Windows Server 2008 R2 file and print services : Administering Distributed File System Services (part 2) - Configuring and administering DFS Replication

Configuring and administering DFS Replication

DFSR is a feature Microsoft developed to provide reliable replication between multiple DFS and standard shared folders. By using DFSR, you can have multiple copies of the same shared folder spread across different servers within your organization. Users can then connect to DFS root, which will then link them to the copy of the shared folder that is closest to them on the network. If the user makes any updates to files within the shared folder, DFSR will replicate those changes to all copies of the shared folder as depicted in Figure 10.

Figure 10. DFS Replication.

DFSR does not necessarily require DFS Namespaces. It can also be used with standard shared folders. For example, you could configure DFSR to replicate data from file servers in branch offices back to a central datacenter. In the event that a branch office file server crashed, you would have a second copy of the data in your primary datacenter. DFSR uses advanced algorithms to compress data and replicate only changes to files making it very efficient for wide area networks (WANs). In the following exercise, we will add the DFSR role services and set up DFSR to create a replica of the ClientData share being published in the DFS Namespace.

1.	Open Server Manager and select the node Roles | File services.
2.	In the middle pane, click the link Add Role Services.
3.	Select the option DFS Replication as seen in Figure 11. Then click Next. Figure 11. DFS Replication Role Service.
4.	Click Install.

DFSR should now be installed. Next, we will set up a new shared folder target to replicate the existing ClientData folder to.

1.	Log on to the server that will host the second copy of the data and create a new shared folder named ClientData.
2.	Log on to the DFS Root host server and open Server Manager and select the node Roles | File Services | File Services | DFS Management | Namespaces | contoso.com\Data | ClientData.
3.	Right click the selected ClientData node and choose the option Add folder target as seen in Figure 12. Figure 12. Adding a Second Target Folder for DFS Namespace Shared Folder.
4.	Enter the UNC Path to the second target folder. Then click OK.
5.	You will be prompted to set up replication to keep the target folders synchronized. Click Yes to enable replication between the two folders. This will launch the Replicate Folder Wizard.
6.	The first step is to name the replication group and enter the replicated folder name as seen in Figure 13. Verify whether this information is correct, and then click Next to continue. Figure 13. DFS Replication Group and Folder.
7.	The wizard should display the two shared folders that were set up in the DFS Namespace (see Figure 14 ). Verify whether these are correct, and then click Next to continue. Figure 14. DFS Replication Eligibility.
8.	You now need to select the primary member of the replication group. This is the initial data source to be used to replicate to the other folder. In our example, the initial ClientData shared folder was set up on labfs1 so we will choose this as our initial source. After choosing the primary member, click Next to continue.
9.	Next you will need to select the topology to use for replication. DFSR supports two primary topology types, Hub and Spoke and Full Mesh: Hub and Spoke—Hub and Spoke requires at least three member servers and should primarily be used when data are published to a master share (hub), and then replicated to spokes. In this topology, the hub share will always overwrite the spokes. Full Mesh—A fully meshed topology means that all folder targets are able to replicate changes to all other folder targets. This means that updates to files can be made on any member server and have them replicated to all other member servers. Full Mesh requires two or more member servers in the replication group. In our example, we will be using Full Mesh as we only have two member servers. We also want either member server to be able to replicate updates to the other member server. Select Full Mesh (see Figure 15) and then click Next. Figure 15. DFS Replication Topology Selection.
10.	The next step is to select any bandwidth limits or schedules for replication to occur. You can use these features to limit the usage of WAN links for replication traffic. In our example, we will not be limiting bandwidth and replication should occur continuously. Click Next to continue.
11.	Verify whether the replication settings are correct on the summary page, and then click Create to set up DFSR. You should see a success confirmation page after replication has been set up. Click Close to close the wizard.

Replication should now begin between the two folders. Depending on the amount of data in the initial source server (labfs1), it may take a while for all data to be replicated to the secondary DFS shared folder. You can verify whether replication is occurring by logging onto the second server and browsing the shared folder. You should see that files from the primary server appear in this folder as replication occurs. After initial replication, you should be able to create a file on either shared folder and it should replicate to the other.

Users accessing the DFS root should now be directed to the server which can most efficiently provide access to the chosen shared folder. If either of the DFS members fails, the target folder, ClientData, should remain online and accessible by sending users to the server still online.

Notes From the Field

DFS save conflicts

If multiple people are accessing the same file from different DFS access points and the file is saved by all of them, which version is saved? To handle conflicts, the last write wins, meaning that the version of the file that was last saved is the one that will be replicated to all copies of the file.