ReadMeFeed / Community knowledge
Real questions. Useful conversations. Find the people who know your stack.
Developer tools · #harbor
Learn how to configure Harbor's built‑in image replication to create a hot standby registry, with a step‑by‑step example, trade‑offs, and verification steps.
A minimal Harbor lifecycle uses per-project retention policies, immutable tag rules for releases, scheduled garbage collection, and scoped robot accounts to bound storage growth while protecting release artifacts.
When CI/CD or Kubernetes pulls hit Docker Hub rate limits, a Harbor proxy cache can solve the problem. This guide walks through creating a cache project, pulling images, and verifying the cache works.
Learn how to enable Harbor’s integrated vulnerability scanner, configure it for automatic scans on image push, and tune performance versus depth with a concrete example and verification steps.
Harbor can front Docker Hub (or other OCI registries) as a pull-through cache, reducing rate-limit hits and adding centralized scanning and retention--but stale 'latest' tags and version-specific behavior require careful handling.
Harbor's pull-based replication lets regional registries fetch images from a central Harbor on a schedule, so CI publishes once and every cluster pulls locally. Here's how the policies work and where the pitfalls are.
Proxy cache and replication both make upstream images available in Harbor, but they fail differently. Compare constraints, storage, and validation steps before choosing.
Set up Harbor v2.8+ replication between two registries for cross-region disaster recovery: robot accounts, a target endpoint, a scheduled tag-filtered policy, digest checks and a promotion checklist.
Enable Harbor's built-in Trivy scanner by editing harbor.yml, restart services, push a test image, and verify vulnerability reports in the UI or via API. Covers offline DB updates, concurrency limits, and common misconfigurations.