How to choose what network card or interface to use for replication?

  • Thread starter Thread starter Jacques Koorts
  • Start date Start date
J

Jacques Koorts

Good day,

I have two laptops and I want to use the Ifrared port for my AD replication.
I know about sites and sitelinks, but where do you tell it what interface to
use?

thanks
jacques
 
Jacques Koorts said:
Good day,

I have two laptops and I want to use the Ifrared port for my AD replication.
I know about sites and sitelinks, but where do you tell it what interface to
use?
Click to expand...

An IP network is the key. You will also have to
get the two DCs registered with their SRV records
in your DNS and both of the DCs using that same
DNS server (set) to resovle each other.

IP is required in order to replicate AD so it will
require you to specifically setup Network and
Dial-Up connectsions using Infrared -- it's a
choice in the wizard for a new connection but
the wording is slightly different depending on
which version of the OS you use (2000 or 2003).

Look for the Advanced or Direct connection
choice in the wizard.

One machine will be "client" and one "host" --
or you can alternatively set these up in RRAS,
one to "dial" (or connect) and the other as server
using the router interface.

Once you can get an IP connection between the
machines it is possible to replicate over the
connection.

You would (likely) want to setup multiple
sites.

I supposed you know it is fairly odd to want to
replicate AD over an Infrared connection --
especially given that you must own two copies
of the server software to even need replication
and in this case it would seem easy enough to
afford two wireless cards or a little router to
hook the machines together.

Afterthought: you can set up a symetric connection
where each calls the other, or you might enable
the setting for "mutual authentication" -- which
was largely designed to support dial-up branch
offices where they collect their own AD changes
and inform the replication partner to collect it's
changes while the connection is invoked.

Normally, AD intersite replication is PULL only.
 
Hi,

I know about the IP, DNS, SRV records, Dialup networking thing, RRAS, mutual
authentication. But where do you physically in windows server say, "ok when
replicating, dont go through the network card go through this dialup
connection that I have setup"

The reason I'm using the infrared port (although I'll probably go through
the COM1 port on my laptop) is because its an old laptop with only pcmcia
slots, and they are already occupied. Why this strange setup? well im
studying for mcse and, also It would be nice to have a clean line just for
replication where i have a dedicated line between the laptops, totally
isolated from the rest of the network. Kinda like building my own secure
replication thingy.

Thanks for any help :)
 
Jacques Koorts said:
Hi,

I know about the IP, DNS, SRV records, Dialup networking thing, RRAS, mutual
authentication. But where do you physically in windows server say, "ok when
replicating, dont go through the network card go through this dialup
connection that I have setup"
Click to expand...

In some sense you cannot.

You setup you DCs, Sites, and make sure the Dynamic DNS gets
updated.

Then the DCs (KCC process) decides which DCs replicate.

Then, it just sends the packets using the network and the ROUTING
decides that a particular network path will be used.

So, if I had to answer that question in some way, I would say:

By setting up the network and especially by setting up your ROUTING.

Perhaps by adding a MANUAL route if the two DCs are not directly
connected -- had I invisioned you were connecting DC1 to DC2 by a
direct I/R connection (no intermediates.)
The reason I'm using the infrared port (although I'll probably go through
the COM1 port on my laptop) is because its an old laptop with only pcmcia
slots, and they are already occupied. Why this strange setup?
Click to expand...

Well, for one, because a Laptop is a DC -- although I do that
fairly regularly. <grin>

Second because, most people don't use I/R for IP networks these
days -- maybe for a printer or some such, or syncing their PDA,
but it is fairly uncommon.
well im
studying for mcse and, also It would be nice to have a clean line just for
replication where i have a dedicated line between the laptops, totally
isolated from the rest of the network.
Click to expand...

Nothing wrong with that, but it is unusual.
Kinda like building my own secure
replication thingy.
Click to expand...

Ok. Do you understand the basics of Routing and especially
adding manual routes?

If not, it won't (likely) work unless the two DCs are the endpoints
of the I/R.
 
Jacques,

You could treat this just as though you were putting an DC in your DMZ
across a firewall. There are a number of registry keys that you can use to
shape and define the flow of replication traffic.

The easiest way, though, would be to have them as different networks and
define them as disparate sites... then it will route that path as Herb
suggests.
 
Back
Top