Multiple WINS Servers Enhancement

[Christopher R. Hertel]

On Jul 8, 12:37am, Elrond wrote:
:
> As Jeremy already wrote, this is a "documentation issue".

A few chapters in this conversation have been missed.  It was one of the other
Chris' (sorry I can't remember which one, but there are three of us chatting)
who pointed this out early yesterday.

> > > - Windows clients will only be able to address two of them.
>
> Well, that's their problems.
>
> But it might even make some sense (provided, the
> replication does the job the right way):
> When you have a big company, you setup a wins-server at
> every big place and the people use the local as primary and
> a nearby as secondary.

Yes, I remember a presentation at one of the LISA/NT conferences in which this
architecture was described.

> > > This also means that we're going to want to implement
> > > WINS replication ourselves.
> >
> > Nope.  Not necessary for this round.
>
> Precisely.

Disagree.  Of course, I've been suggesting the WINS replication for a while so
perhaps I'm just stuck in a rut.

If you think of Samba as a product, then adding the ability to use failover
without adding the ability to provide failover is poor marketing.  I'm not
saying we need to do it for the next release, but we should be thinking about
it, *and* thinking about how to enhance the capability.

> Can you (Chris) try and see, wether the patch also applies
> in TNG?

I have not looked at the TNG code so I may need some help with it.  I don't
even have it checked out.  Let me get this into HEAD and then I'll let you
know.

> Just a few minutes ago, I've merged namequery.c from HEAD,
> so applying the patch shouldn't be a big problem (nmbd/ is
> completely in sync, except for processlogon, which uses
> different unicode-handling, and shouldn't be affected)

Shouldn't be an issue then.

Chris -)-----

-- 
Christopher R. Hertel -)-----                   University of Minnesota
crh at nts.umn.edu              Networking and Telecommunications Services