Windows 2008 configuration cluster continuous replication (1)

Although configuring Cluster Continuous Replication (CCR) on Windows Server 2008 is similar to the configuration process on Windows Server 2003, there are some significant differences. Before configuring CCR, we recommend that you read the article "Cluster Continuous Replication" in its entirety. Also, be sure to meet all of the requirements in the article "Deploying Cluster Continuous Replication" in advance.

To install CCR on Windows Server 2008, complete the following steps:

1. Configure the hardware installation and start the cluster network construction and configuration process.

2. Construct a cluster and generate the first and second nodes in turn.

3. Configure the tolerance of the cluster network and lost cluster heartbeat.

4. Configure and protect the file share witness.

5. Set up cluster activity and inactive Mailbox server roles. During the setup of the active Mailbox server role, the

Clustered Mailbox Server (CMS) will be created.

Note:

We recommend following the steps below after completing each step. After completing all the steps, it's a good idea to verify the validity of the CCR solution before applying it.

After the installation is complete, perform the following tasks:

· Adjust the control settings for failover.

· Adjust the default configuration of the transfer dumpster.

· Verify the ability to move CMS between cluster nodes.

· Make the multi-network option for continuous replication active valid.

Before starting any of the steps mentioned above, you must first ensure that the target computer has the operating system components required for Windows Server 2008. For specific steps to install the necessary components of Microsoft Exchange on Windows Server 2008, see the article "How to install the necessary components of Exchange 2007 SP1 on Windows Server 2008."

The following sections explain each step of the installation process in detail.

Network Construction and Configuration

When creating a CMS in a 2-node CCR environment for Windows Server 2008, you must have a sufficient number of available IP addresses. Significantly unlike legacy clusters, Windows Server 2008 failover clustering provides new networking capabilities. For example, Windows Server 2008 failover clustering provides support for multiple subnets, Dynamic Host Configuration Protocol (DHCP), IPv4, and IPv6. When a Windows Server 2008 failover cluster is running, Microsoft Exchange Server 2007 Service Pack 1 (SP1) supports failover between two subnets in a geographically dispersed cluster. This support includes Single Copy Cluster (SCC) and Mailbox servers in a CCR environment.

Note:

Although Windows Server 2008 failover clusters support DHCP IPv4, we still recommend using static IP addresses in real-world environments. If you must use DHCP IPv4 in a failover cluster, it is recommended to configure the DHCP server to ensure that there is no time limit for the use of IP addresses.

Start configuring Windows Server 2008 failover clusters. A single cluster node can now be placed in a separate network of routes. This requires resources that rely on IP addresses (such as network name resources) to implement an OR logic because it is not possible for each cluster node to have a direct local connection to connect to each network adjacent to the cluster. This will help to distribute the IP address and network name resources online when a business or application fails over to a remote node.

All online IP addresses associated with the Network Name resource will be dynamically registered in the Domain Name System (DNS) by a custom list (if set to Dynamic Update), thus enabling those online IP address resources. Return to the customer at the first time. Since the cluster nodes are placed in different routing networks, and the communication mechanism is changed in order to implement a reliable session protocol (unicast) on top of the User Datagram Protocol (UDP), the networking requirements of the geographically dispersed cluster are no longer applicable. . As a result, organizations can configure failover clusters between two physical data centers without the need to use virtual local area network (VLAN) technology to extend cluster subnets between the two locations.

When CMS mobility and failover occur in a geographically dispersed multi-subnet failover cluster, the CMS name will be preserved. However, the IP address assigned to the name will not be retained. Whether the server has availability to users and other servers depends on the propagation of new IP addresses in the DNS. DNS propagation may take some time. Because of this, we recommend configuring the time-to-live value (TTL) of the CMS DNS host record to 5 minutes (300 seconds). For detailed steps on configuring the DNS TTL value of the CMS, please see the article "How to configure the DNS TTL value of the network name resource". After the CMS DNS TTL value is configured successfully, you must stop and restart the CMS for the changes to take effect.

Although the internal Microsoft Office Outlook client does not need to update or reconfigure the configuration files when using the new IP address connection, they still need to wait for the local DNS cache to be emptied before parsing the CMS name from the old IP. The address is transferred to the new IP address. When the IP address is propagated to the appropriate DNS server, run the following command on the client command line to clear the DNS cache of the Outlook client.

ipconfig /flushdns

Both public and private networks require an IP address. The requirements for private and public addresses are as follows:

l Private Address: Each network adapter of each node of the cluster private network requires an IP address to correspond to it. You can use a static IPv4 address or a dynamically assigned IPv6 address. However, be aware that you cannot use the IP address of the same network on the same subnet or public network. The recommended private addresses are 10.10.10.10 and 10.10.10.11, and the subnet mask is 255.255.255.0.

l Public address: A cluster public network, sometimes called a hybrid network. Each network adapter of each of its nodes also requires an IP address to correspond to it. Unlike private networks, failover clusters and CMSs also require IP addresses for client and administrator access. Also, you should be careful not to use the IP address of the same network in the same subnet or private network. You can use an IPv4 address, a DHCP IPv4 address, or a static IPv6 address.

Important:

All network adapters for a clustered network must use the same version of TCP/IP. That is, they must all use only IPv4 or both IPv4 and IPv6.