opschain state doctor

Included from Basic tier

Classifies every database property owner in the local store against the repository's current config.yaml, as a table (OWNER VERSION STATUS):

• declared — the owner node exists in config.yaml.
• ORPHANED (not in config.yaml) — the owner exists in the store but not in the current config.yaml (a renamed or removed node).
• other project (per-user store) — the owner belongs to another project sharing this per-user store.

Every run then names the store file and the keyset in force — store: <opschain home>/local/opschain.db, then keyset: <opschain home>/local/keys.env (present, mode 0600), or (present, mode <mode> — tighten to 0600) when the file is readable more widely than the generated posture, or (missing — generated on the next local write; values encrypted with a previous keyset are unreadable) — and ends with the scoping note the store operates under: owners are keyed by project path in the per-user store: two repositories declaring the same project code share these layers. -o json|yaml emits the full result document (project path, owners, removals, store with path, keyset with path/present/mode, note).

With --gc the orphaned owners' layers and history are deleted, reported as removed <n> orphaned owner(s): <paths>.

This verb is repo-scoped: against a server target it fails with inspecting state is repo-scoped and runs locally — it has no remote form; run it in the repository (or pass --local).

Flags

Flag	Meaning
--gc	Delete the orphaned owners' layers and history.

Exit codes

Code	Meaning
0	Report printed (and orphans removed, with --gc).
1	The store read, config.yaml read or GC failed.
2	Usage error.

Examples

opschain state doctor
opschain state doctor --gc

See also

• opschain docs cli/state-reset — the project-granularity clobber, and how a reset treats the keyset.
• opschain docs storage — "Local mode": the store's location beside its keyset, and the keyset's rules.
• opschain learn 1.3 — properties and settings, including DB property layers.