fix(gateway): preserve systemd handoff recovery

2026-08-22 07:51:54 +08:00
parent 5b024c7ccc
commit 91fb175188
4 changed files with 146 additions and 9 deletions
@@ -756,6 +756,81 @@ class TestGatewaySystemServiceRouting:
        assert "21627" not in out  # must use the mocked budget, not live defaults
        assert "27" in out

+    def test_systemd_restart_forces_recovery_only_when_handoff_has_no_replacement(
+        self, monkeypatch, capsys
+    ):
+        calls = []
+
+        monkeypatch.setattr(gateway_cli, "_select_systemd_scope", lambda system=False: False)
+        monkeypatch.setattr(gateway_cli, "_require_service_installed", lambda action, system=False: None)
+        monkeypatch.setattr(gateway_cli, "_preflight_user_systemd", lambda **kwargs: None)
+        monkeypatch.setattr(gateway_cli, "refresh_systemd_unit_if_needed", lambda system=False: None)
+        monkeypatch.setattr(gateway_cli, "_get_restart_exit_wait_budget", lambda: 27.0)
+        monkeypatch.setattr("gateway.status.get_running_pid", lambda: 654)
+        monkeypatch.setattr(gateway_cli, "_graceful_restart_via_sigusr1", lambda pid, timeout: True)
+        waits = iter((False, True))
+        monkeypatch.setattr(
+            gateway_cli,
+            "_wait_for_systemd_service_restart",
+            lambda system=False, previous_pid=None: next(waits),
+        )
+        monkeypatch.setattr(gateway_cli, "_systemd_service_is_start_limited", lambda system=False: False)
+        monkeypatch.setattr(
+            gateway_cli,
+            "_read_systemd_unit_properties",
+            lambda system=False, properties=None: {"ActiveState": "inactive", "MainPID": "0"},
+        )
+        monkeypatch.setattr(
+            gateway_cli,
+            "_run_systemctl",
+            lambda args, **kwargs: calls.append((args, kwargs))
+            or SimpleNamespace(returncode=0, stdout="", stderr=""),
+        )
+
+        gateway_cli.systemd_restart()
+
+        assert [call[0][0] for call in calls] == ["reset-failed", "restart"]
+        assert "did not relaunch" in capsys.readouterr().out
+
+    def test_systemd_restart_does_not_force_an_unready_replacement(self, monkeypatch):
+        calls = []
+
+        monkeypatch.setattr(gateway_cli, "_select_systemd_scope", lambda system=False: False)
+        monkeypatch.setattr(gateway_cli, "_require_service_installed", lambda action, system=False: None)
+        monkeypatch.setattr(gateway_cli, "_preflight_user_systemd", lambda **kwargs: None)
+        monkeypatch.setattr(gateway_cli, "refresh_systemd_unit_if_needed", lambda system=False: None)
+        monkeypatch.setattr(gateway_cli, "_get_restart_exit_wait_budget", lambda: 27.0)
+        monkeypatch.setattr("gateway.status.get_running_pid", lambda: 654)
+        monkeypatch.setattr(gateway_cli, "_graceful_restart_via_sigusr1", lambda pid, timeout: True)
+        monkeypatch.setattr(
+            gateway_cli,
+            "_wait_for_systemd_service_restart",
+            lambda system=False, previous_pid=None: False,
+        )
+        monkeypatch.setattr(gateway_cli, "_systemd_service_is_start_limited", lambda system=False: False)
+        monkeypatch.setattr(
+            gateway_cli,
+            "_read_systemd_unit_properties",
+            lambda system=False, properties=None: {"ActiveState": "active", "MainPID": "777"},
+        )
+        monkeypatch.setattr(gateway_cli, "_run_systemctl", lambda args, **kwargs: calls.append(args))
+
+        gateway_cli.systemd_restart()
+
+        assert calls == []
+
+    def test_systemd_restart_wait_timeout_includes_supervisor_budgets(self, monkeypatch):
+        monkeypatch.setattr(
+            gateway_cli,
+            "_read_systemd_unit_properties",
+            lambda system=False, properties=None: {
+                "RestartUSec": "5s",
+                "TimeoutStartUSec": "1min 30s",
+            },
+        )
+
+        assert gateway_cli._systemd_restart_wait_timeout() == 155.0
+