There are three ways to run the Scribe services, and every way grows the same deployment in the same order: one storage server, then shared authentication, then a gateway in front of more storage servers. The web interface runs beside it in Docker.
| Way | Start here | What you run |
|---|---|---|
With scribe-ctl, natively or in Docker | Tutorials 1–5, then the cluster guide | One definition file; scribe-ctl creates, starts, renews, upgrades and diagnoses everything, on this machine or another over ssh |
| By hand, on bare metal | Bare metal | scribe-server, scribe-auth and scribe-gateway under your own service manager |
| By hand, with Docker Compose | Docker Compose | A Compose project you write, from the images in docker/ |
The manual guides are also the reference for what scribe-ctl does for you:
each tutorial stage names the files it creates, and the matching manual stage
creates the same files with the executables.
| Task | With scribe-ctl | By hand |
|---|---|---|
| One storage server | Tutorial 1 | Bare metal §1, Docker §1 |
| Add authentication, people and grants | Tutorial 2 | Bare metal §2, Docker §2 |
| Add a gateway and more storage servers | Tutorial 3 | Bare metal §3–4, Docker §3–4 |
| Web interface | Tutorial 5 | Web interface |
| Renew credentials and certificates, rotate CAs | Tutorial 4 §5 | Certificates and credentials |
| Upgrade to a new release | Tutorial 4 §6 | Upgrades |
| Back up and restore | Backups | Backups |
| A team's first connection, login and trust | Client trust and login | the same |
Before a team uses any of them, go through the production checklist.