Docker image constantly restarting

[lnlyssg]

cc'ing @benyafai

I'm using the below docker-compose to run fedifetcher however I've noticed that it's constantly running, ending and running again within minutes (or even seconds) as per the below screenshot from my Healthchecks instance. It seems that it's not picking up the delay in the restart_policy but I'm not sure why?

services:
  fedifetcher:
    container_name: fedifetcher
    stdin_open: true
    tty: true
    image: ghcr.io/nanos/fedifetcher:latest
    command: "--access-token=${TOKEN} --server=${SERVER} --config=/app/artifacts/config.json"
    # Persist our data
    volumes:
      - /opt/fedifetcher/:/app/artifacts
    # Use the `deploy` option to enable `restart_policy`
    deploy:
      # Don't go above 1 replica to avoid multiple overlapping executions of the script
      replicas: 1
      restart_policy:
        # The `any` condition means even after successful runs, we'll restart the script
        condition: any
        # Specify how often the script should run - for example; after 1 hour.
        delay: 1h

[benyafai]

It seems docker compose ignores the delay command, it only works when deployed as a swarm/stack. I apologise I didn't see this, in my limited testing I thought I had it working.

I 'solved' it by creating my own entrypoint in the compose file and adding a basic unix sleep command like below.

entrypoint: ["sh", "-c", "python /app/find_posts.py --access-token=<TOKEN> --server=<SERVER>; sleep 3600; "]

I'll create a pull for this, but @nanos can feel free to remove the example entirely if it doesn't fit as well any more!

[nanos]

Thanks @benyafai ! I will take a look at this now