fix: restart-to-apply never re-execs — drop stop_polling race in restart_bot (v0.14.1)

Author: dervish666

.claude/structure/root.yaml

@@ -33,8 +33,8 @@ files:
     exports:
       - name: restart_bot
         kind: function
-        description: Stops polling, optionally broadcasts a notice, then re-execs `python -m src.main`.
-        params: [{name: bot, type: Bot}, {name: dp, type: Dispatcher}, {name: chat_id_store, type: ChatIdStore}, {name: notice, type: "str | None"}]
+        description: Optionally broadcasts a notice, then re-execs `python -m src.main`. Does NOT stop polling/close the session (that would race main()'s teardown and cancel execv); os.execv replaces the whole process.
+        params: [{name: bot, type: Bot}, {name: chat_id_store, type: ChatIdStore}, {name: notice, type: "str | None"}]
         returns: Coroutine
     imports:
       - from: src.alerts.manager

CHANGELOG.md

@@ -2,6 +2,11 @@
 
 All notable changes to UnraidMonitor will be documented in this file.
 
+## [0.14.1] - 2026-06-05
+
+### Fixed
+- **Restart-to-apply never came back** — enabling image updates (and the setup-wizard restart) called `dp.stop_polling()` before `os.execv`, which unwound `main()`'s polling loop and ran its shutdown `finally`, cancelling the restart coroutine before it re-exec'd. The bot shut down cleanly but never restarted (it only recovered if Docker's restart policy happened to be set). `restart_bot()` now re-execs directly without stopping polling or closing the session — `os.execv` replaces the whole process, and Python's close-on-exec defaults drop the in-flight long-poll so Telegram frees the getUpdates slot for the fresh process.
+
 ## [0.14.0] - 2026-06-05
 
 ### Added

CLAUDE.md

@@ -92,7 +92,7 @@ src/utils/version_store.py - read/write data/announced_version.json for startup
 
 ## Project Overview
 
-Unraid Server Monitor Bot (v0.14.0) - A Docker-based Telegram bot for monitoring Unraid servers. Monitors Docker containers (events, logs, resources) and Unraid server health (CPU, memory, disks, array, UPS). Uses multi-provider LLM support (Anthropic, OpenAI, Ollama) for AI-powered diagnostics and natural language interaction. Sends alerts via Telegram with quick-action buttons.
+Unraid Server Monitor Bot (v0.14.1) - A Docker-based Telegram bot for monitoring Unraid servers. Monitors Docker containers (events, logs, resources) and Unraid server health (CPU, memory, disks, array, UPS). Uses multi-provider LLM support (Anthropic, OpenAI, Ollama) for AI-powered diagnostics and natural language interaction. Sends alerts via Telegram with quick-action buttons.
 
 **Version string lives in TWO places** — `pyproject.toml` (`version`) and `src/__init__.py` (`__version__`). Both must be bumped together on release; `tests/test_version.py` guards against drift. `BOT_VERSION` (in `src/bot/health_command.py`) resolves from installed package metadata when available, else falls back to `src.__version__` — the package is not installed in the Docker image, so the `__version__` fallback is what runs in production.
 

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "unraid-monitor-bot"
-version = "0.14.0"
+version = "0.14.1"
 requires-python = ">=3.11"
 dependencies = [
     "docker>=7.0.0,<8.0.0",

src/__init__.py

@@ -1 +1 @@
-__version__ = "0.14.0"
+__version__ = "0.14.1"

src/main.py

@@ -83,7 +83,7 @@ async def on_wizard_complete() -> None:
             """Called when the wizard saves config.yaml for the first time."""
             logger.info("Setup wizard complete -- restarting to apply config")
             await restart_bot(
-                bot, dp, chat_id_store,
+                bot, chat_id_store,
                 notice="✅ Setup complete! Restarting to apply configuration...",
             )
 
@@ -123,7 +123,7 @@ async def on_rerun_complete() -> None:
                 """Called when a /setup re-run saves updated config."""
                 logger.info("Setup wizard re-run complete -- restarting to apply config")
                 await restart_bot(
-                    bot, dp, chat_id_store,
+                    bot, chat_id_store,
                     notice="✅ Configuration updated! Restarting to apply changes...",
                 )
 

src/restart.py

@@ -5,34 +5,47 @@
 src.main`` so the supervisor (Docker) sees the same process.
 """
 
-import asyncio
 import logging
 import os
 import sys
 from typing import TYPE_CHECKING
 
 if TYPE_CHECKING:
-    from aiogram import Bot, Dispatcher
+    from aiogram import Bot
     from src.alerts.manager import ChatIdStore
 
 logger = logging.getLogger(__name__)
 
 
 async def restart_bot(
     bot: "Bot",
-    dp: "Dispatcher",
     chat_id_store: "ChatIdStore",
     notice: str | None = None,
 ) -> None:
-    """Stop polling and re-exec the process so new config takes effect.
+    """Replace the running process with a fresh ``python -m src.main``.
 
     Args:
         bot: The aiogram Bot, used to broadcast ``notice`` if provided.
-        dp: The dispatcher to stop polling on.
         chat_id_store: Source of chat IDs for the optional broadcast.
         notice: If set, sent to every known chat before restarting. Callers
             that have already messaged the user (e.g. an inline-button handler
             editing its own message) should pass ``None``.
+
+    We deliberately do NOT call ``dp.stop_polling()`` or ``bot.session.close()``
+    first: either would unwind ``main()``'s polling loop and run its
+    ``finally`` (``bg.shutdown()`` + loop teardown), which cancels this
+    coroutine *before* it reaches ``os.execv`` — the restart then silently
+    degrades to a plain exit that only Docker's restart policy (if any) would
+    recover. ``os.execv`` replaces the whole process image, so polling, tasks
+    and sockets all stop regardless. Python marks file descriptors
+    close-on-exec by default, so the in-flight long-poll connection drops on
+    exec and Telegram frees the getUpdates slot immediately (no 409 conflict
+    for the fresh process).
+
+    The broadcast ``await`` below resolves once Telegram has the message, and
+    there is intentionally no awaitable between it and ``execv`` so nothing can
+    preempt the restart. ``execv`` only returns on failure, in which case the
+    bot keeps running normally.
     """
     logger.info("Restarting bot to apply configuration changes")
     if notice:
@@ -41,6 +54,4 @@ async def restart_bot(
                 await bot.send_message(cid, notice)
             except Exception:
                 pass
-    await dp.stop_polling()
-    await asyncio.sleep(1)
     os.execv(sys.executable, [sys.executable, "-m", "src.main"])

src/startup.py

@@ -463,7 +463,7 @@ async def start_monitoring(
     )
 
     async def _restart_to_apply() -> None:
-        await restart_bot(bot, dp, chat_id_store)
+        await restart_bot(bot, chat_id_store)
 
     controller, diagnostic_service = register_commands(
         dp,

tests/test_restart.py

@@ -8,54 +8,75 @@
 
 
 @pytest.mark.asyncio
-async def test_restart_broadcasts_notice_stops_polling_and_execs():
+async def test_restart_broadcasts_notice_and_execs():
     bot = MagicMock()
     bot.send_message = AsyncMock()
-    dp = MagicMock()
-    dp.stop_polling = AsyncMock()
     store = MagicMock()
     store.get_all_chat_ids.return_value = [1, 2]
 
-    with patch("src.restart.os.execv") as execv, \
-            patch("src.restart.asyncio.sleep", new=AsyncMock()):
-        await restart_bot(bot, dp, store, notice="restarting")
+    with patch("src.restart.os.execv") as execv:
+        await restart_bot(bot, store, notice="restarting")
 
     assert bot.send_message.await_count == 2
-    dp.stop_polling.assert_awaited_once()
     execv.assert_called_once()
 
 
 @pytest.mark.asyncio
 async def test_restart_without_notice_skips_broadcast():
     bot = MagicMock()
     bot.send_message = AsyncMock()
-    dp = MagicMock()
-    dp.stop_polling = AsyncMock()
     store = MagicMock()
     store.get_all_chat_ids.return_value = [1, 2]
 
-    with patch("src.restart.os.execv") as execv, \
-            patch("src.restart.asyncio.sleep", new=AsyncMock()):
-        await restart_bot(bot, dp, store)
+    with patch("src.restart.os.execv") as execv:
+        await restart_bot(bot, store)
 
     bot.send_message.assert_not_awaited()
-    dp.stop_polling.assert_awaited_once()
+    execv.assert_called_once()
+
+
+@pytest.mark.asyncio
+async def test_restart_does_not_stop_polling_or_close_session():
+    """Regression: stopping polling/closing the session unwinds main()'s loop
+    and cancels this coroutine before execv runs (the restart silently becomes
+    a clean exit). restart_bot must touch neither."""
+    bot = MagicMock()
+    bot.send_message = AsyncMock()
+    bot.session = MagicMock()
+    bot.session.close = AsyncMock()
+    store = MagicMock()
+    store.get_all_chat_ids.return_value = [1]
+
+    with patch("src.restart.os.execv") as execv:
+        await restart_bot(bot, store, notice="hi")
+
+    bot.session.close.assert_not_awaited()
     execv.assert_called_once()
 
 
 @pytest.mark.asyncio
 async def test_restart_swallows_send_failures():
     bot = MagicMock()
     bot.send_message = AsyncMock(side_effect=Exception("network"))
-    dp = MagicMock()
-    dp.stop_polling = AsyncMock()
     store = MagicMock()
     store.get_all_chat_ids.return_value = [1]
 
-    with patch("src.restart.os.execv") as execv, \
-            patch("src.restart.asyncio.sleep", new=AsyncMock()):
-        await restart_bot(bot, dp, store, notice="hi")
+    with patch("src.restart.os.execv") as execv:
+        await restart_bot(bot, store, notice="hi")
 
     # A failed broadcast must not prevent the restart
-    dp.stop_polling.assert_awaited_once()
     execv.assert_called_once()
+
+
+@pytest.mark.asyncio
+async def test_restart_execs_current_module():
+    bot = MagicMock()
+    store = MagicMock()
+    store.get_all_chat_ids.return_value = []
+
+    with patch("src.restart.os.execv") as execv:
+        await restart_bot(bot, store)
+
+    args = execv.call_args.args
+    # os.execv(python, [python, "-m", "src.main"])
+    assert args[1][-2:] == ["-m", "src.main"]