[indiana-discuss] Default Partitions sizes

[Shawn Walker]

On 07/10/2007, Peter Tribble <peter.tribble at gmail.com> wrote:
> On 10/7/07, Mike Gerdts <mgerdts at gmail.com> wrote:
> > On 10/7/07, Shawn Walker <swalker at opensolaris.org> wrote:
> > > I completely agree, perhaps "syspool" and "userpool" would be the most
> > > appropriate here? With /usr/local falling under userpool perhaps?
> >
> > Not if pkgadd (or its successor) put stuff there or is likely to do so
> > in the future.  If pkgadd or its successor has a monolithic package
> > database, anything the package database touches must be part of the
> > root pool.  To me the root pool minimally contains the set of
> > directories that are affected by the packaging database that delivers
> > /.
>
> I would go further and want to see that set of directories as a single dataset.
>
> The case of /usr/local is arguable, though. I wouldn't expect that to be
> part of the packaging domain. On my systems /usr/local is just a
> symlink pointing to a relevant automounted nfs filesystem that is
> shared amongst all machines. (I shared the same one amongst all versions
> of Solaris until Solaris 10, which shipped enough utilities that I could
> clean /usr/local out. The fact that half the stuff I've built for Solaris 10
> doesn't work on Solaris Express means I'll have to build up another
> instance.)

Mike's point does bring up an interesting consideration though. Yes, I
know that packages can deliver to multiple paths, etc. However, one
does wonder for monolithic-db backed package systems if the ability to
have a database per dataset and set up the datasets correspondingly
would have a tangible benefit.

-- 
Shawn Walker, Software and Systems Analyst
binarycrusader at gmail.com - http://binarycrusader.blogspot.com/

"Beware of bugs in the above code; I have only proved it correct, not
tried it. " --Donald Knuth