Re: [gentoo-user] separate / and /usr to require initramfs 2013-11-01

[Neil Bothwick]

On Sun, 29 Sep 2013 12:07:44 +0000, Alan Mackenzie wrote:

> Hello, Neil.
> > In what way is it patronising?
>
> It talks down to people.  It insinuates that the readers don't have the
> wherewithal to appreciate that they have been deliberately hurt by
> _somebody_ rather than something "just happening"; that the idea of an
> abstraction "moving" is any sort of justification for anything.

That only applies if you start from the position that this is a
deliberate action against users, it's not, it's just the way the Linux
ecosystem has developed. You call my attitude patronising, but from my
viewpoint your attitude is paranoid.

> Somebody, somewhere was the first person to decide to put early boot
> software into /usr.  Others may have followed him, sooner or later, but
> there was a single person (or perhaps a conspiracy) that did this first.

Not necessarily. It most likely happened that it happened the other way
round, that and increasing amount of software already in /usr became
important during early boot.

> Who?  There was no public discussion of this momentous change, not that
> I'm aware of.  Why?

It was discussed to death on this list several times, going back at
least a year.

> > I think that is entirely the right time to learn of it. If you want to
> > know about the devs' discussions before reaching the decision, you
> > should read gentoo-dev. Until then it was a dev issue, now it is being
> > implemented it is a user issue.
>
> Please be aware the change I was talking about was the decision to break
> separate /usr, not the Gentoo devs' reaction to this breakage.  Why did
> we only become aware of the decision to break separate /usr after it was
> too late to do anything about it?  How could such a thing happen, if not
> through conspiracy?

Ignorance? Not paying attention? This comes as no surprise to those that
read this list. Users of other distros aren't even affected by it as they
have been using initramfs/initrds for many years.

> > I disagree, but then I have actually tried doing it.
>
> I tried, and gave up after a couple of hours.  It was a challenge, but
> I've grown out of being fascinated by challenges for their own sake.
> Then I installed dracut, only to find it won't work on my system.  I
> haven't tried genkernel.  In the end, with regrets, I took /usr out of
> my LVM area and put it into a new partition which became the root
> partition.

Why didn't you try genkernel? That has been creating Gentoo initrds for
longer than I have been using Gentoo. But things would be easier if the
kernel supported LVM.

> > This whole discussion reminds me of a conversation I had with a senior
> > SUSE engineer earlier this year, someone of a similar age to myself.
> > His comment was along the lines of "I remember when Linux users wanted
> > the latest bleeding edge, now they complain every time something
> > changes".
>
> The particular change is not progress, it's not a new feature, it's not
> something useful for users.  It's pure breakage for no good reason.  If
> this is what "bleeding edge" now means, no surprise that people complain
> about it.

The comment wasn't about early boot, I think we were talking abut Unity
at the time, but it seems relevant. Now Unity fits in with your
arguments, a single organisation developed it and sprang t upon their
users without warning. The same is not true of the usr/initramfs
situation.


-- 
Neil Bothwick

Would a fly without wings be called a walk?