core+cli: castle apply — one converge verb replaces deploy/start/enable/…

Reconcile the running system to declared config in a single operation,
replacing the old deploy && start plus the per-kind enable/disable/
install/uninstall verbs. The mechanism varies by manager (systemd unit /
PATH install / gateway route); the verb never does.

Core (castle_core.deploy.apply):
- render (units, Caddyfile, tunnel) then reconcile: activate every
  enabled deployment that's down, restart any whose unit bytes changed,
  deactivate the disabled ones. Returns ApplyResult (the enacted diff).
- --plan computes the diff and touches nothing.
- restart-on-change is exact: _render_unit_files is the single source of
  truth for unit bytes, used both to write units and to predict restarts,
  so the plan can't drift from what deploy writes.

Desired state:
- DeploymentBase.enabled (default True) — the declarative on/off. apply
  converges to it; a disabled deployment is defined but not running and
  gets no gateway route (else it would 502). Registry omits enabled when
  True, keeping existing registries byte-identical.

CLI:
- add `castle apply [name] [--plan]`; keep `restart` as the one
  imperative bounce, plus status/doctor/logs/gateway/program.
- retire deploy, start, stop (top-level); service/job deploy|enable|
  disable|start|stop; tool install|uninstall; program install|uninstall.
- next-step strings + doctor hints now say `castle apply`.

Verified: core 129 + cli 31 green; live `castle apply` on a converged
node reports "nothing to do" with no spurious restarts; --plan shows an
empty diff; gateway + api stay up.
This commit is contained in:
2026-07-02 11:35:34 -07:00
parent 9028d15ec6
commit d0a206f7d6
18 changed files with 406 additions and 240 deletions

View File

@@ -6,7 +6,7 @@ of checks grouped into Environment, Configuration, Runtime, and TLS & exposure;
each check reports ok / warn / fail with a one-line hint when action is needed.
Exit code: 0 when nothing FAILed (warnings are allowed), 1 otherwise — so it
doubles as a scriptable smoke test after `./install.sh` or `castle deploy`.
doubles as a scriptable smoke test after `./install.sh` or `castle apply`.
"""
from __future__ import annotations
@@ -206,7 +206,7 @@ def _check_runtime(config) -> list[Check]:
Check(
FAIL,
"gateway not running",
hint="castle deploy && castle start",
hint="castle apply",
)
)
@@ -227,7 +227,7 @@ def _check_runtime(config) -> list[Check]:
checks.append(Check(OK, "castle-api running", detail=detail))
else:
checks.append(
Check(FAIL, "castle-api not running", hint="castle deploy && castle start")
Check(FAIL, "castle-api not running", hint="castle apply")
)
# Generated artifacts.
@@ -244,7 +244,7 @@ def _check_runtime(config) -> list[Check]:
FAIL,
"generated specs missing",
detail=", ".join(missing),
hint="castle deploy",
hint="castle apply",
)
)
@@ -347,7 +347,7 @@ def _check_tls_exposure(config) -> list[Check]:
WARN,
"castle-tunnel not running",
detail="public routes are down",
hint="castle service start castle-tunnel",
hint="enable castle-tunnel in its deployment, then: castle apply",
)
)
checks.append(_check_public_dns(config))
@@ -364,7 +364,7 @@ def _check_public_dns(config) -> Check:
records. The required permission is a single DNS:Edit (Cloudflare's 'Edit zone
DNS' template), which also grants the zone lookup — so a correctly-scoped token
passes this probe. Write itself isn't exercised (that would mutate); a
DNS:Read-only token would false-pass here but `castle deploy` then surfaces a
DNS:Read-only token would false-pass here but `castle apply` then surfaces a
403 with the fix. Absent token → WARN (CNAMEs stay manual), not a failure.
"""
import urllib.error