Hello, I have a mumble server running as a docker container.
I'm using lego to generate let's encrypt certificates using dns challenge as an unprivileged user.
My problem is that I need to restart the mumble server so that it can reload the ssl certificate, which means restarting the docker container.
Is there any way to do this withuot adding the user to docker group, since there are security issues with that group where it's easy to escalate to a root user.
I thought about maybe a user level systemd service could be able to do it, but wan't to hear your opinions?
EDIT: I mention mumble specifically here, but this can apply to any server that requires manual reload of ssl keys, like postfix for example. I'm also not againd the idea of converting to something like kubernetes or even podman ( which would be harder cause of file permissions on all my binds ) if they can allow for woking on the container without root privileges.
I think you can send a SIGUSR1 signal to mumble process to tell it to reload the ssl certificate without actually restarting mumble's process. You can use docker kill --signal="SIGUSR1" <container name or id>, but then you still need to give your user access to docker group. Maybe you can setup a monthly cron on root user to run that command every months?
Did you think about automating the certificate generation? You could put everything into a cron script. That's more or less the reason why certificates have such a short lifetime. To force automation.
I am using it in a cron script, but the mumble server doesn't support automatic reload of the ssl certs, so I need to restart tge mumblr server to load the new certs. That's where my problem comes cause I run it inside docker.
Not sure how your stack works together, but sudo will let you run particular commands as a different user and you can be pretty specific with the privileges. For example you can have a script that’s only allowed to run docker compose -f /path/to/compose.yml restart containername as a user in the docker group. Maybe there’s some docker-specific approach, but this should work with traditional Unix tools and a little scripting.
Kubernetes has user accounts that you could use to restart containers in an unprivileged way. Create a role and role binding that gives the "delete Pod" permission to a service account. Kind makes it very easy to run Kubernetes without any setup. You'd just need to convert your docker compose files to Deployments, Services, and PersistentVolumes.
If converting to a kubernetes setup is too big of a leap, you could maybe try to write a C program that uses setuid to gain docker privileges in a restricted way.
Probably easiest to just have a cronjob that restarts the container regularly, though.
Yeah, there's definitely a learning curve since it's so different from docker, but there are some good tutorials and everything just makes sense. All the error messages are googlable and everything fits together so well.