Docker restart on CLI?

[Micky]

Anybody knows how to restart docker from shell? Not the containers but the package itself.

And next question: anybody knows where I can find daemon.json that is loaded by the docker package or when not existing already where I've to create this file for docker?

 

[Rusty]

Anybody knows how to restart docker from shell? Not the containers but the package itself.

synoservice --restart pkgctl-Docker

Also you can use start and stop instead of restart

 

[one-eyed-king]

And next question: anybody knows where I can find daemon.json that is loaded by the docker package or when not existing already where I've to create this file for docker?

The Syno docker package stores the file in a different path with a different name: /var/packages/Docker/etc/dockerd.json

 

[Micky]

Thanks to both!

 

[mcnahum]

I was looking exactly the same but "synoservice" doesn't look to be command that exist on my Synology .. Am I missing something there?

 

[Micky]

Maybe you didn't switch to root user after opening the SSH console?

 

[Rusty]

I was looking exactly the same but "synoservice" doesn't look to be command that exist on my Synology .. Am I missing something there?

If you are on dsm7 command is different. Examples:

usage: synopkgctl <command> [...]

command:
    enable <package>
    disable <package>
    setup <package>
    start <package>
    stop <package>
    teardown <package>
    check-startable <package>
    
systemctl restart pkgctl-WebStation
systemctl restart nginx

 

[mcnahum]

Thanks! ---> sudo synopkgctl stop Docker / sudo synopkgctl start Docker did it

 

[otravers]

synopkgctl stop Docker stopped working for me months ago (around DSM 7.1 maybe?), instead I get this error:

Failed to load info [0x0000 (null):0]

I'm not sure it should list docker, put I noticed that synopkg list --name doesn't include docker in the results, even though Docker is definitely running on my DS920+

It was convenient to be able to restart the entire docker daemon as I have a couple containers that sometimes turn into unresponsive but unkillable zombies.

 

[one-eyed-king]

Try sudo systemctl restart pkg-Docker-dockerd

Update: with Container Manager it is sudo systemctl restart pkg-ContainerManager-dockerd

If the service should ever change its name, this command should help to identify the new name: systemctl list-units --type=service --all | grep -i docker

 

[one-eyed-king]

It was convenient to be able to restart the entire docker daemon as I have a couple containers that sometimes turn into unresponsive but unkillable zombies.

You should fix the images maintained by you, and raise an issue in the github repository for images of other maintainers.

See Dockerfile reference for differences between the shell form and exec form.
The shell form implicitly wraps the command with /bin/sh or /bin/bash. Some image maintainers explicitly use sh or bash in their ENTRYPOINT and/or CMD, or use the exec form to run a shell script, but don't hand over pid1 to the main process. As a result the SIGTERM signal does not reach the main process, and it can not be gracefully stopped. Some processes require a different signal for termination - it should be described in the documentation of the application/service.

 

[jucor]

Try sudo systemctl restart pkg-Docker-dockerd

Update: with Container Manager it is sudo systemctl restart pkg-ContainerManager-dockerd

If the service should ever change its name, this command should help to identify the new name: systemctl list-units --type=service --all | grep -i docker

Thanks @one-eyed-king so much for this, just helped fix a problem I'd been banging my head against today