I opened a similar topic a long time ago, but I did not get to the bottom of it. More docker experience and testing over 2 days prompted me to revisit.

I am trying to deploy docker containers through ansible. This is working mosly as expected, and I can perform basic functions through the dockerman interface. However there is an issue on reboot.

I am using "unless-stopped" as restart-policy. The way it is supposed to work (confirmed by testing on other linux OS) is that the containers will restart on docker daemon restart, unless they were intentionally stopped with the docker stop command. However, whenever I reboot, all the containers are stopped after reboot.

If I switch to the "always" restart policy, all the containers will be started after reboot, even the one intentionally stopped. This works as expected.

Is it possible that Unraid is not applying the expected logic when restarting the docker daemon and is setting all containers to the "stopped" state? Like issuing docker stop command to all containers before rebooting?

On 9/10/2025 at 10:54 PM, PierreT said:

Is it possible that Unraid is not applying the expected logic when restarting the docker daemon and is setting all containers to the "stopped" state? Like issuing docker stop command to all containers before rebooting?

Yes I believe that the unRAID explicitly stops all containers before stopping the docker daemon.