Hatchery Local

Use case

If you are a CDS user, you can use shared.infra Worker to run your job.

But, if you want to use your host for various good reasons as OS, Network, etc… you can launch CDS Worker on your host.

Two prerequisites:

your host must reach your CDS API on HTTP port defined on your CDS Configuration
the basedir (default is /var/lib/cds-engine) must allow to execute binaries.

The worker binary is downloaded from CDS API at the start of the hatchery, it’s placed into basedir directory.

Start Local hatchery

Generate a token:

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

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

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

Then start hatchery:

engine start hatchery:local --config config.toml

This hatchery will now start worker binary on your host. You can manage settings, as max workers in the hatchery configuration file.