Gentoo Archives: gentoo-user

From: Alan Mackenzie <acm@×××.de>
To: gentoo-user@l.g.o
Subject: Re: [gentoo-user] separate / and /usr to require initramfs 2013-11-01
Date: Sat, 28 Sep 2013 19:06:33
Message-Id: 20130928190441.GB11317@acm.acm
In Reply to: Re: [gentoo-user] separate / and /usr to require initramfs 2013-11-01 by William Hubbs
1 Hi, William.
2
3 On Sat, Sep 28, 2013 at 11:01:59AM -0500, William Hubbs wrote:
4 > On Fri, Sep 27, 2013 at 07:32:20PM -0500, Bruce Hill wrote:
5 > > On Fri, Sep 27, 2013 at 05:57:06PM -0500, Dale wrote:
6 > > > Bruce Hill wrote:
7 > > > > On Fri, Sep 27, 2013 at 05:33:02PM -0500, Dale wrote:
8 > > > >> I'm hoping that since I use eudev, I don't have to worry about this.
9 > > > >> If I do, this could get interesting, again. Dale
10 > > > > Do you have /usr separate from / ?
11
12 > > > Yep. From my understanding tho, eudev is not supposed to be affected by
13 > > > this problem tho.
14
15 > > > One reason for this being seperate, I have / and /boot on a regular
16 > > > partition and everything else on LVM. Sometimes that /usr gets a bit
17 > > > full. It's not so bad after I moved all the portage stuff out and put
18 > > > it in /var. Now I have to watch /var too. lol
19
20 > > > Dale
21
22 > > You need to read the blog post listed in the news item, as it's not just
23 > > specific to udev anymore.
24
25 > Bruce is correct; This issue is not specific to udev/eudev/mdev.
26
27 > I suppose that what I am about to say isn't really relevant, but it is
28 > unfortunate over the past year that people blamed udev specifically for
29 > this. It is true that it does things that don't work if /usr isn't
30 > mounted, but eudev does as well, since it is basically the same code.
31
32 Who else is there to blame? We are continually being told that a
33 separate /usr is "broken", as though this were some unfortunate act of
34 <insert your deity here>, much like an earthquake. This gets patronising
35 really quickly. (Please note, I'm NOT blaming you here. I appreciate
36 that you're as much victim as Dale or me or anyone else round here.)
37
38 No, this breaking of separate /usr was done by some specific project,
39 some specific person, even, in a supreme display of incompetence, malice,
40 or arrogance. How come this project and this person have managed to
41 maintain such a low profile? There seems to have been some sort of
42 conspiracy to do this breakage in secret, each member of the coven
43 pushing the plot until the damage was irrevocable. Who was it?
44
45 > If you read flameeyes' blog post, you will get a better idea of what the
46 > issue involves. It is the entire boot process and how to deal with which
47 > software is considered critical for booting.
48
49 > There is no reason to rebuild your server; we aren't telling you you
50 > have to merge /usr into /. The only thing we are saying is that you will
51 > need to use an initramfs if you are going to keep them separate.
52
53 "Only"? ONLY??? You say that as though creating an initramfs were a
54 trifle, trivial, and of no moment.
55
56 An initramfs is a highly complicated, fragile contraption, and has all
57 the aesthetic appeal of a car crash. It is a desperate expedient, an
58 ugly kludge, made necessary (for binary distributions) by the design
59 deficiencies of the Linux kernel. Who in their right mind (other than a
60 specialist at a binary distribution) would want to spend evenings and
61 weekends battling this abortion just trying to get their machine to boot?
62 The alternative is to install some magic, effectively binary blob,
63 generated by genkernel or dracut or whatever. Who knows what these blobs
64 will do during booting?
65
66 Consider how ridiculous booting Linux is. Firstly, on power up, the bios
67 initialises then loads the program from the HDD's boot sector, namely
68 grub or lilo. This loads its main part. Then it loads the kernel,
69 which starts, then the init sequence. Each element of this sequence can
70 be individually justified, but the whole lot together just look
71 incompetent - why can't the kernel just start? And now, on top of all
72 this the conspirators want to force us to use an initramfs.
73
74 Ah yes, the deficiencies of the kernel. It can only mount one file
75 system when it starts. It's incapable of mounting LVM2 systems (even
76 though it contains LVM2 code). It's incapable of mounting encrypted
77 partitions (even though it contains encryption code), ...... So because
78 of these holes, a system must either be constrained in it's makeup (as
79 mine is) or use an ugly hack. It can (still) mount RAID partitions, I
80 suppose.
81
82 > I have a pretty simple setup, but I have been using an initramfs which I
83 > built some time ago with genkernel and I barely know it is there.
84
85 Until, after some update, it reminds you of its presence by not booting
86 your machine. That's the sort of excitement I can do without.
87
88 > I recommend that you familiarize yourself with genkernel or dracut and
89 > build an initramfs. Since nothing is changing until at least
90 > Nov 1, you can test your initramfs by adding an entry to your boot
91 > loader configuration that uses it and get it set up correctly while you
92 > can still fall back on booting without it.
93
94 > I do not recommend that anyone who has separate /usr "do nothing" at
95 > this point. Please re-read the second paragraph of the news item.
96
97 I dismantled my separate /usr partition some while ago in anticipation of
98 what has transpired. Previously, it was in an LVM2 partition, where I'd
99 prefer it still to be. Now, /usr is just in my root partition, /dev/md6.
100 At least RAID is still available.
101
102 > Thanks,
103
104 > William
105
106 --
107 Alan Mackenzie (Nuremberg, Germany).

Replies

Subject Author
Re: [gentoo-user] separate / and /usr to require initramfs 2013-11-01 Neil Bothwick <neil@××××××××××.uk>
Re: [gentoo-user] separate / and /usr to require initramfs 2013-11-01 Tanstaafl <tanstaafl@×××××××××××.org>