Serverless with network storage
Hi all,
I am trying to setup a serverless worker for comfyui (currently using customized template from this https://github.com/blib-la/runpod-worker-comfy .
I have a several large models which I would like not to bake into the image.
I see there is an option to mount network storage to serverless worker, I tried to mount it (with the required models to run the workflow) to the serverless comfy worker, but when I send a request with the workflow I see in the worker logs that it does not see any of the models in the mounted storage.
There are also no customization for the network storage mount for a serverless worker, so I am not even sure if the paths are mounted correctly.
Therefore I want to ask:
- Is this type of functionality/usecase is supported/feasible?
- If it is, what am I missing or not doing correctly?
Thanks in advance!
GitHub
GitHub - blib-la/runpod-worker-comfy: ComfyUI as a serverless API o...
ComfyUI as a serverless API on RunPod. Contribute to blib-la/runpod-worker-comfy development by creating an account on GitHub.
13 Replies
on serverless network volume isn't mounted on /workspace, not really clear in docs
check /runpod-volume i think
maybe the paths?, where does your comfy reads your model from
I also assume this is probably because of the incorrectly mounted paths, but I do not see the option to specify the network storage path neither in the serverless deploy options, not in the template creation screen (which is weird as there is an option to specify network storage mount path for a non-serverless option, but when you select serverless the option is no longer there), nor in the documentation. Maybe someone from runpod team can clarify/help with this?
its always at /runpod-volume for serverless
roger, will try to add a symlink in the Dockerfile for /runpod-volume and see if it works
did it work @kranas4755 ?
yes, seems to work when I add a symlink to the start.sh file in the docker
rm -rf /comfyui/models && ln -s /runpod-volume/ComfyUI/models /comfyui/modelsneeded to figure out the path to runpod-volume myself, I think it should be documented somewhere thanks @nerdylive @tzushi
@kranas4755 I'm interested to know wether this works with little cold start lag for you ? Do you have quick start time or does the network storage takes a minute or more to boot up
Have only tested once so far, didnt seem to be much different loading wise than without network storage, but will test some more later
So you just deploy a normal pod with the volume and download models to runpod-volume/ path, right? Then we can use this storage volume with serverless @kranas4755
Yes
oh, normal pod doesn't have runpod-volume/ path itself, i have to create a dir for this
Unknown User•3w ago
Message Not Public
Sign In & Join Server To View