The Container Structure Tests provide a powerful framework to validate the structure of a container image. These tests can be used to check the output of commands in an image, as well as verify metadata and contents of the filesystem.

Tests can be run either through a standalone binary, or through a Docker image.

Note: container-structure-test is not an officially supported Google project, and is currently in maintainence mode. Contributions are still welcome!

Install via brew:

$ brew install container-structure-test

curl -LO https://github.com/GoogleContainerTools/container-structure-test/releases/latest/download/container-structure-test-darwin-arm64 && chmod  x container-structure-test-darwin-arm64 && sudo mv container-structure-test-darwin-arm64 /usr/local/bin/container-structure-test

curl -LO https://github.com/GoogleContainerTools/container-structure-test/releases/latest/download/container-structure-test-linux-amd64 && chmod  x container-structure-test-linux-amd64 && sudo mv container-structure-test-linux-amd64 /usr/local/bin/container-structure-test

If you want to avoid using sudo:

curl -LO https://github.com/GoogleContainerTools/container-structure-test/releases/latest/download/container-structure-test-linux-amd64 && chmod  x container-structure-test-linux-amd64 && mkdir -p $HOME/bin && export PATH=$PATH:$HOME/bin && mv container-structure-test-linux-amd64 $HOME/bin/container-structure-test

Additionally, a container image for running tests through Google Cloud Builder can be found at gcr.io/gcp-runtimes/container-structure-test:latest.

To use container structure tests to validate your containers, you'll need the following:

• The container structure test binary or docker image
• A container image to test against
• A test .yaml or .json file with user defined structure tests to run inside of the specified container image

Note that the test framework looks for the provided image in the local Docker daemon (if it is not provided as a tar). The --pull flag can optionally be provided to force a pull of a remote image before running the tests.

An example run of the test framework:

container-structure-test test --image gcr.io/registry/image:latest \
--config config.yaml

Tests within this framework are specified through a YAML or JSON config file, which is provided to the test driver via a CLI flag. Multiple config files may be specified in a single test run. The config file will be loaded in by the test driver, which will execute the tests in order. Within this config file, four types of tests can be written:

• Command Tests (testing output/error of a specific command issued)
• File Existence Tests (making sure a file is, or isn't, present in the file system of the image)
• File Content Tests (making sure files in the file system of the image contain, or do not contain, specific contents)
• Metadata Test, singular (making sure certain container metadata is correct)

Command tests ensure that certain commands run properly in the target image. Regexes can be used to check for expected or excluded strings in both stdout and stderr. Additionally, any number of flags can be passed to the argument as normal. Each command in the setup section will run in a separate container and then commits a modified image to be the new base image for the test run.

NOTE: schemaVersion must be specified in all container-structure-test yamls. The current version is 2.0.0.

• Name (string, required): The name of the test
• Setup ([][]string, optional): A list of commands (each with optional flags) to run before the actual command under test.
• Teardown ([][]string, optional): A list of commands (each with optional flags) to run after the actual command under test.
• Command (string, required): The command to run in the test.
• Args ([]string, optional): The arguments to pass to the command.
• EnvVars ([]EnvVar, optional): A list of environment variables to set for the individual test. See the Environment Variables section for more info.
• Expected Output ([]string, optional): List of regexes that should match the stdout from running the command.
• Excluded Output ([]string, optional): List of regexes that should not match the stdout from running the command.
• Expected Error ([]string, optional): List of regexes that should match the stderr from running the command.
• Excluded Error ([]string, optional): List of regexes that should not match the stderr from running the command.
• Exit Code (int, optional): Exit code that the command should exit with.

Example:

commandTests:
  - name: "gunicorn flask"
    setup: [["virtualenv", "/env"], ["pip", "install", "gunicorn", "flask"]]
    command: "which"
    args: ["gunicorn"]
    expectedOutput: ["/env/bin/gunicorn"]
  - name:  "apt-get upgrade"
    command: "apt-get"
    args: ["-qqs", "upgrade"]
    excludedOutput: [".*Inst.*Security.* | .*Security.*Inst.*"]
    excludedError: [".*Inst.*Security.* | .*Security.*Inst.*"]

Depending on your command the argument section can get quite long. Thus, you can make use of YAML's list style option for separation of arguments and the literal style option to preserve newlines like so:

commandTests:
  - name: "say hello world"
    command: "bash"
    args:
      - -c
      - |
         echo hello &&
         echo world

To avoid unexpected behavior and output when running commands in the containers, all entrypoints are overwritten by default. If your entrypoint is necessary for the structure of your container, use the setup field to call any scripts or commands manually before running the tests.

commandTests:
  ...
  setup: [["my_image_entrypoint.sh"]]
  ...

Each command test run creates either a container (with the docker driver) or tar artifact (with the tar driver). By default, these are deleted after the test run finishes, but the --save flag can optionally be passed to keep these around. This would normally be used for debugging purposes.

File existence tests check to make sure a specific file (or directory) exist within the file system of the image. No contents of the files or directories are checked. These tests can also be used to ensure a file or directory is not present in the file system.

• Name (string, required): The name of the test
• Path (string, required): Path to the file or directory under test
• ShouldExist (boolean, required): Whether or not the specified file or directory should exist in the file system
• Permissions (string, optional): The expected Unix permission string (e.g. drwxrwxrwx) of the files or directory.
• Uid (int, optional): The expected Unix user ID of the owner of the file or directory.
• Gid (int, optional): The expected Unix group ID of the owner of the file or directory.
• IsExecutableBy (string, optional): Checks if file is executable by a given user. One of owner, group, other or any

Example:

fileExistenceTests:
- name: 'Root'
  path: '/'
  shouldExist: true
  permissions: '-rw-r--r--'
  uid: 1000
  gid: 1000
  isExecutableBy: 'group'

File content tests open a file on the file system and check its contents. These tests assume the specified file is a file, and that it exists (if unsure about either or these criteria, see the above File Existence Tests section). Regexes can again be used to check for expected or excluded content in the specified file.

• Name (string, required): The name of the test
• Path (string, required): Path to the file under test
• ExpectedContents (string[], optional): List of regexes that should match the contents of the file
• ExcludedContents (string[], optional): List of regexes that should not match the contents of the file

Example:

fileContentTests:
- name: 'Debian Sources'
  path: '/etc/apt/sources.list'
  expectedContents: ['.*httpredir\.debian\.org.*']
  excludedContents: ['.*gce_debian_mirror.*']

The Metadata test ensures the container is configured correctly. All of these checks are optional.

• EnvVars ([]EnvVar): A list of environment variable key/value pairs that should be set in the container. isRegex (optional) interpretes the value as regex.
• UnboundEnvVars ([]EnvVar): A list of environment variable keys that should NOT be set in the container.
• Labels ([]Label): A list of image labels key/value pairs that should be set on the container. isRegex (optional) interpretes the value as regex.
• Entrypoint ([]string): The entrypoint of the container.
• Cmd ([]string): The CMD specified in the container.
• Exposed Ports ([]string): The ports exposed in the container.
• Unexposed Ports ([]string): The ports NOT exposed in the container.
• Volumes ([]string): The volumes exposed in the container.
• UnmountedVolumes ([]string): The volumes NOT exposed in the container.
• Workdir (string): The default working directory of the container.
• User (user): The default user of the container.

Example:

metadataTest:
  envVars:
    - key: foo
      value: baz
  labels:
    - key: 'com.example.vendor'
      value: 'ACME Incorporated'
    - key: 'build-date'
      value: '^\d{4}-\d{2}-\d{2}T\d{2}:\d{2}:\d{2}\.\d{6}$'
      isRegex: true
  exposedPorts: ["8080", "2345"]
  volumes: ["/test"]
  entrypoint: []
  cmd: ["/bin/bash"]
  workdir: "/app"
  user: "luke"

License tests check a list of copyright files and makes sure all licenses are allowed at Google. By default it will look at where Debian lists all copyright files, but can also look at an arbitrary list of files.

• Debian (bool, required): If the image is based on Debian, check where Debian lists all licenses.
• Files (string[], optional): A list of other files to check.

Example:

licenseTests:
- debian: true
  files: ["/foo/bar", "/baz/bat"]

A list of environment variables can optionally be specified as part of the test setup. They can either be set up globally (for all test runs), or test-local as part of individual command test runs (see the Command Tests section above). Each environment variable is specified as a key-value pair. Unix-style environment variable substitution is supported.

To specify, add a section like this to your config:

globalEnvVars:
  - key: "VIRTUAL_ENV"
    value: "/env"
  - key: "PATH"
    value: "/env/bin:$PATH"

The following fields are used to control various options and flags that may be desirable to set for the running container used to perform a structure test against an image. This allows an image author to validate certain runtime behavior that cannot be modified in the image-under-test such as running a container with an alternative user/UID or mounting a volume.

Note that these options are currently only supported with the docker driver.

The following list of options are currently supported:

containerRunOptions:
  user: "root"                  # set the --user/-u flag
  privileged: true              # set the --privileged flag (default: false)
  allocateTty: true             # set the --tty flag (default: false)
  envFile: path/to/.env         # load environment variables from file and pass to container (equivalent to --env-file)
  envVars:                      # if not empty, read each envVar from the environment and pass to test (equivalent to --env/e)
    - SECRET_KEY_FOO
    - OTHER_SECRET_BAR
  capabilities:                 # Add list of Linux capabilities (--cap-add)
    - NET_BIND_SERVICE
  bindMounts:                   # Bind mount a volume (--volume, -v)
    - /etc/example/dir:/etc/dir

This tool is released as a builder image, tagged as gcr.io/gcp-runtimes/container-structure-test, so you can specify tests in your cloudbuild.yaml:

steps:
# Build an image.
- name: 'gcr.io/cloud-builders/docker'
  args: ['build', '-t', 'gcr.io/$PROJECT_ID/image', '.']
# Test the image.
- name: 'gcr.io/gcp-runtimes/container-structure-test'
  args: ['test', '--image', 'gcr.io/$PROJECT_ID/image', '--config', 'test_config.yaml']

# Push the image.
images: ['gcr.io/$PROJECT_ID/image']

Container images can be represented in multiple formats, and the Docker image is just one of them. At their core, images are just a series of layers, each of which is a tarball, and so can be interacted with without a working Docker daemon. While running command tests currently requires a functioning Docker daemon on the host machine, File Existence/Content tests do not. This can be useful when dealing with images which have been docker exported or saved in a different image format than the Docker format, or when you're simply trying to run structure tests in an environment where Docker can't be installed.

To run tests without using a Docker daemon, users can specify a different "driver" to use in the tests, with the --driver flag.

An example test run with a different driver looks like:

container-structure-test test --driver tar --image gcr.io/registry/image:latest \
--config config.yaml

The currently supported drivers in the framework are:

• docker: the default driver. Supports all tests, and uses the Docker daemon on the host to run them. You can set the runtime to use (by example runsc to run with gVisor) using --runtime flag.
• tar: a tar driver, which extracts an image filesystem to wherever tests are running, and runs file/metadata tests against it. Does not support command tests.

Structure tests can also be run through bazel.

With Bazel 6 and bzlmod, just see https://registry.bazel.build/modules/container_structure_test. Otherwise, load the rule and its dependencies in your WORKSPACE, see bazel/test/WORKSPACE.bazel in this repo.

Load the rule definition in your BUILD file and declare a container_structure_test target, passing in your image and config file as parameters:

load("@container_structure_test//:defs.bzl", "container_structure_test")

container_structure_test(
    name = "hello_test",
    configs = ["testdata/hello.yaml"],
    image = ":hello",
)

container-structure-test test -h

  -c, --config stringArray             test config files
      --default-image-tag string       default image tag to used when loading images to the daemon. required when --image-from-oci-layout refers to a oci layout lacking the reference annotation.
  -d, --driver string                  driver to use when running tests (default "docker")
  -f, --force                          force run of host driver (without user prompt)
  -h, --help                           help for test
  -i, --image string                   path to test image
      --image-from-oci-layout string   path to the oci layout to test against
      --metadata string                path to image metadata file
      --no-color                       no color in the output
  -o, --output string                  output format for the test report (available format: text, json, junit) (default "text")
      --platform string                Set platform if host is multi-platform capable (default "linux/amd64")
      --pull                           force a pull of the image before running tests
  -q, --quiet                          flag to suppress output
      --runtime string                 runtime to use with docker driver
      --save                           preserve created containers after test run
      --test-report string             generate test report and write it to specified file (supported format: json, junit; default: json)

See this example repo for a full working example.

Reports are generated using one of the following output formats: text, json or junit. Formats like json and junit can also be used to write a report to a specified file using the --test-report.

====================================
====== Test file: config.yaml ======
====================================
=== RUN: File Existence Test: whoami
--- PASS
duration: 0s
=== RUN: Metadata Test
--- PASS
duration: 0s

=====================================
============== RESULTS ==============
=====================================
Passes:      2
Failures:    0
Duration:    0s
Total tests: 2

PASS

The following sample has been formatted.

{
  "Pass": 2,
  "Fail": 0,
  "Total": 2,
  "Duration": 0,
  "Results": [
    {
      "Name": "File Existence Test: whoami",
      "Pass": true,
      "Duration": 0
    },
    {
      "Name": "Metadata Test",
      "Pass": true,
      "Duration": 0
    }
  ]
}

The following sample has been formatted.

<?xml version="1.0"?>
<testsuites failures="0" tests="2" time="0">
  <testsuite>
    <testcase name="File Existence Test: whoami" time="0"/>
    <testcase name="Metadata Test" time="0"/>
  </testsuite>
</testsuites>