1 |
On Mon, Nov 19, 2012 at 5:07 AM, Steven J. Long |
2 |
<slong@××××××××××××××××××.uk> wrote: |
3 |
> On Sun, Nov 18, 2012 at 05:16:18PM +0200, Samuli Suominen wrote: |
4 |
>> I'm still happy enough with building udev out from systemd tree and |
5 |
>> letting sep. /usr consept from 90s to finally die in favour of |
6 |
>> simplifying the system. |
7 |
> |
8 |
> It's from a lot earlier than the 90s. Perhaps we should get rid of pipes, |
9 |
> too- they're /so/ 1970s. |
10 |
> |
11 |
> Torvalds expressed it well[1]: |
12 |
> You made the statement that you want to get away from 30 years of |
13 |
> history, but the fact is, "30 years of history" is a damn good argument |
14 |
> for "that thing worked". |
15 |
> |
16 |
>> The BIOSes have been upgraded last century to support booting from |
17 |
>> larger partitions, the need has long past. |
18 |
>> Nobody has ever provided a valid reason for using sep. /usr in the ML |
19 |
>> either. |
20 |
>> |
21 |
> So at the same time as you have never heard a valid reason, the need (which |
22 |
> you have never understood) is long past? Pfft. |
23 |
> |
24 |
> To reiterate again: many of us are used to keeping /usr on a separate |
25 |
> partition for security and backup purposes (other use-cases for a separate |
26 |
> partition include mounting /usr over NFS, or a common partition for |
27 |
> virtual machines.) I'm sure other people would have more. Irrespective, |
28 |
> it should be about mechanism, not policy, and certainly not about breaking |
29 |
> existing setups. |
30 |
> |
31 |
> Many of the advertised benefits of the "merge /bin and /sbin to /usr" |
32 |
> concept are in fact just benefits of having a separate /usr partition, |
33 |
> though of course they're presented as new and unique to the merge. So, if |
34 |
> you accept those as benefits, you cannot logically deny them as benefits of |
35 |
> a traditional /usr split. |
36 |
> |
37 |
> Further, as one user pointed out[2]: |
38 |
> "I'd be more likely to just stick /usr back on / than create an initramfs |
39 |
> - it was only because the LVM guide recommended that /usr go on its own |
40 |
> partition that I now face this situation." |
41 |
> |
42 |
> We were given a simple requirement ages ago: udev requires /usr and |
43 |
> /var (and going forward may require /opt for 3rd party drivers) mounted |
44 |
> before it starts, not for udev itself but for helper scripts that it may |
45 |
> call. Why get bogged down in anything else? |
46 |
> |
47 |
> The recommended (which became "supported") method to do this was |
48 |
> to use an initramfs, since apparently any chicken-and-egg problems (such |
49 |
> as requiring a udev-initiated device to mount /usr) are easier to handle |
50 |
> there. (Note that no-one has ever provided a description of how they did |
51 |
> that, eg for their bluetooth keyboard, so that the issues, and how they |
52 |
> are addressed, could be made more transparent to everyone.) |
53 |
> |
54 |
> A few us are happily running initramfs-less machines by fulfilling the |
55 |
> requirement with a couple of patches to openrc initscripts[2]. (It's |
56 |
> been working fine here since last August.) This works in the case where |
57 |
> you didn't need an initramfs before, ie have modules for local-disks and |
58 |
> filesystems built-in, root not on LVM or encrypted, but other partitions |
59 |
> might be-- which includes LVM users who followed the Gentoo docs. |
60 |
> |
61 |
> If that's not enough, there is *already* a forked udev version which quite |
62 |
> a few users have been using[3] and reporting success with. I'd hope the |
63 |
> new fork would just collaborate with them, but it's entirely up to them |
64 |
> what they do. |
65 |
> |
66 |
> All of this argues that, irrespective of your views of the layouts admins |
67 |
> prefer to use, they will still use them regardless, and indeed users will |
68 |
> put in the work you yourself told them to.[4] I really don't understand |
69 |
> why people are getting beat up on, just for going ahead and doing what |
70 |
> they've been told to: provide code for an alternative. |
71 |
> |
72 |
> If a Gentoo developer doesn't understand what a Gentoo project means, |
73 |
> that's his problem. Nor should Gentoo projects suddenly change what they |
74 |
> are because "the internet" doesn't understand them. That's a ridiculous |
75 |
> basis for any change. |
76 |
> |
77 |
> The thing none of the proponents of an initramfs, with no other support |
78 |
> for a separate /usr (apart from busybox sep /usr which does NOT fulfil the |
79 |
> requirements presented by Chainsaw, which Council voted to approve before: |
80 |
> for a start it doesn't handle /usr on LVM), have ever answered is: |
81 |
> |
82 |
> How do you deal with the mismatch problem? |
83 |
> |
84 |
> That is, when you have programs or libs upgraded, just as part of normal |
85 |
> system upgrades, and they are now different, potentially incompatible, to |
86 |
> the initramfs. (The potential exists, therefore it must be addressed for |
87 |
> a solution even to be considered as robust.) |
88 |
> |
89 |
> Do you now need another script to run after every emerge to check that |
90 |
> files in your initramfs are in sync? After all, we've been told the |
91 |
> initramfs is "the new root": iow that we should think of it as our rescue |
92 |
> system, so this matters. |
93 |
|
94 |
Debian / Ubuntu have a tool that basically does this. Its |
95 |
update-initramfs. I believe it is called from..the postinst of |
96 |
packages that are supposed to be in the initramfs? honestly I'd have |
97 |
to look up how they implemented it. |
98 |
|
99 |
-A |
100 |
|
101 |
> |
102 |
> At the same time, we've been told "it's only a few hundred kilobytes at |
103 |
> most." That contradiction has never been answered, either. |
104 |
> |
105 |
> [1] http://lwn.net/Articles/492134/ |
106 |
> [2] http://forums.gentoo.org/viewtopic-t-901206.html |
107 |
> [3] http://forums.gentoo.org/viewtopic-t-934678.html |
108 |
> [4] http://forums.gentoo.org/viewtopic-p-6987380.html#6987380 |
109 |
> -- |
110 |
> #friendly-coders -- We're friendly, but we're not /that/ friendly ;-) |
111 |
> |