Fix env save-and-restart by deferring PM2 restart until after HTTP response.

Co-authored-by: Cursor <cursoragent@cursor.com>

lib/common/static/instance_settings_prefs.js

@@ -335,7 +335,11 @@
   }
 
   async function restartInstance() {
+    try {
       await fetchJson("/api/admin/restart", { method: "POST" });
+    } catch (_) {
+      // 重启会中断当前 HTTP 连接;只要后续 health 恢复即视为成功.
+    }
     const deadline = Date.now() + 90000;
     while (Date.now() < deadline) {
       await new Promise((r) => setTimeout(r, 2000));

lib/instance/instance_pm2_lib.py

@@ -2,6 +2,7 @@
 from __future__ import annotations
 
 import os
+import shlex
 import subprocess
 import sys
 from typing import Any
@@ -23,10 +24,34 @@ def resolve_pm2_app_name(exchange_key: str) -> str:
     return default_pm2_app_name(exchange_key)
 
 
-def restart_instance_pm2(exchange_key: str) -> dict[str, Any]:
+def schedule_pm2_restart(app_name: str, *, delay_seconds: float = 1.0) -> dict[str, Any]:
+    """延迟触发 PM2 重启,便于 HTTP 响应先返回(避免重启当前进程导致请求中断)."""
+    if not sys.platform.startswith("linux"):
+        return {"ok": False, "msg": "仅 Linux 服务器支持 PM2 重启", "app": app_name}
+    if not (app_name or "").strip():
+        return {"ok": False, "msg": "未指定 PM2 应用名", "app": app_name}
+    app_name = app_name.strip()
+    try:
+        cmd = f"sleep {delay_seconds} && exec pm2 restart {shlex.quote(app_name)} --update-env"
+        subprocess.Popen(
+            ["bash", "-c", cmd],
+            stdout=subprocess.DEVNULL,
+            stderr=subprocess.DEVNULL,
+            start_new_session=True,
+        )
+        return {"ok": True, "app": app_name, "msg": "重启已触发", "deferred": True}
+    except FileNotFoundError:
+        return {"ok": False, "msg": "未找到 bash 或 pm2 命令", "app": app_name}
+    except Exception as e:
+        return {"ok": False, "msg": str(e), "app": app_name}
+
+
+def restart_instance_pm2(exchange_key: str, *, defer: bool = False) -> dict[str, Any]:
     if not sys.platform.startswith("linux"):
         return {"ok": False, "msg": "仅 Linux 服务器支持 PM2 重启", "app": None}
     app_name = resolve_pm2_app_name(exchange_key)
+    if defer:
+        return schedule_pm2_restart(app_name)
     try:
         proc = subprocess.run(
             ["pm2", "restart", app_name, "--update-env"],

lib/instance/instance_settings_register.py

@@ -140,7 +140,7 @@ def register_instance_settings_routes(
     @app.route("/api/admin/restart", methods=["POST"])
     @api_auth
     def api_admin_restart():
-        result = restart_instance_pm2(exchange_key)
+        result = restart_instance_pm2(exchange_key, defer=True)
         code = 200 if result.get("ok") else 500
         return jsonify({"ok": bool(result.get("ok")), **result}), code
 

manual_trading_hub/hub_env_lib.py

@@ -2,8 +2,6 @@
 from __future__ import annotations
 
 import os
-import subprocess
-import sys
 from pathlib import Path
 from typing import Any
 
@@ -13,6 +11,7 @@ from lib.env.shared_env_lib import (
     build_ai_env_payload,
     restart_instances_then_hub_pm2,
 )
+from lib.instance.instance_pm2_lib import schedule_pm2_restart
 
 HUB_DIR = Path(__file__).resolve().parent
 
@@ -46,25 +45,4 @@ def restart_all_pm2() -> dict[str, Any]:
 
 def restart_hub_pm2() -> dict[str, Any]:
     app_name = (os.getenv("PM2_APP_NAME") or "").strip() or "manual-trading-hub"
-    if not sys.platform.startswith("linux"):
+    return schedule_pm2_restart(app_name)
-        return {"ok": False, "msg": "仅 Linux 服务器支持 PM2 重启", "app": app_name}
-    try:
-        proc = subprocess.run(
-            ["pm2", "restart", app_name, "--update-env"],
-            capture_output=True,
-            text=True,
-            timeout=120,
-        )
-        ok = proc.returncode == 0
-        return {
-            "ok": ok,
-            "app": app_name,
-            "msg": (proc.stdout or proc.stderr or "").strip()[:500],
-            "returncode": proc.returncode,
-        }
-    except FileNotFoundError:
-        return {"ok": False, "msg": "未找到 pm2 命令", "app": app_name}
-    except subprocess.TimeoutExpired:
-        return {"ok": False, "msg": "pm2 restart 超时", "app": app_name}
-    except Exception as e:
-        return {"ok": False, "msg": str(e), "app": app_name}

tests/test_instance_pm2_lib.py

@@ -0,0 +1,35 @@
+from unittest.mock import MagicMock, patch
+
+from lib.instance.instance_pm2_lib import restart_instance_pm2, schedule_pm2_restart
+
+
+@patch("lib.instance.instance_pm2_lib.sys.platform", "linux")
+@patch("lib.instance.instance_pm2_lib.subprocess.Popen")
+def test_schedule_pm2_restart_returns_before_pm2(mock_popen):
+    result = schedule_pm2_restart("crypto_okx")
+
+    assert result["ok"] is True
+    assert result["deferred"] is True
+    mock_popen.assert_called_once()
+
+
+@patch("lib.instance.instance_pm2_lib.sys.platform", "linux")
+@patch("lib.instance.instance_pm2_lib.schedule_pm2_restart")
+def test_restart_instance_pm2_defer_uses_schedule(mock_schedule):
+    mock_schedule.return_value = {"ok": True, "app": "crypto_okx", "deferred": True}
+
+    result = restart_instance_pm2("okx", defer=True)
+
+    mock_schedule.assert_called_once_with("crypto_okx")
+    assert result["deferred"] is True
+
+
+@patch("lib.instance.instance_pm2_lib.sys.platform", "linux")
+@patch("lib.instance.instance_pm2_lib.subprocess.run")
+def test_restart_instance_pm2_sync_runs_pm2(mock_run):
+    mock_run.return_value = MagicMock(returncode=0, stdout="ok", stderr="")
+
+    result = restart_instance_pm2("okx", defer=False)
+
+    mock_run.assert_called_once()
+    assert result["ok"] is True