docker debug

DescriptionGet a shell into any container or image. An alternative to debugging with `docker exec`.
Usagedebug [OPTIONS] {CONTAINER|IMAGE}

Description

Note

Docker Debug requires a Pro, Team, or Business subcription. You must sign in to use this command.

Docker Debug is a CLI command that helps you follow best practices by keeping your images small and secure. With Docker Debug, you can debug your images while they contain the bare minimum to run your application. It does this by letting you create and work with slim images or containers that are often difficult to debug because all tools have been removed. For example, while typical debug approaches like docker exec -it my-app bash may not work on a slim container, docker debug will work.

With docker debug you can get a debug shell into any container or image, even if they don't contain a shell. You don't need to modify the image to use Docker Debug. However, using Docker Debug still won't modify your image. Docker Debug brings its own toolbox that you can easily customize. The toolbox comes with many standard Linux tools pre-installed, such as vim, nano, htop, and curl. Use the builtin install command to add additional tools available on https://search.nixos.org/packages. Docker Debug supports bash, fish, and zsh. By default it tries to auto-detect your shell.

Custom builtin tools:

  • install [tool1] [tool2]: Add Nix packages from: https://search.nixos.org/packages, see example.
  • uninstall [tool1] [tool2]: Uninstall Nix packages.
  • entrypoint: Print, lint, or run the entrypoint, see example.
  • builtins: Show custom builtin tools.

Note

For images and stopped containers, all changes are discarded when leaving the shell. At no point, do changes affect the actual image or container. When accessing running or paused containers, all filesystem changes are directly visible to the container. The /nix directory is never visible to the actual image or container.

Options

OptionDefaultDescription
--shellautoSelect a shell. Supported: bash, fish, zsh, auto.
-c, --commandEvaluate the specified commands instead of starting an interactive session, see example.
--hostDaemon docker socket to connect to. E.g.: ssh://root@example.org, unix:///some/path/docker.sock, see example.

Examples

Debugging containers that have no shell (slim containers)

The hello-world image is very simple and only contains the /hello binary. It's a good example of a slim image. There are no other tools and no shell.

Run a container from the hello-world image:

$ docker run --name my-app hello-world

The container exits immediately. To get a debug shell inside, run:

$ docker debug my-app

The debug shell allows you to inspect the filesystem:

docker > ls
dev  etc  hello  nix  proc  sys

The file /hello is the binary that was executed when running the container. You can confirm this by running it directly:

docker > /hello

After running the binary, it produces the same output.

Debugging (slim) images

You can debug images directly by running:

$ docker debug hello-world
...
docker > ls
dev  etc  hello  nix  proc  sys

You don't even need to pull the image as docker debug will do this automatically like the docker run command.

Modifying files of a running container

Docker debug lets you modify files in any running container. The toolbox comes with vim and nano pre-installed.

Run an nginx container and change the default index.html:

$ docker run -d --name web-app -p 8080:80 nginx
d3d6074d0ea901c96cac8e49e6dad21359616bef3dc0623b3c2dfa536c31dfdb

To confirm nginx is running, open a browser and navigate to http://localhost:8080. You should see the default nginx page. Now, change it using vim:

vim /usr/share/nginx/html/index.html

Change the title to "Welcome to my app!" and save the file. Now, reload the page in the browser and you should see the updated page.

Managing your toolbox using the install command

The builtin install command lets you add any tool from https://search.nixos.org/packages to the toolbox. Keep in mind adding a tool never modifies the actual image or container. Tools get added to only your toolbox. Run docker debug and then install nmap:

$ docker debug nginx
...
docker > install nmap
Tip: You can install any package available at: https://search.nixos.org/packages.
installing 'nmap-7.93'
these 2 paths will be fetched (5.58 MiB download, 26.27 MiB unpacked):
/nix/store/brqjf4i23fagizaq2gn4d6z0f406d0kg-lua-5.3.6
/nix/store/xqd17rhgmn6pg85a3g18yqxpcya6d06r-nmap-7.93
copying path '/nix/store/brqjf4i23fagizaq2gn4d6z0f406d0kg-lua-5.3.6' from 'https://cache.nixos.org'...
copying path '/nix/store/xqd17rhgmn6pg85a3g18yqxpcya6d06r-nmap-7.93' from 'https://cache.nixos.org'...
building '/nix/store/k8xw5wwarh8dc1dvh5zx8rlwamxfsk3d-user-environment.drv'...

docker > nmap --version
Nmap version 7.93 ( https://nmap.org )
Platform: x86_64-unknown-linux-gnu
Compiled with: liblua-5.3.6 openssl-3.0.11 libssh2-1.11.0 nmap-libz-1.2.12 libpcre-8.45 libpcap-1.10.4 nmap-libdnet-1.12 ipv6
Compiled without:
Available nsock engines: epoll poll select

You can confirm nmap is now part of your toolbox by getting a debug shell into a different image:

$ docker debug hello-world
...
docker > nmap --version

Nmap version 7.93 ( https://nmap.org )
Platform: x86_64-unknown-linux-gnu
Compiled with: liblua-5.3.6 openssl-3.0.11 libssh2-1.11.0 nmap-libz-1.2.12 libpcre-8.45 libpcap-1.10.4 nmap-libdnet-1.12 ipv6
Compiled without:
Available nsock engines: epoll poll select

docker > exit

nmap is still there.

Understanding the default startup command of a container (entry points)

Docker Debug comes with a builtin tool, entrypoint. Enter the hello-world image and confirm the entrypoint is /hello:

$ docker debug hello-world
...
docker > entrypoint --print
/hello

The entrypoint command evaluates the ENTRYPOINT and CMD statement of the underlying image and lets you print, lint, or run the resulting entrypoint. However, it can be difficult to understand all the corner cases from Understand how CMD and ENTRYPOINT interact. In these situations, entrypoint can help.

Use entrypoint to investigate what actually happens when you run a container from the Nginx image:

$ docker debug nginx
...
docker > entrypoint
Understand how ENTRYPOINT/CMD work and if they are set correctly.
From CMD in Dockerfile:
 ['nginx', '-g', 'daemon off;']

From ENTRYPOINT in Dockerfile:
 ['/docker-entrypoint.sh']

By default, any container from this image will be started with following   command:

/docker-entrypoint.sh nginx -g daemon off;

path: /docker-entrypoint.sh
args: nginx -g daemon off;
cwd:
PATH: /usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin

Lint results:
 PASS: '/docker-entrypoint.sh' found
 PASS: no mixing of shell and exec form
 PASS: no double use of shell form

Docs:
- https://docs.docker.com/reference/dockerfile/#cmd
- https://docs.docker.com/reference/dockerfile/#entrypoint
- https://docs.docker.com/reference/dockerfile/#understand-how-cmd-and-entrypoint-interact

The output tells you that on startup of the nginx image, a script /docker-entrypoint.sh is executed with the arguments nginx -g daemon off;. You can test the entrypoint by using the --run option:

$ docker debug nginx
...
docker > entrypoint --run
/docker-entrypoint.sh: /docker-entrypoint.d/ is not empty, will attempt to perform configuration
/docker-entrypoint.sh: Looking for shell scripts in /docker-entrypoint.d/
/docker-entrypoint.sh: Launching /docker-entrypoint.d/10-listen-on-ipv6-by-default.sh
10-listen-on-ipv6-by-default.sh: info: Getting the checksum of /etc/nginx/conf.d/default.conf
10-listen-on-ipv6-by-default.sh: info: Enabled listen on IPv6 in /etc/nginx/conf.d/default.conf
/docker-entrypoint.sh: Sourcing /docker-entrypoint.d/15-local-resolvers.envsh
/docker-entrypoint.sh: Launching /docker-entrypoint.d/20-envsubst-on-templates.sh
/docker-entrypoint.sh: Launching /docker-entrypoint.d/30-tune-worker-processes.sh
/docker-entrypoint.sh: Configuration complete; ready for start up
2024/01/19 17:34:39 [notice] 50#50: using the "epoll" event method
2024/01/19 17:34:39 [notice] 50#50: nginx/1.25.3
2024/01/19 17:34:39 [notice] 50#50: built by gcc 12.2.0 (Debian 12.2.0-14)
2024/01/19 17:34:39 [notice] 50#50: OS: Linux 5.15.133.1-microsoft-standard-WSL2
2024/01/19 17:34:39 [notice] 50#50: getrlimit(RLIMIT_NOFILE): 1048576:1048576
2024/01/19 17:34:39 [notice] 50#50: start worker processes
2024/01/19 17:34:39 [notice] 50#50: start worker process 77
...

This starts nginx in your debug shell without having to actually run a container. You can shutdown nginx by pressing Ctrl+C.

Running commands directly (e.g., for scripting)

Use the --command option to evaluate a command directly instead of starting an interactive session. For example, this is similar to bash -c "arg1 arg2 ...". The following example runs the cat command in the nginx image without starting an interactive session.

$ docker debug --command "cat /usr/share/nginx/html/index.html" nginx

<!DOCTYPE html>
<html>
<head>
<title>Welcome to nginx!</title>
<style>
html { color-scheme: light dark; }
body { width: 35em; margin: 0 auto;
font-family: Tahoma, Verdana, Arial, sans-serif; }
</style>
</head>
<body>
<h1>Welcome to nginx!</h1>
<p>If you see this page, the nginx web server is successfully installed and
working. Further configuration is required.</p>

<p>For online documentation and support please refer to
<a href="http://nginx.org/">nginx.org</a>.<br/>
Commercial support is available at
<a href="http://nginx.com/">nginx.com</a>.</p>

<p><em>Thank you for using nginx.</em></p>
</body>
</html>

Remote debugging using the --host option

The following examples shows how to use the --host option. The first example uses SSH to connect to a remote Docker instance at example.org as the root user, and get a shell into the my-container container.

$ docker debug --host ssh://root@example.org my-container

The following example connects to a different local Docker Engine, and gets a shell into the my-container container.

$ docker debug --host=unix:///some/path/docker.sock my-container