refactor(docker): rename OrchestratorService -> DockerInfraService, move to backend/docker

OrchestratorService wasn't the orchestrator — it's the host-side lifecycle of
the docker infra *container* (the one that runs the Orchestrator). It read as
"the orchestrator as a service" and lived in orchestrator/lifecycle.py, while
its siblings (macOS MacosInfraService, Firecracker infra_vm) live under their
backend package. Rename it DockerInfraService and move it to
backend/docker/infra.py alongside the docker backend, with its docker-only
constants (INFRA_*/ORCHESTRATOR_* image + container names, daemon list, mount
paths).

orchestrator/lifecycle.py keeps only the backend-neutral pieces the other infra
services share — DEFAULT_PORT, DEFAULT_STARTUP_TIMEOUT_SECONDS,
OrchestratorStartError, source_hash — which macOS / firecracker / client still
import from there. backend/docker/infra.py imports those (backend -> orchestrator
is an allowed direction). Renamed the unit test to test_docker_infra.py.

Full unit suite green (2243).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

tests/unit/test_docker_infra.py

@@ -0,0 +1,231 @@
+"""Unit: infra container lifecycle — idempotent singleton (PRD 0070)."""
+
+from __future__ import annotations
+
+import tempfile
+import unittest
+import urllib.error
+from pathlib import Path
+from unittest.mock import MagicMock, Mock, patch
+
+from bot_bottle.gateway import GatewayError
+from bot_bottle.backend.docker.infra import (
+    INFRA_NAME,
+    INFRA_SOURCE_HASH_LABEL,
+    DockerInfraService,
+)
+from bot_bottle.orchestrator.lifecycle import (
+    OrchestratorStartError,
+    source_hash,
+)
+from bot_bottle.paths import GATEWAY_CA_DIRNAME
+from tests.unit import use_bottle_root
+
+_URLOPEN = "bot_bottle.backend.docker.infra.urllib.request.urlopen"
+_RUN = "bot_bottle.backend.docker.infra.run_docker"
+_SLEEP = "bot_bottle.backend.docker.infra.time.sleep"
+_MONOTONIC = "bot_bottle.backend.docker.infra.time.monotonic"
+
+
+def _health(status: int) -> MagicMock:
+    m = MagicMock()
+    m.__enter__.return_value.status = status
+    return m
+
+
+def _proc(returncode: int = 0, stdout: str = "", stderr: str = "") -> Mock:
+    return Mock(returncode=returncode, stdout=stdout, stderr=stderr)
+
+
+class TestDockerInfraService(unittest.TestCase):
+    def setUp(self) -> None:
+        self._tmp = tempfile.TemporaryDirectory()
+        self.addCleanup(self._tmp.cleanup)
+        self.addCleanup(use_bottle_root(Path(self._tmp.name)))
+        self.svc = DockerInfraService(port=8099)
+
+    def test_url(self) -> None:
+        self.assertEqual("http://127.0.0.1:8099", self.svc.url)
+
+    def test_is_healthy(self) -> None:
+        with patch(_URLOPEN, return_value=_health(200)):
+            self.assertTrue(self.svc.is_healthy())
+        with patch(_URLOPEN, side_effect=urllib.error.URLError("refused")):
+            self.assertFalse(self.svc.is_healthy())
+
+    def test_ensure_running_noop_when_healthy_and_source_unchanged(self) -> None:
+        # A healthy container on current source is left alone — recreating it
+        # on every launch drops in-memory egress tokens (#381).
+        current = source_hash(self.svc._repo_root)
+        calls: list[list[str]] = []
+
+        def fake(argv: list[str], **_kw: object) -> Mock:
+            calls.append(argv)
+            if argv[:2] == ["docker", "ps"]:
+                return _proc(stdout=INFRA_NAME)
+            if argv[:2] == ["docker", "inspect"]:
+                return _proc(stdout=current)
+            return _proc()
+
+        with patch(_URLOPEN, return_value=_health(200)), \
+                patch(_RUN, side_effect=fake), patch(_SLEEP):
+            self.assertEqual(self.svc.url, self.svc.ensure_running())
+        runs = [c for c in calls if c[:2] == ["docker", "run"]]
+        rms = [c for c in calls if c[:3] == ["docker", "rm", "--force"] and INFRA_NAME in c]
+        self.assertEqual([], runs)
+        self.assertEqual([], rms)
+
+    def test_ensure_running_recreates_when_source_changed(self) -> None:
+        calls: list[list[str]] = []
+
+        def fake(argv: list[str], **_kw: object) -> Mock:
+            calls.append(argv)
+            if argv[:2] == ["docker", "ps"]:
+                return _proc(stdout=INFRA_NAME)
+            if argv[:2] == ["docker", "inspect"]:
+                return _proc(stdout="stale-hash")
+            return _proc()
+
+        with patch(_URLOPEN, return_value=_health(200)), \
+                patch(_RUN, side_effect=fake), patch(_SLEEP):
+            self.assertEqual(self.svc.url, self.svc.ensure_running())
+        runs = [c for c in calls if c[:2] == ["docker", "run"]]
+        self.assertEqual(1, len(runs))
+        self.assertIn(INFRA_NAME, runs[0])
+        current = source_hash(self.svc._repo_root)
+        self.assertIn(f"{INFRA_SOURCE_HASH_LABEL}={current}", runs[0])
+
+    def test_ensure_running_starts_infra_container_when_absent(self) -> None:
+        calls: list[list[str]] = []
+
+        def fake(argv: list[str], **_kw: object) -> Mock:
+            calls.append(argv)
+            if argv[:2] == ["docker", "ps"]:
+                return _proc(stdout="")
+            return _proc()
+
+        with patch(_URLOPEN, side_effect=[urllib.error.URLError("down"), _health(200)]), \
+                patch(_RUN, side_effect=fake), patch(_SLEEP):
+            self.assertEqual(self.svc.url, self.svc.ensure_running())
+        runs = [c for c in calls if c[:2] == ["docker", "run"]]
+        self.assertEqual(1, len(runs))
+        argv = runs[0]
+        self.assertIn(INFRA_NAME, argv)
+        # Published on loopback — not exposed on external interfaces.
+        self.assertEqual("127.0.0.1:8099:8099", argv[argv.index("--publish") + 1])
+        # Both processes in one container — no separate entrypoint override.
+        self.assertNotIn("--entrypoint", argv)
+        # Gateway daemons + orchestrator explicitly opted in.
+        daemons_flag = "BOT_BOTTLE_GATEWAY_DAEMONS=egress,git-http,supervise,orchestrator"
+        self.assertIn("orchestrator", argv[argv.index(daemons_flag)])
+        # The mitmproxy CA persists on a HOST bind-mount under the app-data root
+        # (not a docker named volume `docker volume prune` would wipe — #450), so
+        # a restarted infra container keeps the CA every running bottle trusts.
+        ca_mounts = [a for a in argv if a.endswith(":/home/mitmproxy/.mitmproxy")]
+        self.assertEqual(1, len(ca_mounts))
+        src = ca_mounts[0].rsplit(":", 1)[0]
+        self.assertTrue(src.startswith(self._tmp.name), src)
+        self.assertTrue(src.endswith("/" + GATEWAY_CA_DIRNAME), src)
+
+    def test_ensure_running_builds_all_images(self) -> None:
+        calls: list[list[str]] = []
+
+        def fake(argv: list[str], **_kw: object) -> Mock:
+            calls.append(argv)
+            if argv[:2] == ["docker", "ps"]:
+                return _proc(stdout="")
+            return _proc()
+
+        with patch(_URLOPEN, side_effect=[urllib.error.URLError("down"), _health(200)]), \
+                patch(_RUN, side_effect=fake), patch(_SLEEP):
+            self.svc.ensure_running()
+        builds = [c for c in calls if c[:2] == ["docker", "build"]]
+        # Gateway base + orchestrator intermediate + infra image — all three built.
+        self.assertEqual(3, len(builds))
+        dockerfiles = [next(a for a in b if "Dockerfile" in a) for b in builds]
+        self.assertIn("Dockerfile.gateway", dockerfiles[0])
+        self.assertIn("Dockerfile.orchestrator", dockerfiles[1])
+        self.assertIn("Dockerfile.infra", dockerfiles[2])
+        # All three images are distinct.
+        tags = [b[b.index("-t") + 1] for b in builds]
+        self.assertEqual(3, len(set(tags)))
+
+    def test_publish_maps_host_port_to_fixed_internal_port(self) -> None:
+        """A non-default self.port is published to the fixed internal port 8099,
+        not to self.port:self.port — the orchestrator always listens on 8099."""
+        calls: list[list[str]] = []
+
+        def fake(argv: list[str], **_kw: object) -> Mock:
+            calls.append(argv)
+            if argv[:2] == ["docker", "ps"]:
+                return _proc(stdout="")
+            return _proc()
+
+        svc = DockerInfraService(port=20001)
+        with patch(_URLOPEN, side_effect=[urllib.error.URLError("down"), _health(200)]), \
+                patch(_RUN, side_effect=fake), patch(_SLEEP):
+            svc.ensure_running()
+        runs = [c for c in calls if c[:2] == ["docker", "run"]]
+        argv = runs[0]
+        self.assertEqual("127.0.0.1:20001:8099", argv[argv.index("--publish") + 1])
+        orch_url = next(a for a in argv if "BOT_BOTTLE_ORCHESTRATOR_URL" in a)
+        self.assertIn(":8099", orch_url)
+
+    def test_ensure_running_raises_on_timeout(self) -> None:
+        with patch(_URLOPEN, side_effect=urllib.error.URLError("down")), \
+                patch(_RUN, return_value=Mock(returncode=0, stdout="", stderr="")), \
+                patch(_SLEEP), patch(_MONOTONIC, side_effect=[0.0, 0.5, 2.0]):
+            with self.assertRaises(OrchestratorStartError):
+                self.svc.ensure_running(startup_timeout=1.0)
+
+    def test_noop_when_healthy_and_inspect_fails(self) -> None:
+        """If docker inspect fails (e.g. docker daemon hiccup), leave the
+        working container alone rather than churning it."""
+        def fake(argv: list[str], **_kw: object) -> Mock:
+            if argv[:2] == ["docker", "ps"]:
+                return _proc(stdout=INFRA_NAME)
+            if argv[:2] == ["docker", "inspect"]:
+                return _proc(returncode=1, stderr="daemon error")
+            return _proc()
+
+        with patch(_URLOPEN, return_value=_health(200)), \
+                patch(_RUN, side_effect=fake), patch(_SLEEP):
+            self.svc.ensure_running()
+        # no docker run — the working container was left alone
+
+    def test_build_failure_raises(self) -> None:
+        with patch(_URLOPEN, side_effect=urllib.error.URLError("down")), \
+                patch(_RUN, return_value=_proc(returncode=1, stderr="no space left on device")):
+            with self.assertRaises(GatewayError):
+                self.svc.ensure_running()
+
+    def test_ensure_network_creates_if_missing(self) -> None:
+        """If the gateway network doesn't exist yet, create it."""
+        calls: list[list[str]] = []
+
+        def fake(argv: list[str], **_kw: object) -> Mock:
+            calls.append(argv)
+            if argv[:3] == ["docker", "network", "inspect"]:
+                return _proc(returncode=1, stderr="not found")
+            if argv[:2] == ["docker", "ps"]:
+                return _proc(stdout="")
+            return _proc()
+
+        with patch(_URLOPEN, side_effect=[urllib.error.URLError("down"), _health(200)]), \
+                patch(_RUN, side_effect=fake), patch(_SLEEP):
+            self.svc.ensure_running()
+        creates = [c for c in calls if c[:3] == ["docker", "network", "create"]]
+        self.assertEqual(1, len(creates))
+
+    def test_stop_removes_infra_container(self) -> None:
+        with patch(_RUN) as run:
+            self.svc.stop()
+        rms = [
+            c.args[0] for c in run.call_args_list
+            if c.args[0][:3] == ["docker", "rm", "--force"]
+        ]
+        self.assertTrue(any(INFRA_NAME in a for a in rms))
+
+
+if __name__ == "__main__":
+    unittest.main()