I’m calling this DAG 2.0 as this is the second release of Exchange that provides database availability groups for high availability of mailbox and now public folder data natively in the product.
In Exchange 2010, database availability groups were a significant shift as it relates to high availability for Exchange services and more directly mailbox data availability for our end users.
Configuring database availability groups in Exchange 2010 was not terribly difficulty as the Exchange team made it easy for administrators that may not be familiar with failover cluster services and the requirements that go along with it. After providing network information, ip addresses for the dag and members to be included, PowerShell did its magic under the hood and provisioned failover clustering for us if it was not already installed and configured on the mailbox servers.
In this article which is the start of a few upcoming topics ultimately focusing on the new high availability change in 2013, I’m going to walk through the entire process of a DAG setup using Exchange 2013 RTM bits and Windows Server 2012 RTM bits. I am using standard edition servers for both Exchange and Windows Server. Exchange has been installed on my mailbox servers and all default databases have been removed leaving a single DB01 database that I’d like to set up in a high availability model. The mailbox servers in my lab are configured as follows:
- They are joined to deploy.com lab domain as EX13-MB1 and EX13-MB2
- Each mailbox server has two network interfaces attached to the vm’s (Hyper-V 2012)
- Exchange traffic is on the 172.16.10.X network
- DAG replication traffic is on the 198.119.110.X network
Each network adapter on the mailbox servers has been renamed to something descriptive signifying what it’s purpose is. For demonstration purposes, the adapters are named “RPCoHTTP” and “Repl“. If you are unaware, RPC is no longer used in Exchange 2013 and all client access utilizes RPC over HTTP (Outlook Anywhere). For more information around this change please go here. The screenshot below shows the current adapter configuration.
After configuring the network adapters on each mailbox server, I want to ensure they can communicate on the replication network before proceeding. To do this, open a command prompt and verify via ping, specifying source IP address as shown below.
Next, the network interfaces need to be configured correctly and setup in the proper binding order. To do so, on the network connections screen, hit the “Alt” key to bring up the menu bar as shown below and click “Advanced“, then “Advanced Settings…” as shown in the image below.
In the advanced settings of the network adapters, I’m going to move the “RPCoHTTP” adapter to the top of the list in the binding order. While still in the advanced settings, I’ll also configure the “Repl” network by removing unnecessary services / protocols from the adapter. The two images below reflect the changes made to each mailbox server.
Lastly, we need to disable the replication network (Repl) adapter from registering its connection suffix in DNS in the TCP/IP properties.
The below screenshot shows the same changes being made on the TCP/IPv6 connection for the replication adapter.
The above configuration is how I have always configured my adapters for setup in a dag and wonder if those best practices still hold true in Exchange 2013. If anyone reads and has any other ideas and/or configurations, please let me know. The next article will focus on setup of the database availability group in the Exchange Admin Center (EAC).