osdc-pypi-cache

Troubleshoot PyPI wheel cache infrastructure for PyTorch CI runners.

17|14|Updated Jun 15, 2023
One-click install
npx skills add https://github.com/pytorch/ci-infra --skill osdc-pypi-cache
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: osdc-pypi-cache
Source: https://github.com/pytorch/ci-infra/tree/main/osdc/.claude/skills/osdc-pypi-cache
Command: npx skills add https://github.com/pytorch/ci-infra --skill osdc-pypi-cache

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill helps engineers operate and troubleshoot the PyPI wheel cache infrastructure that supplies fast, reliable Python and PyTorch package access to CI runners.

Core Features & Use Cases

  • Architecture Guidance: Explains the per-CUDA cache deployments, shared EFS wheelhouse, nginx caching, pypiserver, wants collector, and wheel syncer.
  • Operational Troubleshooting: Covers pip failures, fallback logs, cache invalidation, S3 synchronization, IPv6 networking, pod placement, resource sizing, and rollout behavior.
  • Infrastructure Maintenance: Guides CUDA slug additions, manifest generation, Kubernetes and Terraform changes, IRSA configuration, NetworkPolicy behavior, and cache-enforcer integration.

Quick Start

Use the osdc-pypi-cache skill to diagnose a PyPI package installation failure for the cu128 cache in a production cluster.

Frequently Asked Questions about osdc-pypi-cache

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

FAQPage Schema
How do I troubleshoot pip package installation failures from a PyPI wheel cache in Kubernetes?

To troubleshoot pip package installation failures from a PyPI wheel cache, inspect fallback logs, verify nginx caching behavior, validate S3 synchronization, and check pod placement and resource sizing within your Kubernetes cluster.

How does a shared EFS wheelhouse work with per-CUDA cache deployments?

A shared EFS wheelhouse provides persistent storage for Python wheels, while per-CUDA cache deployments isolate specific CUDA slugs. nginx and pypiserver route requests, with a wheel syncer and wants collector managing package availability.

Can I use Terraform to configure IRSA and NetworkPolicy for a pypiserver cache?

Yes, you can use Terraform to apply infrastructure changes configuring IRSA and NetworkPolicy behavior for a pypiserver cache, ensuring secure S3 access and controlled network traffic flow across OSDC clusters.

What is the best way to add a new CUDA slug and generate a manifest for a PyPI cache?

The best way to add a new CUDA slug and generate a manifest for a PyPI cache involves updating infrastructure configurations and triggering external wheel-building workflows to populate the shared EFS storage with the correct packages.

Why is PyPI cache invalidation not working over IPv6 networking in my CI runners?

PyPI cache invalidation over IPv6 networking may fail due to misconfigured network policies or pod placement issues. Verify IPv6 routing, check cache-enforcer integration, and ensure S3 synchronization completes successfully.