First attempt at serverless endpoint - "Initializing" for a long time
Hi.
New to RunPod, trying to run a serverless endpoint with a worker based on https://github.com/blib-la/runpod-worker-comfy and not able to get it past the "Initializing" status.
There are NO logs anywhere in the console
Here's what I did:
- cloned the github repo and modified the Dockerfile to include the model tensorfiles required
- Built the docker successfully, and pushed it to dockerhub as a public container (at least for now - to avoid authentication problems). On dockerhub it is: "ofertal/runpod-worker-comfy:pixelwave" and is about 9.5G in size
- Created a template that uses the container image "ofertal/runpod-worker-comfy:pixelwave"
- Deployed a serverless GPU Endpoint that uses that template with RTX3090 GPUs, 0 active workers, 4 max workers
- sent a run request to the Endpoint url
Here's what I observed:
- The worker has been "Initializing" for a very long time so far
- There are no logs in the logs tab ("No Logs Found. Try changing the date range or query.")
- I can see that my request was queued on the requests tab, but its status is IN_QUEUE and has been for a long time
- The workers tab has no workers available.
Any idea what I might be doing wrong here?
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.
2 Replies
nevermind - I figured it out. I was too restrictive on the GPU types in my endpoint selection. as soon as I allowed more types, the initialization happened and the workers were able to run.
It would have been nice to have some indication in the management console...
I've been actively building with Serverless runpod for the past few months for a client, and has been extremely frustrating for debugging. Not to late to find switch to something else