Testcontainers

Testcontainers is a NodeJS library that supports tests, providing lightweight, throwaway instances of common databases, Selenium web browsers, or anything else that can run in a Docker container.

Install

npm i -D testcontainers

Configuration

The following environment variables are supported:

Logs

• DEBUG=testcontainers Enable testcontainers logs
• DEBUG=testcontainers:containers Enable container logs
• DEBUG=testcontainers* Enable all logs

Docker

• DOCKER_HOST=tcp://docker:2375 Sets the URL of the docker daemon
• DOCKER_TLS_VERIFY=1 When set to 1, enables TLS communication with the docker daemon
• DOCKER_CERT_PATH=/some/path Configures the path to the ca.pem, cert.pem, and key.pem files used for TLS verification
• DOCKER_CONFIG=/some/path Configures the path to the config.json
• TESTCONTAINERS_HOST_OVERRIDE=docker.svc.local Docker's host on which ports are exposed
• TESTCONTAINERS_DOCKER_SOCKET_OVERRIDE=/var/run/docker.sock Path to Docker's socket. Used by ryuk and other auxiliary containers that need to perform Docker actions

Testcontainers

• TESTCONTAINERS_RYUK_PRIVILEGED=true Run ryuk as a privileged container
• TESTCONTAINERS_RYUK_DISABLED=true Disable ryuk
• TESTCONTAINERS_RYUK_PORT=65515 Explicitly set ryuk host port (not recommended)
• TESTCONTAINERS_SSHD_PORT=65515 Explicitly set SSHd host port (not recommended)
• RYUK_CONTAINER_IMAGE=registry.mycompany.com/mirror/ryuk:0.3.0 Custom image for ryuk
• SSHD_CONTAINER_IMAGE=registry.mycompany.com/mirror/sshd:1.0.0 Custom image for SSHd

Modules

Check if there already exists a pre-prepared module for your use-case.

Examples

Using a pre-built Docker image, note that omitting the tag will use latest:

const redis = require("async-redis");
const { GenericContainer } = require("testcontainers");

describe("GenericContainer", () => {
  let container;
  let redisClient;

  beforeAll(async () => {
    container = await new GenericContainer("redis")
      .withExposedPorts(6379)
      .start();

    redisClient = redis.createClient(
      container.getMappedPort(6379),
      container.getHost(),
    );
  });

  afterAll(async () => {
    await redisClient.quit();
    await container.stop();
  });

  it("works", async () => {
    await redisClient.set("key", "val");
    expect(await redisClient.get("key")).toBe("val");
  });
});

Using TypeScript:

import { 
  TestContainer, 
  StartedTestContainer, 
  StoppedTestContainer, 
  GenericContainer
} from "testcontainers";

const container: TestContainer = new GenericContainer("alpine");
const startedContainer: StartedTestContainer = await container.start();
const stoppedContainer: StoppedTestContainer = await startedContainer.stop();

Using a specific image version:

const { GenericContainer } = require("testcontainers");

const container = await new GenericContainer("alpine:3.10")
  .start();

Building and using your own Docker image:

const path = require("path");
const { GenericContainer } = require("testcontainers");

const buildContext = path.resolve(__dirname, "dir-containing-dockerfile");

const container = await GenericContainer.fromDockerfile(buildContext)
  .withBuildArg("ARG_KEY", "ARG_VALUE")
  .build();

const startedContainer = await container
  .withExposedPorts(8080)
  .start();

Using a custom Dockerfile name:

const { GenericContainer } = require("testcontainers");

const container = await GenericContainer.fromDockerfile(buildContext, "my-dockerfile")
  .build();

Creating a container with multiple exposed ports:

const { GenericContainer } = require("testcontainers");

const container = await new GenericContainer("alpine")
  .withExposedPorts(22, 80, 443)
  .start();

Enabling container reuse, note that two containers are considered equal if their options (exposed ports, commands, mounts, etc) match. This also works across multiple processes:

const { GenericContainer } = require("testcontainers");

const container1 = await new GenericContainer("alpine")
  .withCmd(["sleep", "infinity"])
  .withReuse()
  .start();

const container2 = await new GenericContainer("alpine")
  .withCmd(["sleep", "infinity"])
  .withReuse()
  .start();

assert(container1.getId() === container2.getId());

Restarting a container:

const { GenericContainer } = require("testcontainers");

const container = await new GenericContainer("alpine")
  .start();
  
await container.restart();

Creating a container with a specified name:

const { GenericContainer } = require("testcontainers");

const container = await new GenericContainer("alpine")
  .withName("custom-container-name")
  .start();

Creating a container with a command:

const { GenericContainer } = require("testcontainers");

const container = await new GenericContainer("alpine")
  .withCmd(["sleep", "infinity"])
  .start();

Execute commands inside a running container:

const { GenericContainer } = require("testcontainers");

const container = await new GenericContainer("alpine")
  .withCmd(["sleep", "infinity"])
  .start();

const { output, exitCode } = await container.exec(["echo", "hello", "world"]);

Stream logs from a running container:

const { GenericContainer } = require("testcontainers");

const container = await new GenericContainer("alpine")
  .start();

const stream = await container.logs();
stream
    .on("data", line => console.log(line))
    .on("err", line => console.error(line))
    .on("end", () => console.log("Stream closed"));

Creating a container with bind mounts:

const { GenericContainer } = require("testcontainers");

const container = await new GenericContainer("alpine")
  .withBindMount("/local/file.txt", "/remote/file.txt")
  .withBindMount("/local/dir", "/remote/dir", "ro")
  .start();

Creating a container with a tmpfs mount:

const { GenericContainer } = require("testcontainers");

const container = await new GenericContainer("postgres")
  .withExposedPorts(5432)
  .withTmpFs({ "/temp_pgdata": "rw,noexec,nosuid,size=65536k" })
  .start();

Copy a file to a container before it is started:

const { GenericContainer } = require("testcontainers");

const container = await new GenericContainer("postgres")
  .withExposedPorts(5432)
  .withCopyFileToContainer("/local/file.txt", "/remote/file1.txt")
  .withCopyContentToContainer("hello world", "/remote/file2.txt")
  .start();

Creating a container with environment variables:

const { GenericContainer } = require("testcontainers");

const container = await new GenericContainer("alpine")
  .withEnv("ENV_KEY", "ENV_VALUE")
  .start();

Creating a container with labels:

const { GenericContainer } = require("testcontainers");

const container = await new GenericContainer("alpine")
  .withLabels({
    label1: "value1",
    label2: "value2",
  })
  .start();

Creating a container with a custom health check command. Note that interval, timeout, retries and startPeriod are optional; the values will be inherited from the image or parent image if omitted. Also note that the wait strategy should be set to Wait.forHealthCheck() for this option to take effect:

const { GenericContainer, Wait } = require("testcontainers");

const container = await new GenericContainer("alpine")
  .withHealthCheck({
    test: "curl -f http://localhost || exit 1",
    interval: 1000,
    timeout: 3000,
    retries: 5,
    startPeriod: 1000
  })
  .withWaitStrategy(Wait.forHealthCheck())
  .start();

Creating a container that connects to a specific network:

const { GenericContainer } = require("testcontainers");

const container = await new GenericContainer("alpine")
  .withNetworkMode("network_name")
  .start();

Create user-defined bridge network and attach the container to it:

const { GenericContainer, Network } = require("testcontainers");

const network = await new Network()
    .start();

const container = await new GenericContainer("alpine")
  .withNetworkMode(network.getName())
  .start();

await container.stop();
await network.stop();

Communicate to containers on the same network via aliases:

const { GenericContainer, Network } = require("testcontainers");

const network = await new Network()
    .start();

const container = await new GenericContainer("alpine")
  .withCmd(["sleep", "infinity"])
  .withNetworkMode(network.getName())
  .start();

const fooContainer = await new GenericContainer("alpine")
  .withCmd(["sleep", "infinity"])
  .withNetworkMode(network.getName())
  .withNetworkAliases("foo", "bar")
  .start();

expect((await container.exec(["nslookup", "foo"])).exitCode).toBe(0);

Add hostname mappings:

const { GenericContainer } = require("testcontainers");

const container = await new GenericContainer("alpine")
  .withExtraHosts(
    { host: "foo", ipAddress: "10.11.12.13" },
    { host: "bar", ipAddress: "11.12.13.14" }
  )
  .start();

expect((await container.exec(["nslookup", "foo"])).exitCode).toBe(0);
expect((await container.exec(["nslookup", "bar"])).exitCode).toBe(0);

Specifying a pull policy. Note that if omitted will use the DefaultPullPolicy which will use a locally cached image if one already exists, this is usually the preferred option. In cases where there is a local image for a given tag but the remote image with the same tag may have changed (for example when using the latest tag), you can tell testcontainers to pull the image again by specifying an AlwaysPullPolicy:

const { GenericContainer, AlwaysPullPolicy } = require("testcontainers");

const container = await new GenericContainer("alpine:latest")
  .withPullPolicy(new AlwaysPullPolicy())
  .start();

Same for images in a Dockerfile:

const { GenericContainer, AlwaysPullPolicy } = require("testcontainers");

const container = await GenericContainer.fromDockerfile(buildContext)
  .withPullPolicy(new AlwaysPullPolicy())
  .build();

Specifying a default log driver. You can override the logging driver used by Docker to be the default one (json-file). This might be necessary when the driver of your docker host does not support reading logs and you want to use the Wait.forLogMessage wait strategy. This is the same as --log-driver json-file on docker run.

const { GenericContainer } = require("testcontainers");

const container = await new GenericContainer("redis")
  .withDefaultLogDriver()
  .start();

Creating and running a container with a specific user, note that the value can be a username or UID (format: <name|uid>[:<group|gid>]):

const { GenericContainer } = require("testcontainers");

const container = await new GenericContainer("alpine")
  .withUser("bob")
  .start();

Creating a container with privileged mode:

const { GenericContainer } = require("testcontainers");

const container = await new GenericContainer("alpine")
  .withPrivilegedMode()
  .start();

Creating a container with IPC mode:

const { GenericContainer } = require("testcontainers");

const container = await new GenericContainer("alpine")
  .withIpcMode("host")
  .start();

Testcontainers will not wait for a container to stop, to override:

const { GenericContainer } = require("testcontainers");

const container = await new GenericContainer("postgres")
  .withExposedPorts(5432)
  .start();

await container.stop({
  timeout: 10000
});

Testcontainers will remove associated volumes created by the container when stopped, to override:

const { GenericContainer } = require("testcontainers");

const container = await new GenericContainer("postgres")
  .withExposedPorts(5432)
  .start();

await container.stop({
  removeVolumes: false
});

Finding a container's IP address in a given network:

const { GenericContainer, Network } = require("testcontainers");

const network = await new Network()
  .start();

const container = await new GenericContainer("alpine")
  .withNetworkMode(network.getName())
  .start();

const networkIpAddress = container.getIpAddress(network.getName());

Exposing host ports to the container:

const { GenericContainer, TestContainers } = require("testcontainers");
const { createServer } = require("http");

const server = createServer((req, res) => {
  res.writeHead(200);
  res.end("hello world");
});
server.listen(8000);

await TestContainers.exposeHostPorts(8000);

const container = await new GenericContainer("alpine")
  .withCmd(["sleep", "infinity"])
  .start();

const { output } = await container.exec(["curl", `http://host.testcontainers.internal:8000`]);
assert(output === "hello world");

Specifying an exact host port to bind to (not recommended, Testcontainers will automatically find an available port):

const { GenericContainer } = require("testcontainers");

const container = await new GenericContainer("alpine")
  .withExposedPorts({
    container: 8000,
    host: 8080
  })
  .start();

Docker Compose

Testcontainers supports docker-compose. For example for the following docker-compose.yml:

version: "3"

services:
  redis:
    image: redis:latest
    ports:
      - 6379
  postgres:
    image: postgres:latest
    ports:
      - 5432

You can start and stop the environment, and interact with its containers:

const path = require("path");
const redis = require("async-redis");
const { DockerComposeEnvironment } = require("testcontainers");

describe("DockerComposeEnvironment", () => {
  let environment;
  let redisClient;

  beforeAll(async () => {
    const composeFilePath = path.resolve(__dirname, "dir-containing-docker-compose-yml");
    const composeFile = "docker-compose.yml";

    environment = await new DockerComposeEnvironment(composeFilePath, composeFile).up();

    const redisContainer = environment.getContainer("redis_1");
    redisClient = redis.createClient(
      redisContainer.getMappedPort(6379),
      redisContainer.getHost(),
    );
  });

  afterAll(async () => {
    await redisClient.quit();
    await environment.down();
  });

  it("works", async () => {
    await redisClient.set("key", "val");
    expect(await redisClient.get("key")).toBe("val");
  });
});

You can supply a list of service names to only start those services:

const { DockerComposeEnvironment } = require("testcontainers");

const environment = await new DockerComposeEnvironment(composeFilePath, composeFile)
  .up(["database_service", "queue_service"]);

Create the containers with their own wait strategies:

const { DockerComposeEnvironment, Wait } = require("testcontainers");

const environment = await new DockerComposeEnvironment(composeFilePath, composeFile)
  .withWaitStrategy("redis_1", Wait.forLogMessage("Ready to accept connections"))
  .withWaitStrategy("postgres_1", Wait.forHealthCheck())
  .up();

Once the environment has started, you can interact with the containers as you would any other GenericContainer:

const { DockerComposeEnvironment } = require("testcontainers");

const environment = await new DockerComposeEnvironment(composeFilePath, composeFile).up();

const container = environment.getContainer("alpine_1");
const { output, exitCode } = await container.exec(["echo", "hello", "world"]);

You can supply multiple compose files to support overriding:

const { DockerComposeEnvironment } = require("testcontainers");

const environment = await new DockerComposeEnvironment(composeFilePath, [composeFile1, composeFile2]).up();

await environment.stop();

If you have multiple docker-compose environments which share dependencies such as networks, you can stop the environment instead of downing it:

const { DockerComposeEnvironment } = require("testcontainers");

const environment = await new DockerComposeEnvironment(composeFilePath, composeFile).up();

await environment.stop();

By default, docker-compose does not re-build Dockerfiles, but you can override this behaviour:

const { DockerComposeEnvironment } = require("testcontainers");

const environment = await new DockerComposeEnvironment(composeFilePath, composeFile)
  .withBuild()
  .up();

Specify the environment file:

const { DockerComposeEnvironment } = require("testcontainers");

const environment = await new DockerComposeEnvironment(composeFilePath, composeFile)
    .withEnvFile(".env.custom")
    .up();

Specify profiles:

const { DockerComposeEnvironment } = require("testcontainers");

const environment = await new DockerComposeEnvironment(composeFilePath, composeFile)
    .withProfiles("profile1", "profile2")
    .up();

Specify not to re-create containers that are already running:

const { DockerComposeEnvironment } = require("testcontainers");

const environment = await new DockerComposeEnvironment(composeFilePath, composeFile)
  .withNoRecreate()
  .up();

Bind environment variables to the docker-compose file. For example if we have the following docker-compose file:

services:
  redis:
    image: redis:${TAG}

Then we can set TAG as follows:

const { DockerComposeEnvironment } = require("testcontainers");

const environment = await new DockerComposeEnvironment(composeFilePath, composeFile)
  .withEnv("TAG", "value")
  .up();

Wait Strategies

Ordinarily Testcontainers will wait for up to 60 seconds for the container's mapped network ports to start listening. If the default 60s timeout is not sufficient, it can be altered with the withStartupTimeout() method:

const { GenericContainer } = require("testcontainers");

const container = await new GenericContainer("redis")
  .withExposedPorts(6379)
  .withStartupTimeout(120000)
  .start();

Log output

Plain text:

const { GenericContainer, Wait } = require("testcontainers");

const container = await new GenericContainer("redis")
  .withExposedPorts(6379)
  .withWaitStrategy(Wait.forLogMessage("Ready to accept connections"))
  .start();

Regular expression:

const { GenericContainer, Wait } = require("testcontainers");

const container = await new GenericContainer("redis")
  .withExposedPorts(6379)
  .withWaitStrategy(Wait.forLogMessage(/Listening on port [0-9]+/))
  .start();

Health check

const { GenericContainer, Wait } = require("testcontainers");

const container = await new GenericContainer("redis")
  .withExposedPorts(6379)
  .withWaitStrategy(Wait.forHealthCheck())
  .start();

Authentication

Testcontainers will automatically pick up and use credentials from $HOME/.docker/config.json, using credential helpers, credential stores, or raw auth as necessary and in that order.

Auxiliary Containers

Testcontainers may need to create auxiliary containers to provide its functionality.

To avoid Docker pull limits, you can host your own images and use them by setting the appropriate environment variables:

| Container | Environment Variable | Default | --- | --- | --- | | ryuk | RYUK_CONTAINER_IMAGE | testcontainers/ryuk:0.3.2 | | SSHd | SSHD_CONTAINER_IMAGE | testcontainers/sshd:1.0.0 |

ryuk

Testcontainers will start ryuk whenever a container, docker-compose environment or network is started.

Once started, this container keeps track of containers/images/networks/volumes created by testcontainers and will automatically clean them up 10s after connectivity with testcontainers is lost. This is useful for example if a test starts a container and then terminates unexpectedly, as it will be automatically removed.

ryuk by default does not run privileged, if necessary this can be overridden by setting the environment variable TESTCONTAINERS_RYUK_PRIVILEGED to true. If necessary, ryuk can be disabled enirely by setting the environment variable TESTCONTAINERS_RYUK_DISABLED to true.

SSHd

Testcontainers will start SSHd when using the expose host port functionality.

Once started, any container that is created will have a host mapping of host.testcontainers.internal that points to the SSHd container, as well as being connected to its network.

When we then expose a host port, we remote port forward our local port to the SSHd container, which the other containers will then be able to access at host.testcontainers.internal:<exposed-port>.

Common Issues

1. Insufficient Docker memory

By default, Docker sets CPU and memory limits, with a default memory limit of 2GB. If exceeded, you will be unable to pull/run Docker images. To see how much memory Docker has used, you can run docker system info

• To remove existing containers and images to clear some space you can run docker system prune
• Alternatively you can increase the memory limit via Docker's settings under the Advanced pane.

2. Insufficient test timeouts

It can take a few seconds up to a few minutes to pull and run certain Docker images, depending on file sizes and network constraints. It's unlikely that the default timeouts set by test frameworks are sufficient.

• Increase the test timeout via the methods provided by the testing framework.