Boilerplate for continuous deployment in Kubernetes
The main 2 tools you need to use is gcloud to manage GCP services and kubectl which
you need to manage the kubernetes cluster.
Most of the common operations are covered by the ./cli
Follow GCP instructions https://cloud.google.com/sdk/install
gcloud components install kubectl
Install instructions follow https://docs.docker.com/engine/install/
Install instructions follow https://docs.docker.com/compose/install/
For a simpler configuration create a file in your user directory called .kube-test.rc with
the following variables setup.
K8S_NAMESPACES=dev stag prod
K8S_DEFAULT_NAMESPACE=dev
GCP_HOSTNAME=gcr.io
GCP_PROJ_ID=<GCP Project id>
GCP_LOCATION=<gcp location>
GCP_CA_JSON=<GCP service token>
GCP_USER=<GCP service user>
GCP_KLUSTER_NAME=<kubernetes-test>
- K8S_NAMESPACES List of desired namespaces
- K8S_DEFAULT_NAMESPACE Default namespace
- GCP_HOSTNAME
- GCP_PROJ_ID
- GCP_LOCATION
- GCP_USER GCP service user e.g. gcp-sa
- GCP_KLUSTER_NAME GCP Kubernetes kluster name
- GCP_CA_JSON Base64 encoded GCP Json certificate for service user
After the steps, simply check the installation with ./cli check
./cli help
./cli deploy Will deploy to K8S_DEFAULT_NAMESPACE
./cli deploy stag Will deploy to stag namespace
...
All folders prefixed by app_ are considered Kubernetes Apps, which for an easier management forces to have a particular structure.
app_xxx
|_> build: all dockerfiles
|_> deploy: all kubernetes configuration files, including databases, services, etc
...
Running kubectl get all and getting this error:
The connection to the server localhost:8080 was refused - did you specify the right host or port?
Use gcloud container clusters list to get the list of clusters available at your GCP account
And then configure kubectl with the desired cluster credentials
gcloud container clusters get-credentials <CLUSTER NAME> --zone <ZONE>