message flow with Exchange module

Question

Our setup:
Windows 2003 SP1
Exchange 2003 SP2
AutoStart 5.2.1
Autostart mod for Exch 2.0.15
Replistor 6.2 SP1

I have been (slowly) working to get our new, internal Exchange cluster (FEBO) running and while there are numerous problems, I am worried mostly about internal message flow from our existing Exchange server (SBS 2003, name=ZEUS) and the new cluster. I can send messages from ZEUS to FEBO but not the other way.

I have turned on message tracking in both servers and I notice that the message IDs contain the name of the active physical cluster ( @PHYS1.domain.local), NOT the name of the virtual machine ( @FEBO.domain.local). I believe that this is the reason that ZEUS never accepts messages from FEBO...because they look to be coming from PHYS1 instead (a member server).

My question is: how SHOULD this work? If the msgID should contain the virtual name, will the emcadmin command help me to reconfigure it properly?

An additional question, both Autostart and Replistor have functionality for Exchange virtual machines. We are using the module for Autostart right now. Can the two be used together at all? Is it possible to use the RS module instead with Autostart still being used?

Thanks in advance,
Ryan

tribicic · Answer

Can't answer the first question as I wouldn't call myself an Exchange expert, but the recent module should at least try to make it seem that there is a virtual exchange server running in there.
Regarding the Autostart/Replistor Exchange support - unless something changed recently Replistor used the same method as the old (1.x) Autostart Exchange module, basically using two distinct physical Exchange servers and moving user mailbox locations during failover. This solution was far from ideal. Now it is possible that the Replistor module was updated as well, but in any case since they will be using the same technology, they should give you the same results, at least as far as Exchange is concerned.

ryanmcfee · Answer

Thanks for your input Tomislav. I assumed the technologies were similar if not the same.

Ryan9 · Answer

Hey Ryan, not sure if this is still an issue for you or not but I ran into a similar issue and it took me a good month to figure it out.

Just to get mailflow working:

-Go to the properties of the Default SMTP Virtual Server Properties in ESM
-Click on the Access tab
-Click on the relay button at the bottom
-You have to add the IP addresses for ALL servers in your organization. YOu could do IP Range but it isn't recommended
-Restart Default SMTP Virtual server (May not be necessary)

Then there is still an issue with SMTP authentication:

Use adsiedit.msc or setspn to change the SPNs for both the Virtual and Physical computer accounts. I recently did an install for a site that had 20 Exchange servers and I had to make sure that all fqdn hostnames in the SPN were lowercase and that the NETBIOS names were all uppercase.

Lastly, I made sure that there was a PTR record for both the physical and virtual name.

After making sure these changes were in place mail is flowing properly and SMTP authentication is working properly. After a month of decoding the base64 strings from network traces and researching SMTP RFC's and what not. I was finally able to resolve the problem.

Replication Manager

message flow with Exchange module

Was this post helpful?