On Wed, Aug 3, 2022, at 10:05 AM, Peter Boy wrote:
> Am 02.08.2022 um 21:34 schrieb Chris Murphy
<lists(a)colorremedies.com>:
>
> Current tech spec file system section:
>
https://fedoraproject.org/wiki/Server/Technical_Specification#File_system
>
> I propose the following change:
>
> The default file system type for Fedora Server installs will be XFS running atop LVM
for all partitions except /boot. [1]
>
> [2]
>
> An option will be provided in the Fedora Server installer to enable disk encryption.
I added the proposal as an alternative to the text at
https://docs.fedoraproject.org/en-US/server-working-group/docs/server-tec...
(It may take some time until the server updates).
I remain convinced that it is indispensable to explain concisely how we
intend to achieve the goal technically. The text describes what we are
doing and how.
The text is confusing because it's using quite a lot of jargon that doesn't help
me understand either what's being done or why or how.
Example: Fedora Server gives the highest priority to maximum reliability and security of
data with a maximum of possible performance.
This is achieved with any of the three file systems presently supported on Fedora. So
what's the effect of this statement? What conclusion is drawn from it? It seems less
non-technical, and more of a marketing statement that could apply to any product. But this
is found at the top of the file system and storage section.
If there is high priority on maximum security, why not encrypt by default? How do you
avoid the inherent contradiction of maximum reliability with maximum performance?
Another example: "strict separation of system and user data". What does strict
mean? What's the difference between "separation" and "strict
separation"? What's the technical distinction and the result from it? What
examples of separation are strict and not strict? And why does the distinction matter in
this document?
The default partitioning is one of the key differentiators of Fedora
Server Edition and is reflected in our documentation and also
determines what Ansible support of services and administration must
accomplish.
How does default partitioning determine anything about what Ansible must accomplish? It
suggests users will have problems if they pick non-default partitioning. Will they? Why?
Chris' wording is not accurate. We are not simply partitioning
XFS over
LVM. We provide an elaborate structuring.
What's elaborate about it? I was involved in the original discussions, and had the
idea of reserving space in the VG for various use cases rather than deciding to use it all
for a single large root LV. I wouldn't call this elaborate, but rather simple yet
non-obvious.
--
Chris Murphy