1 |
On Wed, 2012-07-18 at 18:24 -0700, Matthew Marlowe wrote: |
2 |
> > It would be nice if a sensible structure could be proposed and |
3 |
> > agreed by ALL Linux distributions (coordinated with BSD). |
4 |
> > |
5 |
> |
6 |
> +1 |
7 |
> |
8 |
> If a new file system standard is required, my preferences based on a |
9 |
> history of what is worked and changed over the last 20-30 years would |
10 |
> be: |
11 |
> |
12 |
> - OK with requiring / and /usr on the same FS. This has become common |
13 |
> practice with virtualization and small system deployments, and I |
14 |
> haven't seen any compelling advantage for keeping separate on larger |
15 |
> boxes. |
16 |
|
17 |
No one proposes that, the only requirement that you have for modern |
18 |
Linux to work well is that if /usr is on a separate partition, you need |
19 |
to mount it before starting your main system (ie, from an initramfs). |
20 |
|
21 |
> - NOT OK with limitation on allowing /var, /opt, /home, or any other |
22 |
> common server mount points to require use of initramfs/initrd. There |
23 |
> is enough disagreement as to the reliability and ease of maintenance |
24 |
> of initramfs/initrd that it should not be needed for common server |
25 |
> deployments. |
26 |
|
27 |
This is clearly not needed, /run was even invented to allow /var to be |
28 |
mounted later. |
29 |
|
30 |
> - It would be nice if the rootfs used a snapshot based filesystem and |
31 |
> if the bootloader was intelligent enough to easily allow admins to |
32 |
> boot to older snapshots as an expectation for any standard modern unix |
33 |
> system. |
34 |
|
35 |
One of the reasons to put everything in /usr is to allow using a |
36 |
snapshot based FS, so you can run a system where /usr is read-only and |
37 |
where when you can do all upgrade atomically by writing your changes to |
38 |
a read-write snapshot and then switch all at once. So you never have any |
39 |
half-upgraded package on your system. |
40 |
|
41 |
> - Ideally, server motherboards would come with flash based storage |
42 |
> where sysadmins could install rescue environments as part of a normal |
43 |
> unix install, and that the boot loader or bios would be smart enough |
44 |
> to provide the option to boot from it automatically whenever a normal |
45 |
> boot failed. |
46 |
> - NOT OK with removing the distinction between user and system |
47 |
> binaries. Essential binaries required to boot and troubleshoot system |
48 |
> problems should be located separately from user binaries. Security |
49 |
> sensitive or paranoid admins should be able to make the system binary |
50 |
> path read-only or completely remove the user binary directory from |
51 |
> roots PATH if they so wish. |
52 |
|
53 |
The rescue system should be entirely separate from the main system, so |
54 |
it survives mishandled upgrades. So having that should not hinder how |
55 |
your main system is built. So you should have it as a separate partition |
56 |
or you can even have it an initramfs (ie, in a single file on the main |
57 |
system). |
58 |
|
59 |
> - OK with merging / and /usr, but in that case...why not just move |
60 |
> everything in /usr to /...but limit /sbin to system binaries and /bin |
61 |
> to user ones? This would be horrible for migrations though and |
62 |
> possibly confuse many scripts. |
63 |
|
64 |
The idea of putting everything in /usr instead of / is that you can then |
65 |
make /usr read-only and you can share /usr between multiple systems, |
66 |
while / is read-write and contains /root and /etc. |
67 |
|
68 |
> - NOT OK with making systemd the default init system anytime in the |
69 |
> next few years, it is way too immature and like most major system |
70 |
> software changes...probably will take much longer before it really has |
71 |
> the standing to propose being a new standard. |
72 |
|
73 |
I fully expect systemd to be the init system of the next iteration of |
74 |
RedHat Enterprise Linux, which is probably the most "enterprise" of all |
75 |
distributions, with the most QA and support and everything. It's not a |
76 |
side project of dude of his basement, it has the full support of a large |
77 |
team of people at RedHat. There has probably already been more testing |
78 |
done on systemd today than on OpenRC... |
79 |
|
80 |
> - What other elements can new filesystems like btrfs offer that should |
81 |
> be considered? ext3/ext4 has worked great with the older |
82 |
> standards...but it essentially mimicked the capabilities of older |
83 |
> file-systems that the original unix standards were based on. Btrfs |
84 |
> might change our expectations. I'm assuming that btrfs will be the |
85 |
> standard production fs in a few years. |
86 |
|
87 |
The big thing that btrfs brings is snapshots and subvolumes... So it |
88 |
makes it possible to do atomic upgrades and such. Also, you can have |
89 |
"apps" be subvolumes and also handled atomically. |
90 |
|
91 |
-- |
92 |
Olivier Crête |
93 |
tester@g.o |
94 |
Gentoo Developer |