[geeks] Solaris 10 / OpenSolaris bits to be in next version of OSX

[Charles Shannon Hendrix]

Thu, 10 Aug 2006 @ 21:56 +1000, Scott Howard said:

> On Wed, Aug 09, 2006 at 07:38:43PM -0500, Jonathan C. Patschke wrote:
> > > Why *not* do it this way?
> > 
> >    1) Cutting a disk into slices is a Good Thing.  Sometimes, different
> >       slices need different filesystems.  Why should only the ZFS
> >       filesystems get the RAID benefits?
> 
> So cut it into slices.  ZFS doesn't stop you doing that. You can run ZFS
> on top of either a whole disk, just a slice, or even a VM device if you
> want.

I thought you couldn't get the benefits of ZFS unless it used its fully
integrated volume management and RAID?




> >    2) It complicates system administration.  Some filesystems are defined
> >       as slices mounted in places according to /etc/vfstab.  Some
> >       filesystems are now defined in wherever ZFS retains its
> >       configuration.
> 
> OK, this is a somewhat fair cop at the moment. It will change once you
> can use ZFS for root disks, but even until then it's an acceptable
> overhead in my mind. The minimal amount this complicates sysadmin work
> is far, far outweighed by the fact that most of the rest of ZFS
> simplifies admin.

I'm not sure this changes things for a lot of admins.  A lot them like
being able to use an editor to admin their boxes, especially when the
tools for the hidden metadata start going bad.

Likewise you get similar feedback about OS and userland add-ons that
auto-edit things like /etc/*fstab.  A lot of people just don't like
things like that.

> > > I *like* setting up an eight-disk RAID-5 array with ONE COMMAND.
> > 
> > That's why we have shell scripts.
> 
> bawhahhaw..  that's a good one.
> 
> Ohh.. wait.. you're serious?

Why shouldn't he be serious?



-- 
shannon "AT" widomaker.com -- ["People should have access to the data which
you have about them.  There should be a process for them to challenge any
inaccuracies." -- Arthur Miller]