1 |
On Wed, May 22, 2013 at 9:39 PM, Daniel Campbell <dlcampbell@×××.com> wrote: |
2 |
> On 05/20/2013 10:34 PM, Canek Peláez Valdés wrote: |
3 |
>> On Tue, May 21, 2013 at 3:03 AM, Daniel Campbell <dlcampbell@×××.com> wrote: |
4 |
>>> On 05/19/2013 01:05 PM, Canek Peláez Valdés wrote: |
5 |
>>>> On Sun, May 19, 2013 at 9:34 AM, Peter Stuge <peter@×××××.se> wrote: |
6 |
>>>>> J. Roeleveld wrote: |
7 |
>>>>>> I don't see how this will avoid the issue of a limited amount of |
8 |
>>>>>> inodes. |
9 |
>>>>>> That is what I usually run out of before the disk is full when |
10 |
>>>>>> storing lots of smaller files. |
11 |
>>>>> |
12 |
>>>>> I guess the number of unit files is on the order of hundreds |
13 |
>>>> |
14 |
>>>> (Sorry, sent email before it was ready). |
15 |
>>>> |
16 |
>>>> Laptop running full GNOME: |
17 |
>>>> |
18 |
>>>> # find /usr/lib/systemd/system -type f | wc |
19 |
>>>> 154 154 7012 |
20 |
>>>> |
21 |
>>>> Server running Apache+MySQL+Mailman+Squid+Other services: |
22 |
>>>> |
23 |
>>>> # find /usr/lib/systemd/system -type f | wc |
24 |
>>>> 121 121 5560 |
25 |
>>>> |
26 |
>>>> And as you said, you can always use INSTALL_MASK. If 154 files are |
27 |
>>>> going to deplete your inodes, I think your problem lies somewhere |
28 |
>>>> else. |
29 |
>>>> |
30 |
>>>> Regards. |
31 |
>>>> -- |
32 |
>>>> Canek Peláez Valdés |
33 |
>>>> Posgrado en Ciencia e Ingeniería de la Computación |
34 |
>>>> Universidad Nacional Autónoma de México |
35 |
>>>> |
36 |
>>> |
37 |
>>> That's missing the point. If you don't run systemd, having unit files is |
38 |
>>> pointless. Thankfully there's INSTALL_MASK and whatnot, but that seems |
39 |
>>> like a hack instead of something more robust. Why include systemd unit |
40 |
>>> files (by default, with no systemd USE flag, thanks to the council...) |
41 |
>>> on a system that's not using it? 154 files isn't negligible unless |
42 |
>>> you're flippant with your system and don't care about bloat. Unused |
43 |
>>> software sitting around *is* a waste of disk-space. |
44 |
>> |
45 |
>> Unit files are not software; they are data. |
46 |
>> |
47 |
>> And I believe you are the one missing the point. I don't run OpenRC; I |
48 |
>> don't need no files in /etc/init.d. But you don't see me (nor any |
49 |
>> other systemd user) complaining about pointless scripts in |
50 |
>> /etc/init.d. I just put /etc/init.d in INSTALL_MASK and go on with my |
51 |
>> life. |
52 |
>> |
53 |
>> Non-systemd users should do the same for files under /usr/lib/systemd, |
54 |
>> if they really are that worried about systemd "infecting" their |
55 |
>> systems. Complaining about a council-decided policy (and, I believe, |
56 |
>> backed up by the developers that matter, including the OpenRC |
57 |
>> maintainers) is just beating on a dead horse. |
58 |
>> |
59 |
>> Get over it. |
60 |
>> |
61 |
>>> Some people (like myself) came to Gentoo to avoid putting systemd on |
62 |
>>> their systems and to make use of the great choice that Gentoo allows. |
63 |
>>> This push to make systemd a "first level citizen" or whatever reeks of |
64 |
>>> marketing. |
65 |
>> |
66 |
>> If Gentoo is about choice, then systemd is one of those choices. And |
67 |
>> systemd will become a first class citizen inside Gentoo, like it or |
68 |
>> not. Support for it has been getting better and better, and more and |
69 |
>> more Gentoo users are running with systemd. |
70 |
>> |
71 |
>> If some fundamentalists users don't want even one file in their |
72 |
>> systems with "systemd" on their paths, they can install eudev/mdev, |
73 |
>> put the necessary directories in INSTALL_MASK, and do the extra work. |
74 |
>> If some other fundamentalists users (like myself) don't want even one |
75 |
>> OpenRC related file on our systems, we can create an overlay to remove |
76 |
>> the dependency of baselayout on OpenRC, put /etc/init.d in |
77 |
>> INSTALL_MASK, and do the extra work. |
78 |
>> |
79 |
>> Neither case covers the average systemd/OpenRC user, who doesn't care |
80 |
>> about a few scattered files in /etc/init.d nor /usr/lib/systemd, and |
81 |
>> just want to run her machine with the init system of her choice. If |
82 |
>> Gentoo is really about choice. |
83 |
>> |
84 |
>>> If there is desire among users for unit files, they can |
85 |
>>> contact upstream or maintain their own set of unit files. It's not like |
86 |
>>> they're hard to write. |
87 |
>> |
88 |
>> So, Gentoo is about choice, but only for the choices you agree with. Great. |
89 |
>> |
90 |
>> Regards. |
91 |
>> -- |
92 |
>> Canek Peláez Valdés |
93 |
>> Posgrado en Ciencia e Ingeniería de la Computación |
94 |
>> Universidad Nacional Autónoma de México |
95 |
>> |
96 |
> |
97 |
> It seems that I've stepped on a few toes in calling INSTALL_MASK a hack. |
98 |
> Hacks aren't necessarily bad; if anything it shows that there's interest |
99 |
> in supporting something but perhaps not enough time or manpower to |
100 |
> implement a more robust solution. If adding one or two directories to |
101 |
> that variable will nuke any unit files, consider me happy. |
102 |
|
103 |
As I was, when I used to put /etc/init.d in INSTALL_MASK. |
104 |
|
105 |
> systemd is certainly a choice, but it is no more deserving of |
106 |
> consideration than any other init system. I don't see anyone calling for |
107 |
> runit to be a 'first level citizen'. I wonder why that is. |
108 |
|
109 |
Probably because is used by a really small number of users, contrary to systemd |
110 |
|
111 |
> Again, if |
112 |
> INSTALL_MASKing openrc dirs will get rid of init scripts for systemd |
113 |
> users, then perhaps INSTALL_MASK is the best we have for now and should |
114 |
> make use of it. I never said that it wasn't suitable to use. |
115 |
|
116 |
Then we agree. |
117 |
|
118 |
> As for "complaining" about policy, what is the proper thing to do in a |
119 |
> situation where someone questions the reasoning behind a decision? |
120 |
|
121 |
Contribute? |
122 |
|
123 |
> Are |
124 |
> there links somewhere on Gentoo's website that outline the process for |
125 |
> each important decision that the council's made? |
126 |
|
127 |
The Council meetings logs are in the Council project page: |
128 |
|
129 |
http://www.gentoo.org/proj/en/council/ |
130 |
http://www.gentoo.org/proj/en/council/#doc_chap6 |
131 |
|
132 |
> I think it'd be |
133 |
> valuable information for people and keep individuals like you from |
134 |
> telling others to "get over it" without any explanation whatsoever. |
135 |
> That's not communication, that's prescription. |
136 |
|
137 |
I think I gave an explanation when I said "Complaining about a |
138 |
council-decided policy (and, I believe, backed up by the developers |
139 |
that matter, including the OpenRC maintainers) is just beating on a |
140 |
dead horse." The important thing being that the ones writing the code |
141 |
(including the OpenRC maintainer) already made their choice. |
142 |
|
143 |
> I'm curious as to why you consider users who want to save disk space |
144 |
> (openrc or systemd, or other packages, it doesn't matter) as |
145 |
> fundamentalists. |
146 |
|
147 |
Because wanting to eradicate some dozens of files that take not even |
148 |
one megabyte (combined!) is, in my (updated) view, fundamentalist. |
149 |
|
150 |
When I wrote the overlay to get rid of OpenRC it was because (back |
151 |
then, a couple of years ago), some udev rules sometimes would try to |
152 |
call scripts in /etc/init.d, and then they would try to start all of |
153 |
OpenRC's services, slowing down (sometimes for close to a minute) the |
154 |
booting time. Now the situation has improved greatly, and systemd on |
155 |
Gentoo works great; I would even venture to say that works better than |
156 |
in Fedora. Now, thanks to the work of people like Fabio Erculiani, |
157 |
both init systems can be installed in parallel without stepping in |
158 |
each other toes, and at boot time you can decide which one to use. |
159 |
|
160 |
I recently converted a server from OpenRC to systemd, and it was the |
161 |
first time I didn't use my overlay. The switch was completely |
162 |
painless, and I decided that putting /etc/init.d in INSTALL_MASK was |
163 |
kinda stupid. It makes no sense (I believe, now) to remove 51 files |
164 |
that utilize 232K just so the server is "100% OpenRC-free". I think I |
165 |
should be able to uninstall OpenRC (and work is progressing in that |
166 |
direction), but I no longer care about files in /etc/init.d. |
167 |
|
168 |
I got over it. |
169 |
|
170 |
> I do not consider Gentoo to be only about my own choices, but as a user, |
171 |
> who else's choices am I going to consider when I administer my system? |
172 |
|
173 |
Who took away any choice from you? |
174 |
|
175 |
> I'm happy for any new choices as long as they don't step on mine. I |
176 |
> think that's fair. |
177 |
|
178 |
What choices of yours were stepped on? |
179 |
|
180 |
Regards. |
181 |
-- |
182 |
Canek Peláez Valdés |
183 |
Posgrado en Ciencia e Ingeniería de la Computación |
184 |
Universidad Nacional Autónoma de México |