Kubernetes Compute

The Kubernetes integration have to be configured by CDS administrator.

This integration allows you to run the Kubernetes Hatchery to start CDS Workers.

As an end-users, this integration allows:

to use Worker Models of type “Docker”
to use Service Prerequisite on your CDS Jobs.

Start Kubernetes hatchery

Generate a token:

$ cdsctl consumer new me \
--scopes=Hatchery,RunExecution,Service,WorkerModel \
--name="hatchery.kubernetes" \
--description="Consumer token for kubernetes hatchery" \
--groups="" \
--no-interactive

Builtin consumer successfully created, use the following token to sign in:
xxxxxxxx.xxxxxxx.4Bd9XJMIWrfe8Lwb-Au68TKUqflPorY2Fmcuw5vIoUs5gQyCLuxxxxxxxxxxxxxx

Edit the section hatchery.kubernetes in the CDS Configuration file. The token have to be set on the key hatchery.kubernetes.commonConfiguration.api.http.token.

Then start hatchery:

engine start hatchery:kubernetes --config config.toml

This hatchery will spawn Pods on Kubernetes in the default namespace or the specified namespace in your config.toml. Each pods is a CDS Worker, using the Worker Model of type ‘docker’.