docker download failure `4fadada67c94` - EOF about localai

[mudler]

@Anto79-ops from your logs looks like you are actually building the image rather then pulling it.

Can you share your docker-compose file?

[Anto79-ops]

Good day, @mudler it might have been a server error, i tried to again and it finished the download and it is now working. My docker compose file is below, if you wanted to see it anyways.

version: '3.9'

services:
  api:
    # See https://localai.io/basics/getting_started/#container-images for
    # a list of available container images (or build your own with the provided Dockerfile)
    # Available images with CUDA, ROCm, SYCL
    # Image list (quay.io): https://quay.io/repository/go-skynet/local-ai?tab=tags
    # Image list (dockerhub): https://hub.docker.com/r/localai/localai
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: 2
              capabilities: ["gpu"]
    image: quay.io/go-skynet/local-ai:master-cublas-cuda12-ffmpeg
    build:
      context: .
      dockerfile: Dockerfile
      args:
      - IMAGE_TYPE=core
      - BASE_IMAGE=ubuntu:22.04
    ports:
      - 8080:8080
    env_file:
      - .env
    environment:
      - MODELS_PATH=/models
      - DEBUG=true
    volumes:
      - /home/antonio/scratch/localai_models:/models
      - ./images/:/tmp/generated/images/
    command:
    # Here we can specify a list of models to run (see quickstart https://localai.io/basics/getting_started/#running-models )
    # or an URL pointing to a YAML configuration file, for example:
    # - https://gist.githubusercontent.com/mudler/ad601a0488b497b69ec549150d9edd18/raw/a8a8869ef1bb7e3830bf5c0bae29a0cce991ff8d/phi-2.yaml
    - phi-2

there is one (maybe unrelated) error in the log about failled authority 509, which could be my firewall. Is there a way to make downloads for LocalAI ignore these certificate requiremets etc? thanks

3:04PM INF loading environment variables from file envFile=.env
3:04PM INF Setting logging to debug
3:04PM INF Starting LocalAI using 2 threads, with models path: /models
3:04PM INF LocalAI version: 0e54942 (0e549424e782e315ee166efdb1cba77a1a4a750b)
3:04PM DBG [startup] resolved model to download: github://mudler/LocalAI/examples/configurations/phi-2.yaml@master
3:04PM INF Downloading "https://raw.githubusercontent.com/mudler/LocalAI/master/examples/configurations/phi-2.yaml"
3:04PM ERR error downloading model error="failed to download file \"/models/1701d57f28d47552516c2b6ecc3cc719.yaml\": Get \"https://raw.githubusercontent.com/mudler/LocalAI/master/examples/configurations/phi-2.yaml\": tls: failed to verify certificate: x509: certificate signed by unknown authority" filepath=/models/1701d57f28d47552516c2b6ecc3cc719.yaml url=github://mudler/LocalAI/examples/configurations/phi-2.yaml@master
3:04PM INF Preloading models from /models

[localai-bot]

Good day, @Anto79-ops. Thank you for providing the additional information. It seems that the issue you encountered was indeed a server error, as you were able to successfully pull the image after a few attempts.

Regarding your question about ignoring certificate requirements during downloads for LocalAI, you can disable certificate verification by adding --insecure-registry flag when pulling the Docker image. In your case, if you are pulling from Quay.io, you can use the following command:

docker pull --insecure-registry quay.io/go-skynet/local-ai:master-cublas-cuda12-ffmpeg

This command will bypass the certificate verification for the specified registry during the image pull process.

As for the 4fadada67c94 failure during the download of the v2.12.4-cublas-cuda12-ffmpeg image, it appears to be a networking or connectivity issue. The error message indicates that the download stopped and started repeatedly before eventually failing at 100%. This could be due to network congestion, intermittent connectivity issues, or a temporary server problem. I would recommend trying again after some time or contacting the Quay.io support if the issue persists.

If you have any further questions or issues, please let me know.