Quickstart
Testcontainers for Go plays well with the native go test
framework.
The ideal use case is for integration or end to end tests. It helps you to spin up and manage the dependencies life cycle via Docker.
1. System requirements¶
Please read the system requirements page before you start.
2. Install Testcontainers for Go¶
We use gomod and you can get it installed via:
go get github.com/testcontainers/testcontainers-go
3. Spin up Redis¶
import (
"context"
"github.com/testcontainers/testcontainers-go"
"github.com/testcontainers/testcontainers-go/wait"
)
func TestWithRedis(t *testing.T) {
ctx := context.Background()
req := testcontainers.ContainerRequest{
Image: "redis:latest",
ExposedPorts: []string{"6379/tcp"},
WaitingFor: wait.ForLog("Ready to accept connections"),
}
redisC, err := testcontainers.GenericContainer(ctx, testcontainers.GenericContainerRequest{
ContainerRequest: req,
Started: true,
})
if err != nil {
log.Fatalf("Could not start redis: %s", err)
}
defer func() {
if err := redisC.Terminate(ctx); err != nil {
log.Fatalf("Could not stop redis: %s", err)
}
}()
}
The testcontainers.ContainerRequest
describes how the Docker container will
look.
Image
is the Docker image the container starts from.ExposedPorts
lists the ports to be exposed from the container.WaitingFor
is a field you can use to validate when a container is ready. It is important to get this set because it helps to know when the container is ready to receive any traffic. In this case, we check for the logs we know come from Redis, telling us that it is ready to accept requests.
When you use ExposedPorts
you have to imagine yourself using docker run -p
<port>
. When you do so, dockerd
maps the selected <port>
from inside the
container to a random one available on your host.
In the previous example, we expose 6379
for tcp
traffic to the outside. This
allows Redis to be reachable from your code that runs outside the container, but
it also makes parallelization possible because if you add t.Parallel
to your
tests, and each of them starts a Redis container each of them will be exposed on a
different random port.
testcontainers.GenericContainer
creates the container. In this example we are
using Started: true
. It means that the container function will wait for the
container to be up and running. If you set the Start
value to false
it won't
start, leaving to you the decision about when to start it.
All the containers must be removed at some point, otherwise they will run until
the host is overloaded. One of the ways we have to clean up is by deferring the
terminated function: defer redisC.Terminate(ctx)
.
Tip
Look at features/garbage_collector to know another way to clean up resources.
4. Make your code to talk with the container¶
This is just an example, but usually Go applications that rely on Redis are using the redis-go client. This code gets the endpoint from the container we just started, and it configures the client.
endpoint, err := redisC.Endpoint(ctx, "")
if err != nil {
t.Error(err)
}
client := redis.NewClient(&redis.Options{
Addr: endpoint,
})
_ = client
We expose only one port, so the Endpoint
does not need a second argument set.
Tip
If you expose more than one port you can specify the one you need as a second argument.
In this case it returns: localhost:<mappedportfor-6379>
.
5. Run the test¶
You can run the test via go test ./...
6. Want to go deeper with Redis?¶
You can find a more elaborated Redis example in our examples section. Please check it out here.