1 |
On Wed, Jul 18, 2012 at 7:04 PM, Olivier Crête <tester@g.o> wrote: |
2 |
> On Wed, 2012-07-18 at 18:24 -0700, Matthew Marlowe wrote: |
3 |
> |
4 |
>> - It would be nice if the rootfs used a snapshot based filesystem and |
5 |
>> if the bootloader was intelligent enough to easily allow admins to |
6 |
>> boot to older snapshots as an expectation for any standard modern unix |
7 |
>> system. |
8 |
> |
9 |
> One of the reasons to put everything in /usr is to allow using a |
10 |
> snapshot based FS, so you can run a system where /usr is read-only and |
11 |
> where when you can do all upgrade atomically by writing your changes to |
12 |
> a read-write snapshot and then switch all at once. So you never have any |
13 |
> half-upgraded package on your system. |
14 |
> |
15 |
|
16 |
I understand that RHEL is promoting this system management model, but |
17 |
I'm not sure it is any way superior than other models that do not |
18 |
require RO /usr. Many enterprises today solve the same issue via |
19 |
automation with puppet, active traffic management via LVS/HA/Load |
20 |
Balancers, or replace /usr snapshots with virtualization where a new |
21 |
VM snapshot or clone is upgraded and then replaces the active VM. |
22 |
Therefore, I'm not convinced we should promote /usr/bin and /usr/sbin |
23 |
over /bin and /sbin. It is a point in their favor to give more |
24 |
flexibility but the simplicity of also just removing /usr would also |
25 |
be compelling if our goal afterall is to simplify the FHS. |
26 |
|
27 |
> |
28 |
>> - OK with merging / and /usr, but in that case...why not just move |
29 |
>> everything in /usr to /...but limit /sbin to system binaries and /bin |
30 |
>> to user ones? This would be horrible for migrations though and |
31 |
>> possibly confuse many scripts. |
32 |
> |
33 |
> The idea of putting everything in /usr instead of / is that you can then |
34 |
> make /usr read-only and you can share /usr between multiple systems, |
35 |
> while / is read-write and contains /root and /etc. |
36 |
> |
37 |
|
38 |
Most enterprises that need the ability to make /usr RO for canonical |
39 |
server configs could just as well get by with automated configuration |
40 |
management tools (e.g. puppet) and smart hypervisors/SAN storage. |
41 |
This allows deployment of new servers within minutes and it reflects |
42 |
at least the true reality that a true controlled system state is a |
43 |
snapshot of server config files, virtual hardware, and all binaries. |
44 |
Just making /usr RO is a cheat that might work well for long lived |
45 |
binary distributions that require major version upgrades anytime |
46 |
software packages update their config behavior dramatically. I |
47 |
haven't found it very helpful for source type systems like Gentoo. Of |
48 |
course, others may disagree. |
49 |
|
50 |
>> - NOT OK with making systemd the default init system anytime in the |
51 |
>> next few years, it is way too immature and like most major system |
52 |
>> software changes...probably will take much longer before it really has |
53 |
>> the standing to propose being a new standard. |
54 |
> |
55 |
> I fully expect systemd to be the init system of the next iteration of |
56 |
> RedHat Enterprise Linux, which is probably the most "enterprise" of all |
57 |
> distributions, with the most QA and support and everything. It's not a |
58 |
> side project of dude of his basement, it has the full support of a large |
59 |
> team of people at RedHat. There has probably already been more testing |
60 |
> done on systemd today than on OpenRC... |
61 |
> |
62 |
|
63 |
Well, we know that systemd will be in the next RHEL release - it was |
64 |
in their recent roadmap presentation. However, that release would be |
65 |
RHEL7 and most established servers are RHEL5 now and RHEL6 is still |
66 |
relatively early in deployment. It will be a few years before most |
67 |
RedHat customers can say whether systemd was a good move for RHEL. |
68 |
And, RedHat has made changes in the past that while supported might |
69 |
not become the default config for other distributions (e.g. SE Linux). |
70 |
I don't see that we need to assume now that systemd will define the |
71 |
default Linux ecosystem in the future, even if it becomes widely |
72 |
deployed on RedHat systems.. |