1 |
Am Freitag 06 Januar 2012, 17:07:20 schrieb Alex Alexander: |
2 |
> On Fri, Jan 06, 2012 at 08:35:32AM -0500, Rich Freeman wrote: |
3 |
> > On Thu, Jan 5, 2012 at 5:36 PM, Alex Alexander <wired@g.o> wrote: |
4 |
> > > If people are really interested in keeping a tight, self contained |
5 |
> > > root, we need to: |
6 |
> > > |
7 |
> > > - establish a [tight] list of software we consider critical for / |
8 |
> > > - fix/patch software in that list so it can run without /usr there |
9 |
> > > - create /bin => /usr/bin/ symlinks for above software (simplifies |
10 |
> > > things if packages start hardcoding /usr/bin here and there) |
11 |
> > > - move everything else in /usr/bin/ |
12 |
> > |
13 |
> > You're missing one thing: |
14 |
> > |
15 |
> > - establish a list of all the configurations that will actually work |
16 |
> > with this self-contained root |
17 |
> > |
18 |
> > I think this is why there is so much disagreement over whether this is |
19 |
> > a good move. If you have a really simple configuration, then the |
20 |
> > self-contained root concept works reasonably well (though apparently |
21 |
> > we'll have to heavily patch newer versions of udev or abandon it to |
22 |
> > sustain this). |
23 |
> > |
24 |
> > However, if you have a very complex configuration the current |
25 |
> > self-contained root is already broken and you need an initramfs |
26 |
> > anyway. For in-between cases things might work now but that is likely |
27 |
> > to change as upstream moves on. |
28 |
> > |
29 |
> > The binary distros don't have users tweaking their kernels and init |
30 |
> > scripts, so they basically have to design for worst-case. Gentoo can |
31 |
> > get away with designing for more of an average case since we just tell |
32 |
> > anybody with a complex case to go read a howto and configure what is |
33 |
> > necessary (and we like to do that stuff anyway). |
34 |
> > |
35 |
> > We can choose not to like it, but it sounds like maintaining a |
36 |
> > self-contained root for even the typical case will become untenable. |
37 |
> > Those who argue that having /usr on a separate partition simply |
38 |
> > shouldn't be supported are basically just saying that our |
39 |
> > "self-contained root" should include everything in /usr which seems to |
40 |
> > defeat the whole point of a "self-contained root" anyway. |
41 |
> > |
42 |
> > It seems to me that the most reasonable approach is to not force the |
43 |
> > issue, but not deviate greatly from upstream either. That means |
44 |
> > accepting that over time the rootfs will become less and less capable |
45 |
> > of working on its own, and immediately improving tools like dracut to |
46 |
> > overcome these limitations. Users who can get away with it can avoid |
47 |
> > using an initramfs, at least for a time. |
48 |
> > |
49 |
> > Sure, it is all open source, and Gentoo can swim upstream if we REALLY |
50 |
> > want to. However, this only works if developers are willing to spend |
51 |
> > the time constantly fixing upstream's tools. It sounds to me like the |
52 |
> > maintainers of packages like udev/systemd/etc want to actually move in |
53 |
> > the same direction as upstream so in practice I don't see that |
54 |
> > happening. |
55 |
> > |
56 |
> > Now, Gentoo is about choice, so one thing we should try to do as much |
57 |
> > as possible is understand the limitations of the various |
58 |
> > configurations and make it clear to users when they do and don't need |
59 |
> > an initramfs. To be honest, tight coupling worries me more than the |
60 |
> > /usr move, since that has a lot more potential to constrain the |
61 |
> > choices we can offer our users (which is a great deal of the value |
62 |
> > that Gentoo offers). I understand its advantages, but it seems |
63 |
> > somewhat contrary to "the unix way." |
64 |
> |
65 |
> That's why I wrote "tight list". I do not expect the self-contained root |
66 |
> to be able to handle everything /usr (or a complete initramfs) would. |
67 |
> What it could and couldn't do is something that needs to be decided, but |
68 |
> some work is already done there - it's just a bit messy and incomplete |
69 |
> and because most people don't care it keeps getting worse. |
70 |
> |
71 |
> The important thing here is to make a clear definition of where we draw |
72 |
> the line and make sure things work the way we want them to. |
73 |
> |
74 |
> I agree with you in that at some point patching may become too time |
75 |
> consuming, but I still believe that if we do this properly, with a |
76 |
> well-defined plan and list of packages we want to keep in / (with |
77 |
> symlinks to be compatible with whatever others are trying to do), we |
78 |
> won't be alone in this. Gentoo may be one of the most hardcore |
79 |
> power-user distros out there, but we're certainly not the only one. |
80 |
> |
81 |
> Now, if only we had people interested enough in doing this... :) |
82 |
|
83 |
I'm interested in everything which prevents such kind of insanity from Gentoo. |
84 |
So count me in as volunteer keeping /bin, /sbin and /lib in Gentoo and systemd |
85 |
away from it. |
86 |
As long as we keep trying and working hard I'm sure we can do the workload |
87 |
that might come across when blind upstreams start adopting those foolish |
88 |
systemd/GnomeOS ideas... |
89 |
|
90 |
-- |
91 |
Lars Wendler (Polynomial-C) |