be protected with a login page. When web authentication is enabled, users have to provide valid credentials, otherwise access is denied.

Web authentication can be enabled by setting the WEB_AUTHENTICATION environment variable to 1.

See the Environment Variables section for more details on how to set an environment variable.

NOTE: Secure connection must be also enabled to use web authentication. See the Security section for more details.

Configuring Users Credentials

Two methods can be used to configure users credentials:

  1. Via container environment variables.
  2. Via password database.

Containers environment variables can be used to quickly and easily configure a single user. Username and pasword are defined via the following environment variables:

  • WEB_AUTHENTICATION_USERNAME
  • WEB_AUTHENTICATION_PASSWORD

See the Environment Variables section for more details on how to set an environment variable.

The second method is more secure and allows multiple users to be configured. The usernames and password hashes are saved into a password database, located at /config/webauth-htpasswd inside the container. This database file has the same format as htpasswd files of the Apache HTTP server. Note that password themselves are not saved into the database, but only their hash. The bcrypt password hashing function is used to generate hashes.

Users are managed via the webauth-user tool included in the container:

  • To add a user password: docker exec -ti <container name or id> webauth-user add <username>.
  • To update a user password: docker exec -ti <container name or id> webauth-user update <username>.
  • To remove a user: docker exec <container name or id> webauth-user del <username>.
  • To list users: docker exec <container name or id> webauth-user user.

Reverse Proxy

The following sections contain NGINX configurations that need to be added in order to reverse proxy to this container.

A reverse proxy server can route HTTP requests based on the hostname or the URL path.

Routing Based on Hostname

In this scenario, each hostname is routed to a different application/container.

For example, let's say the reverse proxy server is running on the same machine as this container. The server would proxy all HTTP requests sent to firefox.domain.tld to the container at 127.0.0.1:5800.

Here are the relevant configuration elements that would be added to the NGINX configuration:

map $http_upgrade $connection_upgrade {
	default upgrade;
	''      close;
}

upstream docker-firefox {
	# If the reverse proxy server is not running on the same machine as the
	# Docker container, use the IP of the Docker host here.
	# Make sure to adjust the port according to how port 5800 of the
	# container has been mapped on the host.
	server 127.0.0.1:5800;
}

server {
	[...]

	server_name firefox.domain.tld;

	location / {
	        proxy_pass http://docker-firefox;
	}

	location /websockify {
		proxy_pass http://docker-firefox;
		proxy_http_version 1.1;
		proxy_set_header Upgrade $http_upgrade;
		proxy_set_header Connection $connection_upgrade;
		proxy_read_timeout 86400;
	}
}

Routing Based on URL Path

In this scenario, the hostname is the same, but different URL paths are used to route to different applications/containers.

For example, let's say the reverse proxy server is running on the same machine as this container. The server would proxy all HTTP requests for server.domain.tld/firefox to the container at 127.0.0.1:5800.

Here are the relevant configuration elements that would be added to the NGINX configuration:

map $http_upgrade $connection_upgrade {
	default upgrade;
	''      close;
}

upstream docker-firefox {
	# If the reverse proxy server is not running on the same machine as the
	# Docker container, use the IP of the Docker host here.
	# Make sure to adjust the port according to how port 5800 of the
	# container has been mapped on the host.
	server 127.0.0.1:5800;
}

server {
	[...]

	location = /firefox {return 301 $scheme://$http_host/firefox/;}
	location /firefox/ {
		proxy_pass http://docker-firefox/;
		location /firefox/websockify {
			proxy_pass http://docker-firefox/websockify/;
			proxy_http_version 1.1;
			proxy_set_header Upgrade $http_upgrade;
			proxy_set_header Connection $connection_upgrade;
			proxy_read_timeout 86400;
		}
	}
}

Shell Access

To get shell access to the running container, execute the following command:

docker exec -ti CONTAINER sh

Where CONTAINER is the ID or the name of the container used during its creation.

Allowing the membarrier System Call

To properly work, recent versions of Firefox need the membarrier system call. Without it, tabs would frequently crash.

Docker uses seccomp profile to restrict system calls available to the container. Before Docker version 20.10.0, the membarrier system call was not allowed in the default profile. If you run a such version, you can use one of the following solutions, from the most to the least secure, to provide the container permission to use this sytem call:

  1. Run the container with a custom seccomp profile allowing the membarrier system call. The latest official seccomp profile can be used. Download the file and then add the following parameter when creating the container: --security-opt seccomp=/path/to/seccomp_profile.json.
  2. Run the container without the default seccomp profile (thus allowing all system calls). Use the following parameter when creating the container: --security-opt seccomp=unconfined.
  3. Run the container in privileged mode. This effectively disables usage of seccomp. Add the --privileged parameter when creating the container.

Sound Support

For Firefox to be able to use the audio device available on the host, /dev/snd must be exposed to the container by adding the --device /dev/snd parameter to the docker run command.

Setting Firefox Preferences Via Environment Variables

Firefox preferences can be set via environment variables passed to the container. During the startup, a script process all these variables and modify the preference file accordingly.

The name of the environment variable must start with FF_PREF_, followed by a string of your choice. For example, FF_PREF_MY_PREF is a valid name.

The content of the variable should be in the format NAME=VAL, where NAME is the name of the preference (as found in the about:config page) and VAL is its value. A value can be one of the following types:

  • string
  • integer
  • boolean

It is important to note that a value of type string should be surrounded by double quotes. Other types don't need them.

For example, to set the network.proxy.http preference, one would pass the environment variable to the container by adding the following argument to the docker run command:

-e "FF_PREF_HTTP_PROXY=network.proxy.http=\"proxy.example.com\""

If a preference needs to be removed, its value should be set to UNSET. For example:

-e "FF_PREF_HTTP_PROXY=network.proxy.http=UNSET"

NOTE: This is an advanced usage and it is recommended to set preferences via Firefox directly.

Troubleshooting

Crashes

If Firefox is crashing frequently, make sure that:

  • The membarrier system call is not blocked by Docker. See the Allowing the membarrier System Call for more details.
  • Make sure the kernel of your Linux distribution is up-to-date.

Support or Contact

Having troubles with the container or have questions? Please create a new issue.

For other great Dockerized applications, see https://jlesage.github.io/docker-apps.