#cannot access comfy UI
69 messages · Page 1 of 1 (latest)
502 usually means the service is still booting up. Comfy can take its time, especially when your /output folder has many images or if you are loading many custom_nodes
just login to jupter, open a terminal and run "ps -aux |grep python". Is comfy amongst the processes?
also, what template are you using?
this seems like you got a bad node. either you are not getting a working GPU or you have a wrong CUDA build ..
try running
nvidia-smi || true ls -la /dev/nvidia* 2>/dev/null || true
and send me the output
but the fastest solution for you is to terminate and deploy a new pod (if you are using network storage)
or just shutdown and start again if you have volume storage
feel free to dm me if u need more help
did that
over and over and over
not the command
but what you are telling me to get another pod
@regal turtle every time i create one from my network storage
EVERY TIME
just created another network storage
US NO 1 is fucked?
it definitely has nothing to do with network storage. I am more concerned about that GPU error on top. What do you see in logs?
but its possible you are landing on a bad node. Terminate / redeploy until the error is gone
or try different gpu
so now i got my netwrok storage
i had used it with a previous GPU
i deployed a pod with it
and wont start
@regal turtle
whats goin gon here guys
i got so much data on this network volume
i just want something that works man
that start.sh script is referring to uv command which is clearly missing
which template are you using?
and what do you actually need? just comfy?
try my template: https://console.runpod.io/deploy?template=gg1utaykxa&ref=o3idfm0n
5090
should work just fine
i understand, it is probably unmaintained and needs a fix
but how is it missing i was running this pod for a whole day with comfy
no idea .. maybe someone pushed a new version of the image which wasnt properly tested?
most likely
cuz i cant create anything with it
can you check ?
it always worked
whats on this template?
this is the problem
on the one I gave you link for? www.lorapilot.com
dm me
or show me the log
it should be running on port 8888 - or just open Control Pilot on 7878 and check services from main menu
i need to know what template are you using. This 502 error from cloudflare means the service didnt start yet. Without seeing the logs I am unable to help you
@regal turtle gusy please help me
Using existing FileBrowser configuration...
Starting FileBrowser on port 8080...
Starting Jupyter Lab on port 8888...
Jupyter Lab started
Updating ComfyUI dependencies...
Checking for custom node dependencies...
Checking dependencies for Bond-Node-Suite/...
Installing requirements.txt for Bond-Node-Suite/
/start.sh: line 242: uv: command not found
Exporting environment variables...
Using existing FileBrowser configuration...
Starting FileBrowser on port 8080...
Starting Jupyter Lab on port 8888...
Jupyter Lab started
Updating ComfyUI dependencies...
Checking for custom node dependencies...
Checking dependencies for Bond-Node-Suite/...
Installing requirements.txt for Bond-Node-Suite/
/start.sh: line 242: uv: command not found
Exporting environment variables...
Using existing FileBrowser configuration...
Starting FileBrowser on port 8080...
Starting Jupyter Lab on port 8888...
Jupyter Lab started
Updating ComfyUI dependencies...
Checking for custom node dependencies...
Checking dependencies for Bond-Node-Suite/...
Installing requirements.txt for Bond-Node-Suite/
/start.sh: line 242: uv: command not found
im getting this again.
why is run pod so unstable
@late citrus can you help support