Operations runbooks

Synced read-only from /home/synnet/mirrors/synvirt-product/docs/operations/README.md. Edit at the source, not here.

Operations runbooks

Deploy procedures, recovery flows, and migration guides for production SynVirt hosts. Each runbook is a step-by-step operator-facing document with binary success criteria, rollback procedures, and the exact commands to copy-paste.

Distinct from docs/modules/<X>/ (engineering-facing per-module deep-dives) and docs/architecture/ (cross-cutting design docs).

Catalog

Runbook	Scope
refactor-r1-network.md	6-step migration of an existing host onto synvirt-network.service (R1)

When to write a new runbook

A new runbook lands here when:

1. A refactor changes the boot order, the systemd unit set, or the on-disk layout in a way that needs operator action on already-installed hosts.
2. A recovery path requires operator intervention beyond what the dashboard surfaces (e.g. console-mode vi edit of a YAML source of truth).
3. A deploy procedure has more than three sequential steps where wrong order breaks state.

For one-off migrations or incident-specific recoveries, use docs/incidents/ instead (per-incident postmortems).

Convention

Every runbook starts with a “Why this exists” section, lists locked decisions before any commands, and ends with explicit acceptance criteria the operator can paste into a smoke check. Rollback is documented inline, not as an afterthought.