Anyone else see issues like this happening when trying to deployment via GitHub actions?
If I am reading this correctly, it looks an instance on fly.io is failing to build the Docker container. Hard to say, since there really isn’t a helpful error message. :-/
I looked at this builder, it looks like it’s crashing the Docker Daemon somehow. If you run fly machine status e148e03b945892 you’ll see exit code 138, which is super strange.
Can you try fly deploy --local-only and see if it works better?
What if flyctl by default ran deploy first with --remote-only, but with --local-only immediately after (on failure, that is)?
This may be irrelevant but…
I had an EOF issue when I deleted an app from the platform with a name like “myterribleapp” and then recreated another app but used exactly the same name again “myterribleapp”. This causes some confusion in the ip lookups or something.
When I used a unique name the eof error went away. This may not relate you but popping in case it solves an EOF error for anyone else.
Thanks for the advice–I ended up running fly deploy straight from the CLI because that was the shortest path to get my code deployed.
I also disabled the GitHub action, because right now, the project is just me, and having that action seems like overkill.
If this issue happens again, I’ll try debugging it with --local-only from the CLI and report back. Thank you!
– Doug
