kblaunch 0.1.0

A CLI tool for launching Kubernetes jobs in EIDF

kblaunch

A CLI tool for launching Kubernetes jobs with environment variable and secret management.

Installation

pip install git+https://github.com/gautierdag/kblaunch.git

Usage

Basic Usage

Launch a simple job:

kblaunch \
    --email [e-mail] \
    --job-name myjob \
    --command "python script.py"

With Environment Variables

1. From local environment:

   kblaunch \
       --job-name myjob \
       --command "python script.py" \
       --local-env-vars PATH \
       --local-env-vars PYTHONPATH
   

2. From Kubernetes secrets:

   kblaunch \
       --job-name myjob \
       --command "python script.py" \
       --secrets-env-vars mysecret1 \
       --secrets-env-vars mysecret2
   

3. From .env file:

   kblaunch \
       --job-name myjob \
       --command "python script.py" \
       --load-dotenv
   

GPU Jobs

Specify GPU requirements:

kblaunch \
    --job-name gpu-job \
    --command "python train.py" \
    --gpu-limit 2 \
    --gpu-product "NVIDIA-A100-80GB-PCIe"

Interactive Mode

Launch an interactive job:

kblaunch \
    --job-name interactive \
    --interactive

Options

• --email: User email [required]
• --job-name: Name of the Kubernetes job [required]
• --docker-image: Docker image (default: "nvcr.io/nvidia/cuda:12.0.0-devel-ubuntu22.04")
• --namespace: Kubernetes namespace (default: "informatics")
• --queue-name: Kueue queue name
• --interactive: Run in interactive mode (default: False)
• --command: Command to run in the container [required]
• --cpu-request: CPU request (default: "1")
• --ram-request: RAM request (default: "8Gi")
• --gpu-limit: GPU limit (default: 1)
• --gpu-product: GPU product (default: "NVIDIA-A100-80GB-PCIe")
• --secrets-env-vars: List of secret environment variables
• --local-env-vars: List of local environment variables
• --load-dotenv: Load environment variables from .env file (default: True)

Features

• Kubernetes job management
• Environment variable handling from multiple sources
• Kubernetes secrets integration
• GPU job support
• Interactive mode
• Automatic job cleanup
• Slack notifications (when configured)