Runlevel-adjust crashes docker.service

[root@NS7DEV2 ~]# cat /etc/redhat-release
CentOS Linux release 7.3.1611 (Core)

Like I discovered when I launch the event runlevel-adjust, it disables and It stops the docker.service

I used the command line to allow the service

systemctl enable docker
systemctl start docker
config set docker service status enabled

I found that the sub is_owned crashes my service because I didn’t set a default entry in /etc/e-smith/db/configuration/defaults/docker

see in /usr/share/perl5/vendor_perl/NethServer/Service.pm

Maybe the sub is_owned could simply look after the status property…as a side not, if a service is enabled by the command line, then it is not displayed in the service page, once the defaults service name is created, then it is displayed as expected.

Not sure if it is a bug

This manual change to the DB creates a service record, but there’s no actually an RPM that “owns” it. Thus the runlevel-adjust stops the service. This behavior is required by package uninstall procedure. When an RPM is uninstalled, the configured services are stopped.

Absolutely not: it’s the expected behavior!

To fix your DB:

config delete docker

…this is not possible. A service must be “owned” by an RPM package or not exist at all!

However there exist and is documented a workaround for opening firewall ports to local services:

http://docs.nethserver.org/en/v7/third_party.html#firewall

1 Like