The Distributed File System on Windows Server 2012 R2 consists of two components, Distributed File System Namespaces and Distributed File System Replication. In this blog post we will review Distributed File System Replication, and then look at how to install and configure DFS-R on Microsoft Windows Server 2012 R2.
Navigation
1. Distributed File System Replication (DFS-R)
1.1 Installing DFS-R
1.2 Creating a Replication Group
1.3 Configuration Options in DFS-R
1.3.1 Editing Replication Schedules and Bandwidth
1.3.2 Managing Connections
1.3.3 Configuring properties of replication group members
1.3.4 Configuring properties of replicated folders
1.3.5 Replication topologies
.
1. Distributed File System Replication (DFS-R)
DFS Replication is an efficient, multiple-master replication engine that you can use to keep folders synchronized between servers. DFS replication can occur across multiple sites and servers that live within the same Active Directory forest. DFS Replication uses remote differential compression (RDC) to locate and replicate only the changed data blocks of a file, instead of replicating a file in its entirety. This reduces the amount of network traffic generated by DFS-R.
.
1.1 Installing DFS-R
1.1.1 You can install the DFS-R role by using the Add Roles And Features Wizard in Server Manager.
1.1.2 Select the Role-Based Or Feature-Based Installation option
1.1.3 On the Select Server Roles page, select DFS Replication
1.1.4 Click Add Features to install management tools as well
1.1.5 On the Select features page, select Remote Differential Compression
1.1.6 Click Next, and then click on Install
.
1.2 Creating a Replication Group
A replication group enables you to configure folder replication so that folder replicas remain synchronized across servers.
1.2.1 Right-click the Replication node in DFS Management, select New Replication Group
1.2.2 Select Multipurpose replication group, click Next
Multipurpose replication group is used in most use cases. If you are going to configure replication of folders between two or more servers, where the same data will reside on all servers that are members of the replication group, then you will select this option.
Replication group for data collection is only used if you want to transfer data from a branch server to a hub server, and then use backup software to back up the data on the hub server.
1.2.3 Name the replication group, click Next
1.2.4 Add members to the replication group, click Next
1.2.5 Select a Topology, click Next
1.2.6 Select the replication schedule and bandwidth, click Next
1.2.7 Select the primary member, click Next
This will be the server that will be authoritative during initial replication. Which means that its content will be replicated to other members. If there is a conflict, this server will win. Post the initial replication, there is no such thing as a primary/master member. But the first replication will need a primary/master member to ensure that all folders have the same starting point.
1.2.8 Select the folders on the primary member, you want to be replicated to the other members of the replication group
You can also edit the name of the replicated folders and NTFS permissions of the replicated folders on this page.
Click Next when you are ready to continue
1.2.9 On the next page(s) you get to define local path and membership status of the replicated folder(s) on the secondary members. Click Edit
1.2.10 You are given following options
Membership status Disabled: The replicated folder will not be replicated to this member.
Membership status Enabled: The replicated folder will be replicated to this member.
Local path of folder: Configure the path the replicated folder will be stored at on this member.
Make the selected replicated folder on this member read-only: This means that this member will only receive data, and not transmit it, because it will not be possible for users to manually append data to the replicated folder stored on this member.
1.2.11 Back on the previous page, click Next
1.2.12 Click Create
1.2.13 Click Close
.
1.3 Configuration options in DFS-R
.
1.3.1 Editing Replication Schedules and Bandwidth
The replication schedule defines when replication can occur for the replicated folders that are part of the replication group. And how much bandwidth replication can consume.
1.3.1.1 To edit the schedule for an entire group right-click the replication group → Edit Replication Group Schedule
1.3.1.2 To edit the schedule and bandwidth for a specific connection between two members, click on the Connections tab of the Replication Group. Then right-click a connection and enter its properties sheet. On the Schedule tab of the properties sheet, you can define a custom schedule for this specific connection.
.
1.3.2 Managing Connections
1.3.2.1 On the Connections tab of the replication group, you will find all send and receive connections between all members of the replication group. By right-clicking a connection, you will have these options on the context menu,
Disable lets you disable replication for this specific connection.
Replicate Now starts replication for the connection. The receiving member will start receiving data.
Properties will let you enter the properties sheet, where you can enable/disable replication and remote differential compression for the connection.
1.3.2.2 If you want to create a new connection between members of the replication group, then right-click the replication group and select New Connection. This will bring up the new connection dialog box, where you can define sending and receiving members. You can also create a custom schedule for the connection, and create a second connection in the opposite direction automatically.
.
1.3.3 Configuring properties of replication group members
1.3.3.1 On the Memberships tab of the replication group, you will find the different folders that are being replicated between members. By right-clicking a folder, you will have these options on the context menu,
Delete Member deletes the member from that replicated folder
Make read-only makes the folder read-only on that particular member. Meaning no files can be written on it by users, and the replication will in practice be one-way, as this member will only receive data from the other members because data on it will never be changed by users.
Disable lets you disable replication for this particular folder on a specific member. After a disabled member is enabled, the member must complete an initial replication of the replicated folder. Initial replication will cause about 1 KB of data to be transferred for each file or folder in the replicated folder, and any updated or new files present on the member will be moved to the DfsrPrivate\PreExisting folder on the member and will be replaced with authoritative files from another member. If all members are disabled, then the first member enabled becomes the primary/authoritative member.
Properties will let you enter the properties sheet,
⇒ General tab gives information only,
⇒ Replicated folder tab also only gives information,
⇒ Staging tab lets you define the path and size of the staging folder.
By default the staging folder is located under the local path of the replicated folder. The staging folder is used to store files before sending and receiving them. When remote differential compression is enabled, calculations are performed on files in the staging folder to determine which blocks need to be replicated. The receiving folder uses its own staging folder to build the modified file.
You can improve performance and reduce the likelihood of replication failing by optimizing the size of the staging folder. Microsoft recommends that at a minimum the staging folder quota should be at least the size of the 32 largest files in the replicated folder, and for best possible performance on the file server, the size of the staging folder quota should be set as close as possible to the size of the replicated folder and it should be located on a different physical disk than the replicated folder.
The disk space allocated to the staging folder will not be reserved, but only be consumed when needed.
⇒ Advanced tab
On this tab you can define the size of the Conflict and Deleted folder. You can also enable/disable moving deleted files to this folder. When the same file is edited on different replicas, the user who performs the most recent save operation has the changes saved and replicated across servers. If files with the same name and path are created on two different replicas, before replication occurs, then the file that was created first is kept and replicated across servers. The losing files are moved to the conflict and deleted folder.
.
1.3.4 Configuring properties of replicated folders
1.3.4.1 On the Replicated Folders tab of the replication group, you will find the different folders that are being replicated between members. By right-clicking a folder, you will have these options on the context menu,
Share and Publish in Namespace option lets you share the replicated folder, or share and publish it in a namespace.
Properties will let you enter the properties sheet,
File filter lets you define which files will not be replicated. By default the following file types will not be replicated,
~*, Microsoft Office temporary files
*.bak, backup files
*.tmp, temporary files
Subfolder filter lets you define which folders will not be replicated.
.
1.3.5 Replication topologies
Replication topology determines how members of a replication group replicate with each other. There are two different replication topologies you can choose between.
Hub And Spoke
This topology requires three or more members. Central servers are designated as hub servers, while branch servers are designated as spoke servers. Spoke servers replicate with hub servers, but do not replicate directly with other spoke servers.
Full Mesh
All servers replicate directly with each other.