I don't use them so my opinion may not be important. I think ntploggps is moving to gpsd.
cron/systemd timers add a lot of clutter in their log files. So I would vote for a daemon. 5 minutes may not be fast enough to see thermal spikes. > nobody added Unix daemonization code (something something double-fork()), Unix has a daemon() call. I poked around a bit. Python has one too. The double-fork stuff is needed if you are opening TTYs and don't want the controlling tty, or something like that. So we may be able to skip that dance. Gary said: > Your distro, your choice. I see no reason why NTPSec cares how you run > it. Just that you do run it. I agree with Gary that we should support however you want to do it. I think systemd supports programs that don't daemionize themselves. (ntpd needs a Type=forking option) I don't support the "Just that you do run it" idea. This sort of data is for geeks. Most users would probably consider it clutter. Your distro, your rules. I'll be happy as long as you make it easy to turn on/off. ------- One complication with long running daemons... You want to break the log files into days or whatever so logrotate can do its thing. I think the program has to open a file rather than writing to stdout. Then it can reopen on SIGHUP. Or it can do like ntpd does for the stats files and include a date string in the log file name and link the root log file name to the current version. -- These are my opinions. I hate spam. _______________________________________________ devel mailing list [email protected] https://lists.ntpsec.org/mailman/listinfo/devel
