This app has been running for several weeks, but after a deploy today all deployments for any machine associated with the app fail repeatedly shutting down and exiting abruptly. The logs just contain repeated instances of
2026-01-14T00:20:40Z app[287e126ce1d678] sjc [info]ERROR Error: an unhandled IO error occurred: No such file or directory (os error 2)
2026-01-14T00:20:40Z app[287e126ce1d678] sjc [info][ 0.878594] reboot: Restarting system
2026-01-14T00:20:40Z app[287e126ce1d678] sjc [warn]Virtual machine exited abruptly
I believe this to be a problem on the fly side, as I’ve pulled the image itself from the fly registry and it runs locally. I also made a dramatic change and now the entrypoint of the Dockerfile is /bin/sh which again, runs locally when the image is pulled from the fly.io registry.
Any ideas what I can do on my side to recover?
Hi… I would try cloning a new Machine, to see whether it fares any better…
(This assumes that your app can handle having an extra Machine, which usually is the case on the Fly.io platform.)
You can also try deploying an older image, with fly releases --image and then fly deploy --image <long-id-from-that-list>.
(This one assumes that you haven’t made any incompatible fly.toml changes, or the like, in the interim.)
Finally, it’s generally advisable to post more of the logs. (Preferably the entire span from a Machine start to the corresponding Machine exit.) Forum readers might see a pattern they recognize.
Hope this helps a little!
1 Like
Thank you for your help! I wasn’t able to get the machine launching via clone or re-deploying because the image would never update. But I was able to get it to work by forcibly destroying the offending machines and relaunch the app with an earlier image. I was also able to wedge the deploy again with a new image, so I have a direction to look on my side.
Thanks again!
1 Like