Using SatL

Containers, services, images, networks, volumes and secrets, day to day.

• Containers and services — the model everything else follows from: every container is a task of a service, and what that changes about start, rm and ps.
• Images — pulling, platform selection, registries and authentication, and where the bytes land as ZFS datasets.
• Linux containers — linux/amd64 images as jails under the linuxulator: what to load on the host, what the emulation gives you, and exactly where it stops.
• Networks — the node bridge, overlays, addresses and container DNS.
• Publishing ports — host mode, ingress mode, and why curl localhost on the publishing host never works.
• Volumes, binds and tmpfs — what persists, what does not, and what a volume costs.
• Secrets and configs — where a payload goes, what it may be, and why rotation is by replacement.
• Resource limits — --memory and --cpus as rctl(8) rules, and the boot-time tunable they need.
• Healthchecks — Docker's semantics, plus the two things health does here that it does not there.
• Rolling updates — the twelve policy flags, the rule that decides what yours did, and automatic rollback.
• Compose files — satl compose up with stack semantics, the supported subset, and what it refuses rather than ignores.
• Reclaiming space — satl system prune, which is manual and node-local, and why a layer sometimes survives the first run.