Rudder

some nodes were in error; and process rudder-cf-serverd o nthe server was not running
last logs were

Jul 29 14:03:35 rudder cf-serverd[1677593]: rudder     info: Path does not exist, it's added as-is in access rules: /var/rudder/shared-files/root/
Jul 29 14:03:35 rudder cf-serverd[1677593]: rudder     info: WARNING: this means that (not) having a trailing slash defines if it's (not) a directory!
Jul 29 14:03:35 rudder cf-serverd[1677593]: rudder     info: Failed to canonicalise filename '/var/rudder/shared-files/root/' (realpath: No such file or directory)
Jul 29 14:03:35 rudder cf-serverd[1677593]: rudder     info: Path does not exist, it's added as-is in access rules: /var/rudder/shared-files/root/
Jul 29 14:03:35 rudder cf-serverd[1677593]: rudder     info: WARNING: this means that (not) having a trailing slash defines if it's (not) a directory!
Jul 29 14:03:35 rudder systemd[1]: rudder-cf-serverd.service: Deactivated successfully.
Jul 29 14:03:36 rudder systemd[1]: rudder-cf-serverd.service: Scheduled restart job, restart counter is at 7.
Jul 29 14:03:36 rudder systemd[1]: Stopped CFEngine file server.
Jul 29 14:03:36 rudder systemd[1]: rudder-cf-serverd.service: Start request repeated too quickly.
Jul 29 14:03:36 rudder systemd[1]: rudder-cf-serverd.service: Failed with result 'start-limit-hit'.
Jul 29 14:03:36 rudder systemd[1]: Failed to start CFEngine file server.

running agent didn't change anything
rudder agent check didn't change anything

This component is essential for rudder, it should be started on the rudder server

EDIT: we are not allowed to force a service up if an user chose to stop it. The only case that we need to manage correctly is "if the service crashed, then systemd manage correctly the restart of the service".