Gentoo Archives: gentoo-user

From: Daniel Wagener <stelf@×××.net>
To: gentoo-user@l.g.o
Subject: Re: [gentoo-user] Re: Anyone switched to eudev yet?
Date: Mon, 24 Dec 2012 22:12:50
Message-Id: 20121224221235.779AE21C0B7@pigeon.gentoo.org
In Reply to: Re: [gentoo-user] Re: Anyone switched to eudev yet? by Dale
1 On Mon, 24 Dec 2012 14:00:39 -0600
2 Dale <rdalek1967@×××××.com> wrote:
3
4 > Alan McKinnon wrote:
5 > > On Mon, 24 Dec 2012 06:58:15 -0600
6 > > Dale <rdalek1967@×××××.com> wrote:
7 > >
8 > >> So, Nuno, everything was fine until they started moving things to a
9 > >> place where it shouldn't be.
10 > > No Dale, that is just flat out wrong.
11 > >
12 > > There is no such thing as "place where stuff should be". There are
13 > > only conventions, and like all conventions, rituals, fashions and
14 > > traditions these are prone to breakage when things move on. Things
15 > > move on because they become way more complex than the designer of
16 > > the convention thought they would (or could).
17 > >
18 > > The truth is simply this (derived from empirical observation):
19 > >
20 > > Long ago we had established conventions about / and /usr; mostly
21 > > because the few sysadmins around agreed on some things. In those
22 > > days there was no concept of a packager or maintainer, there was
23 > > only a sysadmin. This person was a lot like me - he decided and if
24 > > you didn't like it that was tough. So things stayed as they were
25 > > for a very long time.
26 > >
27 > > Thankfully, it is not like that anymore and the distinction between
28 > > / and /usr is now so blurry there might as well not be a
29 > > distinction. Which is good as the distinction wasn't exactly a good
30 > > thing from day 1 either - it was useful for terminal servers (only
31 > > by convention) and let the sysadmin keep his treasured uptime
32 > > (which only proves he isn't doing kernel maintenance...)
33 > >
34 > > I'm sorry you bought into the crap about / and /usr that people of
35 > > my ilk foisted on you, but the time for that is past, and things
36 > > move on. If there is to be a convention, there can be only one that
37 > > makes any sense:
38 > >
39 > > / and /usr are essentially the same, so put your stuff anywhere you
40 > > want it to be. ironically this no gives you the ultimate in choice,
41 > > not the false one you had for years. So if your /usr is say 8G, then
42 > > enlarge / bu that amount, move /usr over and retain all your mount
43 > > points as the were. Now for the foreseeable future anything you
44 > > might want to hotplug at launch time stands a very good chance of
45 > > working as expected.
46 > >
47 > > You will only need an initrd if you have / on striped RAID or LVM or
48 > > similar, but that is a boot strap problem not a /usr problem (and
49 > > you do not have such a setup). Right now you need an initrd anyway
50 > > to boot such setups.
51 > >
52 > > The design of separate / and /usr on modern machines IS broken by
53 > > design. It is fragile and causes problems in the large case. This
54 > > doesn't mean YOUR system is broken and won't boot, it means it
55 > > causes unnecessary hassle in the whole ecosystem, and the fix is to
56 > > change behaviour and layout to something more appropriate to what
57 > > we have today.
58 > >
59 >
60 > The problems with that is these: It worked ALL these years, why
61 > should it not now? I have / on a traditional partition which is not
62 > going to resize easily. If I put / on LVM, I need a init thingy. I
63 > don't want a init thingy or I would have put / on LVM too. I made /
64 > large enough that I would not fill it up in the lifetime of this
65 > system but not large enough to absorb /usr. If I am going to have to
66 > redo all my partitions yet again, I will not use LVM. I use LVM to
67 > eliminate this EXACT problem. I got tired of running out of space
68 > and having to move stuff around all the time.
69 >
70 > So, worked for ages, then it breaks when people change where they put
71 > things. Answer is, don't change where you put things. Then things
72 > still work for most everyone, including me. I'm not a programmer nor
73 > am I a rocket scientist but even I can see that. If I can see it, I
74 > have no idea why a programmer can't other than being willingly
75 > blinded. ;-)
76 >
77 > Udev/systemd seems to be the problem. How do I come to that
78 > conclusion, eudev people says they will support separate /usr with no
79 > init thingy. Either the eudev folks are rocket scientist type
80 > programmers and the udev/systemd people are playing with fire
81 > crackers or there is a way for this to work with udev/systemd to, IF
82 > they wanted it to work. Thing is, they have some grand scheme to
83 > force people to their way of doing things, which includes a init
84 > thingy. Since there is a way to continue with the old way, which has
85 > worked for decades, guess what I am going to do? Yep, I'm going to
86 > jump off the udev ship and onto the eudev ship. The eudev ship may be
87 > old and traditional but it works like I expect. Now if others want to
88 > stay on the current ship, works for me too. I'm just not liking the
89 > meals served on the udev ship anymore.
90 >
91 > I might add, one of the reasons I left Mandriva was because of the
92 > init thingy that kept giving me grief. If I have to use that thing on
93 > Gentoo, the first time it breaks, I'm going to a binary install. If I
94 > am going to put up with that mess, I may as well have something that
95 > installs quickly. That was one thing I liked about Mandriva, install
96 > was really easy. It still is. Ubuntu is too. Actually, they look a
97 > lot alike to me.
98 >
99 > Everyone can have their opinion but I also have mine. This worked
100 > fine for ages until udev/systemd came along. That's my opinion and I
101 > don't think I am alone on that.
102 >
103 > Dale
104 >
105 > :-) :-)
106 >
107
108 Is this actually about something being broken like in "the code does
109 not do what it is supposed to" or about something no longer being the
110 tool of choice for everyone?
111
112 --
113 () ascii ribbon campaign - against html e-mail
114 /\ www.asciiribbon.org - against proprietary attachments