lrz-docker-enroot-refresh

Build, push, and safely swap LRZ Enroot container images on the cluster.

Updated Oct 23, 2025
One-click install
npx skills add https://github.com/felixmin/robot-learning-from-video --skill lrz-docker-enroot-refresh
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: lrz-docker-enroot-refresh
Source: https://github.com/felixmin/robot-learning-from-video/tree/main/.codex/skills/lrz-docker-enroot-refresh
Command: npx skills add https://github.com/felixmin/robot-learning-from-video --skill lrz-docker-enroot-refresh

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill automates the complex process of building, pushing, and deploying updated Docker container images to the LRZ cluster for robot learning workflows, ensuring consistency and reducing manual errors.

Core Features & Use Cases

  • End-to-End Container Management: Handles the entire lifecycle from local build to cluster deployment.
  • Profile-Based Workflows: Supports distinct workflows for stage-1/2 (LAM, foundation models) and stage-3 (LeRobot, policy deployment) containers.
  • Safe Deployment: Implements a safe swapping mechanism for cluster images, minimizing downtime and risk.
  • Use Case: A researcher needs to deploy the latest version of the stage-3 robot policy container to the LRZ cluster. This Skill will build the new Docker image, push it to the registry, import it as an Enroot image on the cluster, and safely replace the currently active image without interrupting ongoing jobs.

Quick Start

Use the lrz-docker-enroot-refresh skill to build and refresh the stage3 container workflow.

Frequently Asked Questions about lrz-docker-enroot-refresh

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I deploy Docker images as Enroot containers on the LRZ cluster?

Deploying Docker images as Enroot containers on the LRZ cluster requires building the image locally, pushing it to a registry, importing it via SSH, and safely swapping the active image. This Skill automates that entire lifecycle for robot learning workflows.

What is the best way to automate container image swaps on Slurm without interrupting jobs?

Automating container image swaps on Slurm without interrupting jobs is achieved through a safe swapping mechanism that builds and pushes the new Docker image, imports it as an Enroot image, and replaces the active image. This minimizes downtime and risk.

Can I use this workflow to manage distinct robot learning stages like foundation models and policy deployment?

Yes, you can manage distinct robot learning stages using profile-based workflows that support stage-1/2 for foundation models and stage-3 for policy deployment. The Skill handles specific container requirements for each stage profile.

Do I need Docker and SSH to build and push LRZ Enroot images?

Yes, you need Docker, SSH, and Slurm job submission capabilities to build and push LRZ Enroot images. These dependencies are required for local image construction, remote cluster communication, and executing the deployment.

Why does my Enroot image deployment fail during the Slurm job submission phase?

Enroot image deployment failures during Slurm job submission often stem from missing prerequisites or environment misconfigurations. This Skill implements robust error handling and backup mechanisms to manage these risks during the image swap.

Does the lrz-docker-enroot-refresh skill support automated backup and error handling for cluster deployments?

Yes, the lrz-docker-enroot-refresh skill supports automated backup and error handling for cluster deployments. It ensures safe deployment by managing the Docker image lifecycle and implementing backup mechanisms during the Enroot image swap.