Changing the IP addresses of Exchange 2013 servers in a DAG
A couple of months ago I had a request from a customer to change the ip addresses of all their Exchange 2013 servers because of network changes that they needed to do in their environment.
After reading the article written by Tzahi Kolber: Changing DAG & DAG members IP addresses we decided to use the actions described below.
The environment
They had a DAG with 4 nodes in two datacenters (EMEA and US) and a FSW in a third datacenter in EMEA. Also they wanted to switch from a single network card used by the DAG servers to 2 network cards, one for MAPI traffic and one for Replication.
All servers have the MBX/CAS roles.
Before making the changes
Table with IP addresses:
|
# |
Server name |
Old Ip address |
New MAPI IP |
New Replication IP |
|
1 |
DAG2013-MBX1 |
10.2.1.101 |
10.2.5.101 |
10.2.2.101 |
|
2 |
DAG2013-MBX2 |
10.2.1.102 |
10.2.5.102 |
10.2.2.102 |
|
3 |
DAG2013-MBX3 |
10.2.3.101 |
10.2.6.101 |
10.2.4.101 |
|
4 |
DAG2013-MBX4 |
10.2.3.102 |
10.2.6.102 |
10.2.4.102 |
|
5 |
DAG |
10.2.1.110, 10.2.3.110 |
10.2.5.110, 10.2.6.110 |
|
Talk to all applications owners and notify them of the change. It is often that applications use the IP address of the servers rather than the VIP or the FQDN of the server.
Make sure that the Old IP addresses are routable to the new MAPI IP addresses.
Firewall rules should be updated in order to provide access to the new IP addresses.
The change
In our case we moved the nodes to different subnets so in this we needed to add Ips for the DAG in those subnets as well.
Set-DatabaseAvailabilityGroup –identity DAG -DatabaseAvailabilityGroupIPAddress 10.2.5.110, 10.2.6.110, 10.2.1.110, 10.2.3.110
1. We took the first server out of the HLB.
2. We moved all the active DBs to other nodes.
3. We’ve put the first node in maintenance mode as per article: https://blogs.technet.com/b/nawar/archive/2014/03/30/exchange-2013-maintenance-mode.aspx
4. Checked if the PAM was not hosted on the server where we did the changes:
Get-ClusterGroup
5. As the OwnerNode was the server on which we are doing the changes. We moved it using:
Move-ClusterGroup “Cluster Group” –Node [node where to move it]
6. Enabled the second NIC which will be used for replication. Configured the IP address, unchecked “Register this connection’s addresses in DNS”.
7. Set static routes for the Replication NIC.
8. Checked NIC bindings putting the MAPI network on top.
9. Switched the RDP connection to the Replication NIC and configured the IP address for the MAPI network.
10. Rebooted the server.
11. Check that we have the new IP in DNS and that it has replicated in the environment.
12. If you do a Get-DatabaseAvailabilityGroup | fl you should see:
13. Put the server in the HLB.
14. Pull out server from maintenance mode as per article: https://blogs.technet.com/b/nawar/archive/2014/03/30/exchange-2013-maintenance-mode.aspx
15. Move active databases on this server.
16. Repeat the previous steps on the other servers.
17. At the end do not forget to remove the old IPs for the DAG using:
Set-DatabaseAvailabilityGroup –identity DAG -DatabaseAvailabilityGroupIPAddress 10.2.5.110, 10.2.6.110
18. Disabled replication from the MAPI network.
Note: The above steps have been done for a DAG that had automatic DAG network configuration. The information regarding the networks is updated every 30 minutes.
As all major changes in a production environment you should test all the steps in a test environment in order to make sure you don't have any particular settings you need to take into consideration.