racetrack-client 2.1.2

CLI client tool for deploying workloads to Racetrack

Racetrack client context

racetrack-client is a CLI client tool for deploying workloads to Racetrack (IKP-RT).

Racetrack system allows to deploy jobs in a one step. It transforms your code to in-operation workloads, e.g. Kubernetes workloads. You write some code according to a set of coventions, you include the manifest file which explains the code, and you submit it to Racetrack. A short while after, the service calling your code is in operation.

Quickstart

1. Install racetrack client: pip3 install racetrack-client
2. Configure access token to your git repository: racetrack config credentials set REPO_URL USERNAME TOKEN
3. Deploy your job to Racetrack: racetrack deploy . https://ikp-rt.platform.example.com/lifecycle
4. You will see the URL of your deployed job in the output.

Installation

Install racetrack-client using pip:

pip3 install racetrack-client

Python 3.8+ is required. So if you have any troubles, try with:

sudo apt install python3.8 python3.8-dev python3.8-venv
python3.8 -m pip install racetrack-client

This will install racetrack CLI tool. Verify installation by running racetrack.

Usage

Run racetrack --help to see usage.

Deploying

To deploy a job, just run in the place where fatman.yaml is located:

racetrack deploy . https://ikp-rt.platform.example.com/lifecycle

racetrack deploy [WORKDIR] [RACETRACK_URL] has 2 optional arguments:

• WORKDIR - a place where the fatman.yaml is, by default it's current directory
• RACETRACK_URL - URL address to Racetrack server, where a job should be deployed. If not given, it will be deployed to a URL configured in a local client config, by default it's set to a local cluster at http://localhost:7002.