[LEDE-DEV] procd and self-daemonizing processes with no useful foreground option

Daniel Dickinson lede at daniel.thecshore.com
Tue May 17 23:48:29 PDT 2016


Hi,

Rather than patching every package that daemonizes itself but doesn't
provide a useful (non-debug mode) option for foregrounding the process
(or not option at all for foreground operation), would it possible to do
as systemd has done and support both a directly supervised instance
(i.e. what it does now where the process is in the foreground and
daemonized by procd) as well as forked processes that daemonize themselves?

I've run into three so far that have this problem, and am thinking there
are probably more I just haven't encountered yet.

This is really the only bit of systemd that I've wished we had (although
"it'd be nice" to replace xinetd too with a .socket like thing).

I think that covers 90% or more of the cases where processes need to be
daemonized and managed.

Regards,

Daniel



More information about the Lede-dev mailing list