Re: [tlug] Current practices for Linux partioning?

[Attila Kinali <attila@example.com>]

On Thu, 12 Apr 2012 14:57:05 +0200
Francois Cartegnie <fcartegnie@example.com> wrote:

> Le 11/04/2012 09:31, Attila Kinali a écrit :
> > No. As i said, for a single user system it doesn't make sense.
> > If i fill up /home, i know i've filled it up and have problems anyways.
> > If i fill up / i want to notice it as soon as possible. And as i am
> > the only user on the system, there is no need to isolate users from
> > each other (or their stupidity for that matter)
> 
> What about the "pool" based fs way (btrfs) which will still allow
> splitting without the drawbacks ?


I'm following the btrfs development for a few years now, but i don't
think it's prime time ready yet. There are still too many open issues
(not having a working fsck yet is one of the biggest). IMHO it will
take another 4-6 years until btrfs becomes stable enough for general
use.

Other than that, yes, having resizeable volumes like btrfs/zfs does would
nicely fix the problem with fs resizes with more conventional fs.

 
> Splitting is still useful to compress filesystems (source code) or do
> efficient snapshot based backups (*), or easily change/revert rootfs.

Yes, that would help. But, to be honest, how many people would use that?
How many people even know that such features exists?


> (*) Although it isn't efficient as zfs. (zfs snapshot, zfs send, and its
> zero stat())

Hmm? Why is zfs more efficient? I thought both use more or less the same
system for snapshotting?

				Attila Kinali

-- 
Why does it take years to find the answers to
the questions one should have asked long ago?