I had the app running with no issues. then decided to start over after cleaning up some of the configuration and code.
I deleted all apps/clusters and started over by using fly launch command.
Now every time I deploy it I can see the machine state as follows:
Machine starting • Source: user
Machine started • Source: flyd
Exit • Source: flyd
Destroying
Destroyed
Checking out the machine logs, there’s no errors in the logs.
Hm… This sounds like it’s just the release_command Machine, actually. Those get created and then auto-destroyed as part of the normal deploy process.
It would probably help to post your full fly.toml and Dockerfile. You can use the </> button in the toolbar to get an area suitable for pasting code, etc.
We readers here in the community forum generally can’t poke around in your app settings, etc., so we mostly only know what you tell us yourself…
Hm… When you run fly deploy from the command line, do you see the message release_command <hex-string> completed successfully? (If that fails, then it won’t proceed to any real (i.e., non-ephemeral) Machines.)
Also, I would try fly m list -a dinestack-saas, since the dashboard can be a little quirky sometimes…
This usually happens when the app starts and then immediately stops running. Fly sees that the process exited, so it just shuts the machine down. Even if there are no errors in the logs, the app might be finishing instantly — for example, if the start command is wrong or it’s running something that doesn’t stay alive (like a script instead of a server). I’d double-check the start command in your fly.toml. Make sure it’s actually starting a server that keeps running. You can also try running that same command locally and see if it exits right away. Another common issue is the app not listening on the right port or not binding to 0.0.0.0. In most cases, it’s not Fly destroying it randomly, it’s just the app stopping, so Fly cleans it up.