Bug 370033 – sys-apps/openrc-0.8.2-r1: start scripts in /etc/local.d in the background

• Product: Gentoo Linux
• Status: RESOLVED INVALID

[Description Toralf Förster 2011-06-04 12:18:16 UTC]

Otherwise "/etc/init.d/local start" doesn't return and therefore there's no login possible at vt1-6.


Reproducible: Always

[Comment 1 Jeroen Roovers (RETIRED) 2011-06-07 20:28:50 UTC]

Why wouldn't it return?

[Comment 2 Toralf Förster 2011-06-07 20:57:57 UTC]

(In reply to comment #1)
> Why wouldn't it return?
I dunno - didn't wrote openrc :-)

If I don't put a "&" behind the call of my script then I do not have any login at vt1-6 :

tfoerste@n22 ~ $ cat /etc/local.d/i915_error_state.start

/home/tfoerste/workspace/bin/check_i915_error_state.sh &


I've these 3 sxcritps defined, the other 2 return immediately :

tfoerste@n22 ~ $ ls -l /etc/local.d/*
-rwxr-xr-x 1 root root 105 Jun  3 21:01 /etc/local.d/hibernated.start
-rwxr-xr-x 1 root root  59 Jun  7 22:53 /etc/local.d/i915_error_state.start
-rwxr-xr-x 1 root root 239 Jun  3 21:02 /etc/local.d/power.start
-rw-r--r-- 1 root root 387 May 14 20:41 /etc/local.d/README


The script check_i915_error_state.sh indeed doesn't return (asap) instead it loops (sometimes for days) until an error occurred (and then it finished).

[Comment 3 SpanKY 2011-06-07 20:59:03 UTC]

if your local scripts dont background/detach themselves properly, that's a bug in your local scripts

[Comment 4 Toralf Förster 2011-06-08 07:34:00 UTC]

(In reply to comment #3)
> if your local scripts dont background/detach themselves properly, that's a bug
> in your local scripts
Well - works I designed I would accept - OTOH I'm wondering why openrc doesn't handle this in a manner that a system isn't in a state where (w/ X) login is not possible.

[Comment 5 SpanKY 2011-06-09 01:41:31 UTC]

i dont think openrc should be going out of it's way to workaround broken user code.  in the common case (which is 99.9% of the time), it isnt necessary and uselessly slows things down.