microsoft/autogen · error · RuntimeError

The group chat is currently running. It must be stopped befo

Error message

The group chat is currently running. It must be stopped before it can be reset.

What it means

reset_state() refuses to run while the team is mid-run: it initializes if needed, then checks _is_running and raises RuntimeError, because resetting participants during an active run would corrupt in-flight message routing and agent state.

Source

Thrown at python/packages/autogen-agentchat/src/autogen_agentchat/teams/_group_chat/_base_group_chat.py:624

                stream = team.run_stream(task="Count from 1 to 10, respond one at a time.")
                async for message in stream:
                    print(message)

                # Reset the team.
                await team.reset()
                stream = team.run_stream(task="Count from 1 to 10, respond one at a time.")
                async for message in stream:
                    print(message)


            asyncio.run(main())
        """

        if not self._initialized:
            await self._init(self._runtime)

        if self._is_running:
            raise RuntimeError("The group chat is currently running. It must be stopped before it can be reset.")
        self._is_running = True

        if self._embedded_runtime:
            # Start the runtime.
            assert isinstance(self._runtime, SingleThreadedAgentRuntime)
            self._runtime.start()

        try:
            # Send a reset messages to all participants.
            for participant_topic_type in self._participant_topic_types:
                await self._runtime.send_message(
                    GroupChatReset(),
                    recipient=AgentId(type=participant_topic_type, key=self._team_id),
                )
            # Send a reset message to the group chat manager.
            await self._runtime.send_message(
                GroupChatReset(),
                recipient=AgentId(type=self._group_chat_manager_topic_type, key=self._team_id),

View on GitHub (pinned to 027ecf0a37)

Solutions

1. Finish or stop the current run first: consume run_stream to completion or await stream.aclose() / cancel the run task.
2. Then call await team.reset_state().
3. Guard resets with the same asyncio.Lock used to serialize runs.

Example fix

# before
stream = team.run_stream(task="hi")
async for msg in stream: break
await team.reset_state()  # RuntimeError

# after
stream = team.run_stream(task="hi")
async for msg in stream: break
await stream.aclose()
await team.reset_state()

async with team_lock:
    # run to completion inside the same critical section
    async for _ in team.run_stream(task=task):
        pass
    await team.reset_state()

try:
    await team.reset_state()
except RuntimeError as e:
    if "must be stopped" in str(e):
        await stream.aclose()  # release the running flag, then retry
        await team.reset_state()
    else:
        raise

When it happens

Trigger: Calling await team.reset_state() (or the team.reset() alias) after starting run_stream but before the stream has fully completed — including when the consuming loop is merely paused between messages.

Common situations: Orchestrator code that starts a run in a background task and resets on a timer; calling reset after breaking out of an async for loop over run_stream without closing the generator.

Related errors

• The team is already running, it cannot run again until it is
• The team cannot be loaded while it is running.
• The group chat has not been initialized. It must be run befo
• The group chat has not been initialized. It must be run befo
• Connection is not open.

AI-assisted analysis of microsoft/autogen@027ecf0a37 (2026-08-15). Data as JSON: /api/errors/2a92591a87f61f6d. Report an issue: GitHub.