Troubles with Exchange 2010 and Exchange 2003 Routing Group Connector

We have an environment which includes a combination of Microsoft exchange server 2010 and Microsoft exchange server 2003 existing on the same domain.

The various reasons at the moment we want to leave both versions of exchange coexisting. What we have noticed is that there are at times delivery problems between the routing group connector that is automatically configured when you add an exchange 2010 server to an existing 2003 domain.

This was evident by checking the queue in the exchange 2003 server. It showed that a number of messages were pending delivery and of the status was in “retry”.

Naturally enough, Google was our first point of call on this matter and we discovered that there were many others are experiencing delivery problems between the two servers.

In our case, the problem may well have initially risen due to some necessary changes that we made with the default routing of e-mail being sent from our exchange 2003 server.

After we made the configuration changes and attempted a short-term fix to another problem by using a smart host server, we found that how playing around with the settings was likely a contributing factor to the routing group connector problem.

To solve this problem we performed the following actions.

1. We log on to the existing 2010 exchange server.

2. We started the application known as exchange management shell

3. We listed our existing exchange connectors by typing the following command

  • Get-RoutingGroupConnector

4. We expanded on this list to confirm the details by typing the following command

  • Get-RoutingGroupCOnnector | fl

the results looking something like this:

Identifying the name of the existing connector as: VSERVER13-DGM56G41

5. We use the following command to delete all remove the existing connector:

  • Remove-RoutingGroupConnector -identity “VSERVER13-DGM56G41″

6. Initially this area out referencing: Multiple Instances  So we had to look in more detail at the screen capture above of the listed connectors lost using be  |f1 command , to find the name referenced against Guid: a long string of numbers looking something like this :
8276G-736b-37698726-3782661b78

We then use the following command referencing the GUID to remove that connector:

  • Remove-RoutingGroupConnector -identity “8276G-736b-37698726-3782661b78”

7. We then checked that the existing connector was removed by a gang typing:

  • Get-RoutingGroupConnector

The showed no results and we also verified in the exchange 2003 server that the exchange routing group connector had been removed from the list there.

    8. We then created a new wrapping group connector called “VSERVER13-DGM56G41”

The two servers that we were connecting were server1 (exchange 2003) and server2 (exchange 2010). We use the following command to complete this:

  • New-RoutingGroupConnector -Name “VSERVER13toDGM56G41” -SourceTransportServers “server2.fserver4.com” -TargetTransportServers “server1.fserver4.com” -Cost 100 -Bidirectional $true -PublicFolderReferralsEnabled $truethis created

9.Then we use the “Get-RoutingGroupConnector” commandto gain verify the existence of the new connector. Once verifiedour problems were solved… or was it???

We did fix the connector and again mail was delivered. What we then found was after some time the connector would fail. WTF you say… as I did! I hit google again and again and again, found loads of people talking about similar problems but none with suggestions other than what I had done above.

Then I discovered a hidden little nugget of gold. Turns out that Exchange 2010 (and possibly Exchange server 2007) will not accept remote mail unless the pickup queue is located on a server with a minimum of 4 GB free space. By default this is the C: drive.

Now there are a few things we found that impact that.

1. After the server is rebooted, and even with less than 4 GB it will accept mail for a short time. (This I suspect is just to trick the heck out of us).)

2. The free space will shrink and grow a little with server maintenance. So don’t just look at a disk with 5 GB and say, all is well.

3. There appears from what I can see to be absolutely NO announcements in the server log files about the queue shutdown. It just stops!

So free some space, or research on moving the queue pickup to another drive. (I will not list it here, but it is easy to find in Google). Good luck and may you be blessed with low disk space as the cause of your problems as I was!

Finally Solved!

10 comments

Leave a Reply