On Mon, 29 Apr 2019 21:19:58 +0200
Jeff <sysinit_at_yandex.com> wrote:
> i came across some interesting claims recently. on
> http://skarnet.org/software/s6/
> it reads
>
> "suckless init is incorrect, because it has no supervision
> capabilities, and thus, killing all processes but init can brick the
> machine."
Oh, that.
First of all, Suckless Init is a PID1 that forks an rc script and then
hangs around reaping zombies, but it's not an entire init system. You
could make it a complete init system by using the forked rc file to run
supervision systems such as daemontools-encore and the supervision part
of runit and s6. And of course you'd need a shutdown script that PID1
can call when it gets signals to reboot or poweroff. So Suckless Init is
the PID1 part of an init system. It's 83 lines of C. It's not an entire
init system.
There are three philosophies:
1) The supervision should be done by PID1: Supported by Laurent Bercot
2) The supervision should be done outside of PID1: Perhaps supported by
Rich Felker in his
http://ewontfix.com/14/ blog.
3) Either is acceptable and greatly superior to systemd, sysvinit,
upstart, etc. This is supported by most people who like process
supervision.
So Laurent's words from
http://skarnet.org/software/s6/ were just part
of a very minor family quarrel, not a big deal, and nothing to get
worked up over.
SteveT
Received on Wed May 01 2019 - 07:33:55 UTC