samba - Dec 2003 - Dual Interface

I was wondering if it is possible to get Samba to act like two servers on
one box?

I need to upgrade our RH 7.2 PDC and would like to make some configuration
changes.  To this end, I have build a solid migration plan, but it hinges
on if Samba can be used in a particular way, and I have not seen any
documentation either way, so I think I am searching for the wrong terms.

The Plan:
  - Use a development box and have it start responding on two IP addresses
  - Allow the dev-box to expose its shares on its traditional address
  - Allow it to be a PDC on the NEW/SECONDARY address (different set of
    shares)

I will also be migrating its database to LDAP while on the dev box so when
the original server is ready to take over as a PDC again, I just install
it; point at the LDAP for its user accounts; remove the PDC configuration;
make it boot on only one IP again; then replicate all the data.

The big question is, has anyone seen any howtos or documentation on setting
a samba server to act as two samba servers, one configured as a PDC and the
other configured with only public shares (restricted by IP not UID)

TIA
Kevin Fries

> >The big question is, has anyone seen any howtos or 
> >documentation on setting
> >a samba server to act as two samba servers, one configured as 
> >a PDC and the
> >other configured with only public shares (restricted by IP not UID)
> 
> I don't see any reason why you couldn't do that, as long as you had
> two instances of samba running, each with it's own seperate config.
The only thing I was concerned about was items such as locking files,
semaphores, named pipes, and the such.  Some programs do not like to have
their core run twice.  That is why I got nervous when it was not mentioned
in the documents.  I would love to hear from someone that has actually run
this configuration in production, or failed in attempting it.
> The only thing I'm not sure about is if you can share the same data,
> since that might lead to data corruption.
Not an issue.  The machines right now are logically independent.  I am only
putting them together on one machine physically so that I can rebuild the
PDC during production hours and not have to spend time over night.  They
would never serve the same data.  As a matter of fact, my goal was to place
all the data from the old PDC on a removable HD, and then just plug that
drive into the dev box.  This way, when I go to move the data back, I will
simply move the removable, get the service up, then worry about moving it
to the faster raid.  So, the data served by the PDC instance will never
physically reside on the temporary machine.