Getting started with Kubernetes and Terraform

I’ve been looking into how to learn terraform. I have also discovered for my project I need to use Kubernetes. It turns out that it is really easy to create a kubernetes cluster on the local desktop to have a play with. Here goes:

I got started using the following tutorial I used Kubernetes in Docker (kind) to test. This turned out to be really easy to install. I already had Docker installed, so I didn’t need to worry about that. I downloaded kind from it’s website - there is a compiled executable which is easiest. The instructions say on Linux:

1
2
3
curl -Lo ./kind https://kind.sigs.k8s.io/dl/v0.11.1/kind-linux-amd64
chmod +x ./kind
mv ./kind /some-dir-in-your-PATH/kind

It also notes that kubectl will be required to control the cluster:

1
2
3
4
curl -LO "https://dl.k8s.io/release/$(curl -L -s https://dl.k8s.io/release/stable.txt)/bin/linux/amd64/kubectl"
curl -LO "https://dl.k8s.io/$(curl -L -s https://dl.k8s.io/release/stable.txt)/bin/linux/amd64/kubectl.sha256"
echo "$(<kubectl.sha256) kubectl" | sha256sum --check
mv kubectl /some-dir-in-your-PATH/kubectl

Then I followed the tutorial to create a cluster:

1
2
 kind create cluster --name terraform-learn --config kind-config.yaml
 kind get-clusters

Secrets - pulling containers

Once I had finished the tutorial, I wanted to be able to start working on some containers I had made in our local docker repo. This requires me to use a secret. To do this, I needed to add the secret to kubernetes:

1
2
3
4
kubectl create secret docker-registry mysecret \
 --docker-username=myusername \
 --docker-password=mypassword \
 --docker-email=me@example.com

Kubernetes replies:

secret/mysecret created

This is a secret which allows me access to the docker registry. So I can now pull containers from the registry if I set it up properly in terraform:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
resource "kubernetes_deployment" "mydeployment" {

 spec {
   template {
     spec {
       image_pull_secrets {
         name  = "mysecret"
       }
       container {
         image = "registry.example.com/path/to/my/container/image/master:latest"
...

The secret sends the username and password I placed in the secret I created above to the registry to log in to pull the container.

I can look at the secret using the following:

1
2
3
4
$ kubectl get secrets
NAME                  TYPE                                  DATA   AGE
default-token-hzz2g   kubernetes.io/service-account-token   3      13d
mysecret              kubernetes.io/dockerconfigjson        1      22m

This lists the secrets, but that doesn’t show the values. I can get more information by specifying an output format:

1
kubectl get secret mysecret -o json

This replies:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
{
    "apiVersion": "v1",
    "data": {
        ".dockerconfigjson": "eyJhdXRocyI6eyJodHRwczovL2luZGV4LmRvY2tlci5pby92MS8iOnsidXNlcm5hbWUiOiJteXVzZXJuYW1lIiwicGFzc3dvcmQiOiJteXBhc3N3b3JkIiwiZW1haWwiOiJtZUBleGFtcGxlLmNvbSIsImF1dGgiOiJiWGwxYzJWeWJtRnRaVHB0ZVhCaGMzTjNiM0prIn19fQ=="
    },
    "kind": "Secret",
    "metadata": {
        "creationTimestamp": "2021-07-23T10:33:22Z",
        "name": "mysecret",
        "namespace": "default",
        "resourceVersion": "2000896",
        "uid": "7b480748-61e0-4141-9518-6822b861a553"
    },
    "type": "kubernetes.io/dockerconfigjson"
}

The real secret is base64 encoded in .dockerconfigjson. I can decode it using base64d like this:

1
2
3
kubectl get secret mysecret -o json \
  | jq -r '."data".".dockerconfigjson" | @base64d' \
  | jq

Which gives:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
{
  "auths": {
    "https://index.docker.io/v1/": {
      "username": "myusername",
      "password": "mypassword",
      "email": "me@example.com",
      "auth": "bXl1c2VybmFtZTpteXBhc3N3b3Jk"
    }
  }
}

I wondered what auth is, apparently it is just the base64 encoded version of username:password

1
2
3
kubectl get secret mysecret -o json \
  | jq -r '."data".".dockerconfigjson" | @base64d' \
  | jq -r '.auths."https://index.docker.io/v1/".auth | @base64d'

Returns

myusername:mypassword

Secrets - Passing through the Environment

So, now I want to create a database password I can use in the container. These are generic secrets. I can create one like this:

1
2
3
4
kubectl create secret generic dbpass \
    --from-literal=mypass1=p@assw0rd \
    --from-literal=mypass2=s0pers3kr3t
secret/dbpass created

I can look at it as before:

1
kubectl get secret dbpass -o json 

which replies with

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
{
    "apiVersion": "v1",
    "data": {
        "mypass1": "cEBhc3N3MHJk",
        "mypass2": "czBwZXJzM2tyM3Q="
    },
    "kind": "Secret",
    "metadata": {
        "creationTimestamp": "2021-07-23T13:47:49Z",
        "name": "dbpass",
        "namespace": "default",
        "resourceVersion": "2020579",
        "uid": "754a42a2-724a-4daa-9992-0824255ab887"
    },
    "type": "Opaque"
}

For some reason while you specify generic when the password is created, kubernetes says the type is opaque when you read it back. The inconsistent terminology can be a little confusing.

Once again these are base64 encoded:

1
2
kubectl get secret dbpass -o json | jq -r '.data.mypass1 | @base64d'
p@assw0rd

I can pass this into my container as an environment variable as follows:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
resource "kubernetes_deployment" "mydeployment" {
  spec {
    template {
      spec {
        container {
          image = "registry.example.com/path/to/my/container/image/master:latest"
          name  = "mycontainer"
          env_from {
              secret_ref {
                name = "dbpass"
              }
            }
          }

Now if I run a container which has bash, I can connect to it and see the environment variables.

1
2
3
4
5
kubectl get pods
kubectl exec --stdin --tty mycontainer-8588bb5c4f-xr9lm --/bin/bash
env | grep pass
mypass2=s0pers3kr3t
mypass1=p@assw0rd

So the secrets are exposed in the environment. They aren’t base64 encoded which is handy!

Secrets - Mounting

We can modify the terraform above to pass the secrets to a file as follows:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
resource "kubernetes_deployment" "mydeployment" {
  spec {
    template {
      spec {
        volume {
          name   = "passwords"
          secret {
            secret_name = "dbpass"
          }
        }
        container {
          image = "registry.example.com/path/to/my/container/image/master:latest"
          name  = "mycontainer"

          volume_mount {
            mount_path = "/mnt"
            name = "passwords"
            read_only = true
          }

Again, we can see these secrets by connecting to the running container:

1
2
3
4
5
6
7
8
9
$ kubectl get pods
$ kubectl exec --stdin --tty mycontainer-8588bb5c4f-xr9lm --/bin/bash
# cd /mnt
# ls
mypass1  mypass2
# cat mypass1
p@assw0rd
# cat mypass2
s0pers3kr3t

A note about config maps

Kubernets has config maps. These work in the same way, as secrets, but aren’t secret. So instead of passing a dozen environment variables, we can pass a config map that contains them all.

1
2
3
4
kubectl create configmap myconfig \
  --from-literal=PORT=1521 \
  --from-literal=HOST=mydesktop \
  --from-literal=DBNAME=DEVELOPMENT

To share it as an environment variable, we do exactly the same as with the secrets, but replace secret_ref with config_map_ref. Similarly to mount a config map we replace the secret configuration:

1
2
3
4
5
6
        volume {
          name = "passwords"
          secret {
            secret_name = "dbpass"
          }
        }

with the config map configuration

1
2
3
4
5
6
        volume {
          name = "passwords"
          config_map {
            name = "dbpass"
          }
        }

Obviously we wouldn’t be storing passwords in the config map, so the names would change as well, but this demonstrates how similar config maps and secrets are. The volume once defined is mounted in exactly the same way as with the secrets.

Conclusion

So now I have learned how to use Secrets in kubernetes. I can mount them as a filesystem, or I can pass them as environment variables. The same applies to config maps. This is nice as we can have a config map and secret pair for each of for development, staging and production.