FreeBSD and Linux shared installation

[Ian Smith]

On Tue, 21 Jan 2014 14:50:20 +0700, Olivier Nicole wrote:
 > Hi,

G'day Olivier,

 > > Should be good.  I tend to disagree with Olivier about not having a
 > > separate /home, and I recall you also being a fan of dump/restore,
 > 
 > I advocate a separate file system for /home, but I was suggesting to
 > join that with the Prim #3 partition below, not having one file system
 > for /home and one other file system for /common (as it reduces the
 > fragmentation, and what is common is the data, usually located in
 > /home).

The main issue there is that from FreeBSD you'd be working with a (say) 
ext2/3 partition as /home, when you really have to be sure that FreeBSD 
handles R/W flawlessly with it rather than with UFS2+SU(+J), especially 
regarding crash recovery.  Perhaps with FUSE that might be solid enough, 
but personally I tend to trust native formats and tools better, whether 
from the FreeBSD or Linux side.

 > >  > Extend. #1
 > >  >   log. dr. #1        Kali Linux      15 GB   /dev/sda5
 > >  >   log. dr. #2        Mageia Linux    15 GB   /dev/sda6
 > >
 > > From FreeBSD accessing my old OS/2 partitions I seem to recall that
 > > /dev/ada0s5 is the ext drive itself, and within would be ada0s6 and s7,
 > > though the above nomenclature would be right from Linux' POV.
 > 
 > In Linux too (Ubuntu) the Extended #1 is partition #4 and being
 > splited into logical partition #5 and #6. Basically what you write
 > Ian, but you missed the #4: /dev/ada0s4 is the ext drive itself, and
 > within would be ada0s5 and s6...

I'm still not sure about that from FreeBSD's perspective.  Remembering 
back to '98-'99 when I salvaged years of OS/2 work, especially code, and 
those disks only had 3 primary partitions ('C:', OS/2 Boot Manager, then 
drives D: through I: or J: on the extended partition, but with no s4 I 
still had to start at s5, with s6 the first mountable partition (after 
having built the HPFS code which is still in the tree, at 9.1 anyway).

However I may be misremembering (non-ECC memory :) so perhaps Polytropon 
could show us an 'ls /dev/ada0*' when it's done?

cheers, Ian