Introduction

Jenesis Repository is a self-hosted artifact repository with no database. One server hosts your own packages and caches the public ones for more than twenty package ecosystems - Java, npm, PyPI, container images, Go, Cargo, NuGet, RubyGems, Helm, Debian and RPM packages, and more - each served to its own native client. It screens what comes in against vulnerability and malware feeds, holds back what fails your policy for review, and runs a web console to look after it all. If you build with the Jenesis build tool, it can also keep a remote build cache for it.

You run it from one Docker image, jenesisbuild/jenesis-repository, and everything it holds lives in one place: a directory on a volume, or a bucket on S3, Google Cloud Storage or Azure Blob.

Three things to know up front

• The store is the only state. Artifacts, indexes, settings and keys all live in the one directory or bucket. Back it up and you have backed up the repository; copy it and you have moved it; replicate it, and a server started over the replica is a fail-over, or one more server to read from. There is no database to install, tune or migrate.
• Artifacts stream through, never into memory. An upload or a download is copied between the network and the store without being held whole, so a small POM and a multi-gigabyte image layer cost the server the same small, fixed amount of memory.
• Everything is a module. Formats, storage backends, screening feeds and console pages are all modules behind a service interface. A feature of your own is one more module beside them - added without a fork to maintain, and without waiting on anyone else to change an integration for you. A feature you do not want is simply left out. Compose a smaller server shows how.

How this section is organised

The chapters follow the console. After the first two, each covers one of its sections, in the order they appear across the top of the page, and the command line follows them:

1. Getting started - run the image, sign in, issue a key, and publish with Maven and npm.
2. Finding your way around - the console's two navigation levels, the pages of a repository, and who sees what.
3. Repositories - what a repository is here, creating one with its settings, and its overview, browse, staging and import pages.
4. Connecting your build tools - the URL and credential form for every client.
5. Proxying upstreams - serving Maven Central, Docker Hub and other registries through your repository.
6. Screening what comes in - the gate, the review queue, and the vulnerability, findings and signer pages.
7. Retention, pins and cleanup - how long a repository keeps what it holds, and reclaiming space.
8. The build cache - projects, and pointing the Jenesis build tool at them.
9. Access - signing people in, members and roles, credentials for build tools, and keyless CI.
10. Operations - metrics, the security posture, scheduled walks, manual uploads, webhooks and rate limits.
11. Settings - the levels a setting is stored at, the first-run wizard, the settings catalogue, modules, and backing settings up.
12. The command line - the jenrepo command, for a terminal, a script or a program.
13. Running in production - object storage, the Helm chart, TLS and backups.
14. Migrating in and out - importing from Nexus, Artifactory and others, and taking everything out again.
15. What it costs to run - where an object store's bill comes from, and how to keep it small.
16. Running from source - for those who want to change the server itself.
17. Configuration reference - every setting in one place.