Postgres Flex machine stuck in “starting” state and cannot be stopped or restarted

My Postgres Flex machine (flyio/postgres-flex:15.8) is stuck in the “starting” state with 0/3 health checks, and I’m unable to stop, restart, or kill it. Commands like fly machines restart, fly machines stop --signal SIGKILL, and fly machines start all fail because Fly reports that the machine is still “starting.” The UI also shows it permanently stuck in this state.

  • Attempts to force stop or restart return errors like:

    “machine still restarting”

    “unable to stop machine, current state invalid, starting”

I need help either force-stopping this machine or recovering the volume so I can get Postgres running again.

Experiencing issues too in region ARN. Machine was suddenly stopped and when trying to start it I get “Machine failed to start: Unknown response while starting machine“. Deployment via flyctl is stuck at “Waiting for depot builder…“.

There was a glitch with the Machines API during that time frame, which would be consistent with the errors that both you and @Oriol saw.

It’s reported as officially resolved now, so give it another try if you haven’t already.

(And if it did work, then OP could mark the thread here as Solved, just as a small touch of courtesy.)

I confirm that everything is working as expected now.

1 Like

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.