Docker container with Samba exits without any message

dockersamba

We have a server running CentOS, on which there are several services in Docker containers to separate them from host changes.

Last week Friday we ran an update on CentOS, and since then one of our services does start and immediately exits. The other 15+ services are fine.

I cloned the container setup, changed the container name and rebuild with the same problem: container starts and then exits without any message.

The container is derived from David Personette's samba container. The actual command that exits now but did not use to is:

exec ionice -c 3 smbd -FS

Should I downgrade docker will that have side effects? Is there anything else I can do to get this running again? We chose containers to abstract from the operating system, is that abstraction not as good as expected?

Best Answer

The container will exit without any comment, unless you increase the debug level. You can use something like --debuglevel=4.

Once you do that, rebuilt and restart the container you will should get something at the end like:

Server exit (EOF on stdin)

And that points to the solution, change the smbd invocation to:

exec ionice -c 3 smbd -FS < /dev/null

and your container should no longer directly exit.

I am not sure what causes this backwards incompatible change, nor why this was not tested against existing dockerhub entries.

I had two of my containers break as well on upgrading to 1.11. and on downgrading to 1.10.3, ran into the dreaded Cannot stop container .... Container does not exist: container destroyed bug. I had to stop docker, clear all of the history, so I cannot really recommend that.

Related Question