1 |
Hey guys and gals, |
2 |
|
3 |
reports for the scheduled in the last week: |
4 |
|
5 |
ON-GOING 1. kill bugs for openrc |
6 |
|
7 |
https://bugs.gentoo.org/buglist.cgi?quicksearch=openrc&list_id=1113873 |
8 |
|
9 |
ON-GOING 1.1 bug 413267 |
10 |
|
11 |
reproduced on my box, investigating the built in "single" |
12 |
runlevel. |
13 |
|
14 |
TODO 1.2 bug 409829 |
15 |
|
16 |
ON-GOING 2. evaluate other rc systems (in order) |
17 |
|
18 |
DONE 2.0 smf (Openindiana) |
19 |
|
20 |
2.0.1 updated in |
21 |
|
22 |
https://wiki.gentoo.org/wiki/OpenRC |
23 |
|
24 |
later mgorny created a page for general comparison, |
25 |
|
26 |
http://wiki.gentoo.org/wiki/Comparison_of_init_systems |
27 |
|
28 |
Further evaluations will go to the new page. Old page needs |
29 |
to be cleaned up. |
30 |
|
31 |
TODO 2.1 launchd (Mac OS X) |
32 |
|
33 |
ON-GOING 2.2 upstart (Ubuntu) |
34 |
as well as trying Gentoo Prefix on top of it (there has |
35 |
been a big block for Prefix on Ubuntu, cf. bug |
36 |
400839). and chaining openrc from upstart. |
37 |
|
38 |
EXTRA NOTE 3. found another thread for introducing OpenRC to |
39 |
debian[1] |
40 |
|
41 |
3.1 in the reply by rra@debian[2], there are really what we are |
42 |
missing in openrc (ecosystem). |
43 |
|
44 |
3.2 rleigh@debian seems to represent the thinking of debian |
45 |
community towards init system[3]: they don't like systemd |
46 |
way of integrating everything. But they really need some |
47 |
event-driven mechanism, which is not well defined by OpenRC |
48 |
(yet). |
49 |
|
50 |
3.3 talked with rleigh via irc, conveying (as patrick did) we are |
51 |
eager to take ideas from debian, and impove OpenRC to suite |
52 |
debian's need provided it's of general interest. And I am |
53 |
eager to put my effort to carry ideas into reality. |
54 |
|
55 |
3.4 discussed with patrick in #openrc considering the thread, and |
56 |
agreed upon that we can extend OpenRC with some simple |
57 |
helper scripts (event-driven, crash restart/statistics, |
58 |
dbus-like API for triggers from outside) and kernel specific |
59 |
hooks on top of the present OpenRC core. Changing OpenRC |
60 |
itself to make a fancy feature (or fanboy's paradigm) is not |
61 |
good practice. |
62 |
|
63 |
In the next step, to improve OpenRC with recent ideas of init system, I |
64 |
am thinking of packaging OpenRC in debian and let it work out of box |
65 |
(maybe the LSB thing is the major block instead) as a practical measure |
66 |
of the usefulness of new features I would have introduced. (I myself is |
67 |
an old debian user for 8 years, therefore motivated to do so). |
68 |
|
69 |
rleigh and rra, could you please be my mentor for the packaging? |
70 |
|
71 |
Schedules for 6.25-7.1 |
72 |
====================== |
73 |
|
74 |
1. evaluating launchd, and finish upstart |
75 |
|
76 |
2. the openrc single/reboot built runlevel switching bug |
77 |
|
78 |
3. try to have a piece of script to prove my concept for using |
79 |
openrc as a "event-driven" init system. try to work with |
80 |
rleigh@debian (rra@debian), as they know what are really |
81 |
needed. and synchronize with openrc herd. |
82 |
|
83 |
Cheers, |
84 |
Benda |
85 |
|
86 |
1. http://thread.gmane.org/gmane.linux.debian.devel.general/171695 |
87 |
2. http://article.gmane.org/gmane.linux.debian.devel.general/171786 |
88 |
3. http://article.gmane.org/gmane.linux.debian.devel.general/171897 |