#TimeoutException and WAN

1 messages · Page 1 of 1 (latest)

ruby vale
# 
18:03:44.862 [Error] [BackendHandler] 7 requests denied due to backend timeout failure. Server backends are failing to respond.
18:03:44.864 [Error] [BackendHandler] Backend request #400 failed: System.TimeoutException: No backend has responded in 20 minutes.
18:03:44.885 [Error] [BackendHandler] Backend request #401 failed: System.TimeoutException: No backend has responded in 20 minutes.
18:03:44.885 [Error] [BackendHandler] Backend request #402 failed: System.TimeoutException: No backend has responded in 20 minutes.
18:03:44.886 [Error] [BackendHandler] Backend request #403 failed: System.TimeoutException: No backend has responded in 20 minutes.
18:03:44.886 [Error] [BackendHandler] Backend request #404 failed: System.TimeoutException: No backend has responded in 20 minutes.
18:03:44.887 [Error] [BackendHandler] Backend request #405 failed: System.TimeoutException: No backend has responded in 20 minutes.
18:03:44.887 [Error] [BackendHandler] Backend request #406 failed: System.TimeoutException: No backend has responded in 20 minutes.```

Can you please tell me how to change/remove this timeout handler for comfyui? I queued 9 images and 7 of them were calceled while I was gone because they "took to long", they were taking 15 minutes each and that's fine. I wanted to let it run while I was away.
finite portal
#

and no they weren't taking 15 minutes each for you, they were over 20 minutes, thus the error

#

that timeout is "backends all refused to take new jobs for over 20 minutes straight" ie any one gen exceed 20 min

#

which in the "image gen only" days was wildly longer than even a shitty computer would ever take

ruby vale
#

None of the metadata said they took that long though. I didn't time it personally, just believed the Metadata on how long they took.

ruby vale
finite portal
#

could be the other 5+ minutes was model loading time

ruby vale
#

That's interesting because when I watch them get made, I use lower frames and steps so they are faster and the models never take longer than a minute to load.

#

I posted the metada image of one of the 2 gens.

#

I guess I could time it somehow, but it doesn't matter as long as it doesn't timeout and cancel my queue

finite portal
#

all i can tell you is when that crash happened, it's because the backend didn't move for 20 minutes

#

the calculation for that is very accurate ftr

ruby vale
#

This was the reported metadata on the second image

#

I believe you

finite portal
#

it'd be the first image not the second

#

which prep_time says 32 min so im guessing that's the time the first took

ruby vale
#

Okay