Like the title says, I want to run an app that’s simply going to contain a directory of scripts that interact with resources in the network that are accessible via our fly.io private net, but not from outside.
What’s the best way to deploy this? Do I need an ENTRYPOINT or RUN command in a Dockerfile with an infinite loop? Can the machine sleep and only start when invoked via fly ssh console?
I haven’t done this myself so if someone else have more tips please join in but here’s the general tips for what you’d need:
Do not put [services] on your toml. Those are for web services and your release would fail unless you listened to a port of your choosing.
Make sure you don’t have public IPs. We generate those only if you have [services]. If you do you can fly ips release [IP] no worries.
Maybe you’d need an entrypoint/cmd but that could just be something like while true; do sleep 1; done (I feel like I’ve seen a better one, maybe sleep inf?).
That used to work in v1 apps (scale count 1), but when I do scale count 0, the machine is destroyed, so there’s no way to start it up again. Is there an easier way to start a new machine for a given app?