aa
Guest
|
Posted:
Wed Jun 29, 2005 8:53 am Post subject:
Multiple connectors with same address space ... a bug in Exc |
|
|
I posted a question to the Exchange Server administration list a few days ago
and have been getting a big headache trying to solve it. I cannot figure out
if I'm doing something wrong or if this is a bug in Exchange Server. Let me
describe my situation:
I want my outbound internet mail to be sent two different ways depending on
the user. For some users, I would like Exchange server to lookup information
using DNS and for other users I would like it to use a smart host to relay
email.
In an attempt to try to make this happen, I set up two SMTP connectors. One
is called "Default" and the other is called "Relay". They both have an
address space of * with a cost 1. They both also have the same bridgehead
server. The only difference between the two is:
( 1.) The "Default" connector is set to use DNS to resolve hosts whereas the
"Relay" connector is set to use a smarthost like "smtp.example.com"
( 2.) The Delivery Restrictions are set differently on each connector. The
"Default" connector is set to ALLOW all users except those in the "Use Relay"
security group, whereas the "Relay" connector is set to REJECT all users
except those int he "Use Relay" security group.
This setup seems pretty straightforward and I was expecting it to work ...
but it didn't! For some reason, even users who are not in the "Use Relay"
group seem to be going through the smart host instead of resolving the host
via DNS.
And yes, I have already enabled the registry settings as instructed by
Microsoft article Q277872 in order to enable delivery restrictions.
Does anyone know why this is happening? Is this a bug in Exchange Server?
By the way, I have also tried an alternate solution that worked partially.
In my alternate solution, I also set up two connectors, but I set them up
slightly differently by trying to use costs. In this solution, they both
also used the same bridgehead server, but they had the following differences:
( 1.) Again, the "Default" connector was set to resolve via DNS whereas the
"Relay" connector used the smarthost.
( 2.) The "Default" connector was configured to allow all connections except
from people in the "Use Relay" group. The "Relay" connector was configured
to allow connections from everyone.
( 3.) The address space on both connectors was * but I made the cost on the
"Relay" connector higher. I made the cost on the "Relay" connector 10
instead of 1.
I tried this solution because I thought Exchange server would always try to
resolve a host using DNS (since it has the lower cost), except when a user
from the "Use Relay" group tried to send mail and was rejected by the Default
connector -- in this case it would have to fall back and use the "Relay"
connector. I found this solution to work for a while, but as soon as I
changed the users in the "Use Relay" group, the behavior became unpredictable.
The behavior of Exchange server just seems very arbitrary when there are two
connectors with the same address space + different delivery restrictions . I
realize most people only use the same address space for load balancing
purposes, but I would really like to use it to control which users use
certain connectors. Does anyone know a solution to this?
Any help would be much appreciated! Thanks!
|
|