I had bug reports about restarting daemon because of changes in sysconfig file. Imagine that you have cron daemon running on cluster or with home's through LDAP. After every restart is possible that something goes wrong like LDAP user's won't be found. Also I don't think it's wise reloading all daemons. Some could take new options immediately, some after restart,... I will have troubles only with cron, but what about more important services, where system administrator have to restart daemons? Please revert this behaviour. Reproducer: change /etc/sysconfig/crond systemctl --system daemon-reload service crond restart Version: systemd-15-1.fc15.x86_64
Hmm, we probably should read that information when spawning a service, not already when loading the configuration.
systemd git now reads those env config files at process spwan time, not already when systemd reads its own configuration. This should resolve the issue. Upload to F15 follows shortly.