Zero GPU issue
I wanted to start up a SD Comfy UI pod I created the other day and when I start it up I get a pop-up with the message stating that I don't have access to any GPUs and that I should consider creating a network volume. I click the link to learn more which goes to this page
https://docs.runpod.io/references/faq?_gl=1*lokxwm*_ga*OTAwNDMyOTA4LjE3MDY5MDYzNDc.*_ga_KMF5V28LQG*MTcwODM0NDg1Ni4xMi4xLjE3MDgzNDQ4NTYuNjAuMC4yMDg3Nzg5MzIz*_gcl_au*MjA5MjM3NzI3Ny4xNzA2OTA2MzQ3#why-do-i-have-zero-gpus-assigned-to-my-pod
And in that section there is a link to Learn how to use them (network volumes) and it links to a page with out any tutorial on how to set one up. Here is that link https://docs.runpod.io/pods/network-storage/create-network-volumes
Can you get that page online and/or help me here?
13 Replies
Very simple. Select Storage, create a Network Volume in a region that has the GPU type that you require and click the Deploy button from it once its created.
thank you!
I definitely suggest getting that page up so other users don't hit this issue.
@ashleyk Ok I set up a network volume and started a RunPod SD Comfy UI with 1 x RTX 4000 Ada
9 vCPU 50 GB RAM. I see that the pod is running with the green "Running"indicator but I do not see a the square icon Stop button anywhere. How do I stop it?
You can't stop a pod with a network volume attached. You can only terminate it.
So terminating will delete the whole thing but my data will remain. Right?
also, the other question is how much I will be charged for this network volume per hour?
Ok I have hit another problem. It seems like my disk space ran out becuase this is what the Container Log says when I start it up
Search
0 matches
2024-02-19T17:15:24.657053547Z Starting Nginx service...
2024-02-19T17:15:24.669111398Z * Starting nginx nginx
2024-02-19T17:15:24.682153303Z ...done.
2024-02-19T17:15:24.682171244Z Running pre-start script...
2024-02-19T17:15:25.434425497Z ln: failed to create symbolic link '/workspace/ComfyUI/models/checkpoints/SDv2-768.ckpt': File exists
2024-02-19T17:15:25.434640344Z ln: failed to create symbolic link '/workspace/ComfyUI/models/checkpoints/sd_xl_base_1.0.safetensors': File exists
2024-02-19T17:15:25.434921034Z ln: failed to create symbolic link '/workspace/ComfyUI/models/checkpoints/sd_xl_refiner_1.0.safetensors': File exists
2024-02-19T17:15:25.435263046Z ln: failed to create symbolic link '/workspace/ComfyUI/models/checkpoints/v1-5-pruned-emaonly.ckpt': File exists
2024-02-19T17:15:25.437535895Z Pod Started
2024-02-19T17:15:25.437617488Z Starting Jupyter Lab...
2024-02-19T17:15:25.437835336Z Jupyter Lab started
2024-02-19T17:15:25.437886027Z Exporting environment variables...
2024-02-19T17:15:25.441920398Z Start script(s) finished, pod is ready to use.
2024-02-19T17:15:25.538089716Z Exception ignored in atexit callback: <function close_log at 0x7f929cbd5510>
2024-02-19T17:15:25.538112967Z Traceback (most recent call last):
2024-02-19T17:15:25.538116907Z File "/workspace/ComfyUI/custom_nodes/ComfyUI-Manager/prestartup_script.py", line 273, in close_log
2024-02-19T17:15:25.539216275Z log_file.close()
2024-02-19T17:15:25.539227436Z OSError: [Errno 122] Disk quota exceeded
Also I see 100% on the Container Volume bar
an d here is the screenshot for the Container Log
Also I keep trying to increase the Container size by going to Edit > Container Disk > and then increasing the number. I think it started off as 5gb. Then I upped it to 20gb. Now I tried 40gb. Nothing works
Ur network and container is way too small
set both to 50gb
Shouldn't need such a large container disk. Also RunPod ComfyUI template sucks and should be removed, everyone has issues with it. Use the community "ComfyUI - Jupyter" template instead.