comfyanonymous
Comfyui stops when I leave my pc
Expected Behavior
keep running
Actual Behavior
stops
Steps to Reproduce
happens at random
Debug Logs
/
Other
I've had it multiple times by now that comfy runs for like 5-15 min after I shut off my screens and then return hours later only for nothing to have happened in the mean time and then suddenly as I turn on my screens to check progress that the queue starts again 1-2 min later only for it not to have generated a single image in the time I was not there, very strange behaviour.
I'm having a similar issue. I'm launching comfyui from within a tmux virtual terminal on linux - but the server seems to die when the client session goes to sleep - despite the virtual terminal a full queue.
`Traceback (most recent call last): File "/usr/lib/python3.10/asyncio/selector_events.py", line 862, in _read_ready__data_received data = self._sock.recv(self.max_size) TimeoutError: [Errno 110] Connection timed out
The above exception was the direct cause of the following exception:
Traceback (most recent call last):
File "/home/[USER]/ComfyUI/main.py", line 270, in
I have the same issue. Running on Runpod with Flux.
2024-09-13T11:02:11.778477815Z File "/usr/lib/python3.10/asyncio/selector_events.py", line 862, in _read_ready__data_received
2024-09-13T11:02:11.778482425Z data = self._sock.recv(self.max_size)
2024-09-13T11:02:11.778486145Z ConnectionResetError: [Errno 104] Connection reset by peer
2024-09-13T11:02:11.778489655Z
2024-09-13T11:02:11.778492965Z The above exception was the direct cause of the following exception:
2024-09-13T11:02:11.778496925Z
2024-09-13T11:02:11.778499905Z Traceback (most recent call last):
2024-09-13T11:02:11.778503015Z File "/workspace/ComfyUI/main.py", line 250, in <module>
2024-09-13T11:02:11.778506325Z loop.run_until_complete(run(server, address=args.listen, port=args.port, verbose=not args.dont_print_server, call_on_start=call_on_start))
2024-09-13T11:02:11.778509785Z File "/usr/lib/python3.10/asyncio/base_events.py", line 649, in run_until_complete
2024-09-13T11:02:11.778513175Z return future.result()
2024-09-13T11:02:11.778516985Z File "/workspace/ComfyUI/main.py", line 163, in run
2024-09-13T11:02:11.778520385Z await asyncio.gather(server.start(address, port, verbose, call_on_start), server.publish_loop())
2024-09-13T11:02:11.778524015Z File "/workspace/ComfyUI/server.py", line 805, in publish_loop
2024-09-13T11:02:11.778527065Z await self.send(*msg)
2024-09-13T11:02:11.778530225Z File "/workspace/ComfyUI/server.py", line 736, in send
2024-09-13T11:02:11.778533275Z await self.send_image(data, sid=sid)
2024-09-13T11:02:11.778536205Z File "/workspace/ComfyUI/server.py", line 773, in send_image
2024-09-13T11:02:11.778539195Z await self.send_bytes(BinaryEventTypes.PREVIEW_IMAGE, preview_bytes, sid=sid)
2024-09-13T11:02:11.778542205Z File "/workspace/ComfyUI/server.py", line 783, in send_bytes
2024-09-13T11:02:11.778545325Z await send_socket_catch_exception(self.sockets[sid].send_bytes, message)
2024-09-13T11:02:11.778555185Z File "/workspace/ComfyUI/server.py", line 42, in send_socket_catch_exception
2024-09-13T11:02:11.778558485Z await function(message)
The only way to avoid a crash seems to be disabling any kind of preview method.
I think i have the same issue I start my server and then i use comfyui all is ok at this point. I stop to use Comfyui it but still stay in up. I come back and the machine is completely crashed i have to restart entirely it manually. When i don't use comfyui but is launch and not used, the machine don't crash.. it's when i use Cmfyuin after that it will crshed the machine. I tried to clean modules and this is the same. I mean I d'ont understand....
I had similar issue of comfyui going idle after ~10m when set to instant generation even when PC remains powered and up. I was able to resolve by excluding comfyui URL from tab power-saving mode in my browser. (Latest comfyui version are set to stop the generation queue if client (browser) disconnects - and tab in powersave mode is disconnected. )
Same error here! Please see the attached log!
We calls the comfyui API to implement the image generation task, and everything works fine. However, if the client stops calling the comfyui API for more than two hours, the problem occurs when calling it again.
Description of the log
line 0 ~ line 816, from 9:42:39 ~ 9:42:40 Client got many 'type=progress_state' logs in 2 seconds.
line 816 ~ line 1211, from 9:42:40 ~ 9:44:02 no log is got from the comfyuiserver more then 1 minute. Actually, from the terminal of comfyui, the tasks is already done! Seems the websocket messages were jammed.
line 1211 ~ line 1272, from 9:44:02 ~ 9:44:02 received the remaining logs
