Before you read this post, I strongly recommends you to go through the following post:-
Configuring Microsoft Cluster on Windows Server 2003
http://www.adshotgyan.com/2016/04/configuring-microsoft-cluster-on.html
Once you have configured a two node cluster on Windows 2003, now is the time to configure an application which can use this cluster
For this lab, I will be configuring DFS on this Cluster
So here we have two disk, one for Quorum and the other for Storage. The storage disk will store all our application related data
Create a Root Folder on this Storage Drive. For now, grant "Everyone" - Full Control
Right click on the cluster name and select the option "Configure Application"
Cluster Application wizard launches
Create a New Virtual Server
Wait.. Virtual Server? Were we not configuring an application?
Virtual Server: Collection of services/applications that runs on multiple server but for represent themselves as a single server. When you deploy an application, it requires a "Name" and an "IP Address" to be associated/linked. The Name and the IP will make a resource group. The resource group will then be a part of virtual server. Once you create a virtual server, its name and IP will be registered in DNS as a machine name.
Resource Group: Collection of resources that runs on a specific node where as a Virtual Server consist of a Resource Group that can run on multiple nodes
As stated above, a virtual group essentially consist of a resource group. So we either need to create a new resource group or can specify an existing resource group created earlier
For simplicity, I will be naming the resource group and virtual server with names which will explain what they are at a later point in time
So we will name the the "Resource Name" as "DFS Resource Name"
"Network Name" as "DFSNETNAME"
Also, provide an IP address. This will be a unique IP address and the network name will be used by the client machines to connect to this virtual server/resource group
We can specify advance settings for the Network Name, IP Address and Resource Group
I will not discuss these settings as this post is concentrated on installation of DFS application on Microsoft Cluster on Windows Server 2003 and for the normal working of DFS on cluster we will keep these settings as "Default". For these settings in details, I will create another post
Back to our create cluster resource wizard
In Windows Server 2003 Cluster, even though we are creating a DFS Root, the type of cluster resource will be "File Share"
A list of all the resource types which can be added in the cluster
Note that DFS is not listed as a Resource Type
Provide a name for the "Application Resource"
We will name it as "DFS_Application_Resource_Name"
Some advance properties of the Resource Name
Time to provide the Share Name which will host our DFS Root
ShareName: MyDFSData
Path: S:\MyDFSData
Permissions Tab
For now, we will grant full control to everyone
Advanced Tab
Now its the time to define what kind of file share this folder is. In Windows Server 2003 Cluster, we do not have any Resource type as DFS Root. That is why we specified a File Share and now we will specify what kind of file share it is (DFS Root)
Caching Tab
To specify it want the data which the users will access in this DFS Root to get cached locally on their client machine or not
You can select any option here for lab purpose, however the option will change in production environment
Ok. So we now have an error while creating the resource. This requires a dependency. And why is that required. It is because in order to bring a resource online (like dfs) its name should be available. So we need to have at least one dependency (Which should be network name)
Lets go into the Resource Name Advanced Properties
Add the Network Name as the dependency
Done
The Resource Group has been created now however this is offline
Right click on the resource group and select the option "Bring Online"
Online Pending
And finally, all the resources (IP Address, Network Name and the File Share is Online now on Node2)
Now on a Windows 2008 R2 Server, open the DFS Snap-In
Wait, you said on a different server which is not a Node of a cluster?
Yes, I said that because this snap-in will just "Connect" and not host the DFS
Add Namespace to Display
Select Node2 and click on Show Namespaces
See, The "Network Name" (DFSNETNAME) is getting displayed
And its done
But you need to understand this. This server is NOT hosting this DFS, its getting hosted on Node2 (which is the active node for this resource right now and then once the failover happens this resource will switch over to Node1)
Lets access this path
Now, where is this path taking us to ? Node 2 ?
The answer is No, Its taking us to SAN as the data is NOT stored on Node2 its stored on the SAN which we have created earlier and which was Presented to Node2
Lets create a folder here
While doing some testing, I have changed the node to Node1
And you can verify that the data is still on the S drive on Node1 which was earlier created on Node2. The data is on S drive which happens to be on SAN and is not local to Node1 or Node2. So you will be able to view the data on both the server which so ever server will be active as the drive will be active only on one server at a time
Now lets move the group back to Node2
Done
And again, we still have the data on the drive, accessible from DFS
Configuring Microsoft Cluster on Windows Server 2003
http://www.adshotgyan.com/2016/04/configuring-microsoft-cluster-on.html
Once you have configured a two node cluster on Windows 2003, now is the time to configure an application which can use this cluster
For this lab, I will be configuring DFS on this Cluster
So here we have two disk, one for Quorum and the other for Storage. The storage disk will store all our application related data
Create a Root Folder on this Storage Drive. For now, grant "Everyone" - Full Control
Right click on the cluster name and select the option "Configure Application"
Cluster Application wizard launches
Create a New Virtual Server
Wait.. Virtual Server? Were we not configuring an application?
Virtual Server: Collection of services/applications that runs on multiple server but for represent themselves as a single server. When you deploy an application, it requires a "Name" and an "IP Address" to be associated/linked. The Name and the IP will make a resource group. The resource group will then be a part of virtual server. Once you create a virtual server, its name and IP will be registered in DNS as a machine name.
Resource Group: Collection of resources that runs on a specific node where as a Virtual Server consist of a Resource Group that can run on multiple nodes
As stated above, a virtual group essentially consist of a resource group. So we either need to create a new resource group or can specify an existing resource group created earlier
For simplicity, I will be naming the resource group and virtual server with names which will explain what they are at a later point in time
So we will name the the "Resource Name" as "DFS Resource Name"
"Network Name" as "DFSNETNAME"
Also, provide an IP address. This will be a unique IP address and the network name will be used by the client machines to connect to this virtual server/resource group
We can specify advance settings for the Network Name, IP Address and Resource Group
I will not discuss these settings as this post is concentrated on installation of DFS application on Microsoft Cluster on Windows Server 2003 and for the normal working of DFS on cluster we will keep these settings as "Default". For these settings in details, I will create another post
Back to our create cluster resource wizard
In Windows Server 2003 Cluster, even though we are creating a DFS Root, the type of cluster resource will be "File Share"
A list of all the resource types which can be added in the cluster
Note that DFS is not listed as a Resource Type
Provide a name for the "Application Resource"
We will name it as "DFS_Application_Resource_Name"
Some advance properties of the Resource Name
Time to provide the Share Name which will host our DFS Root
ShareName: MyDFSData
Path: S:\MyDFSData
Permissions Tab
For now, we will grant full control to everyone
Advanced Tab
Now its the time to define what kind of file share this folder is. In Windows Server 2003 Cluster, we do not have any Resource type as DFS Root. That is why we specified a File Share and now we will specify what kind of file share it is (DFS Root)
Caching Tab
To specify it want the data which the users will access in this DFS Root to get cached locally on their client machine or not
You can select any option here for lab purpose, however the option will change in production environment
Ok. So we now have an error while creating the resource. This requires a dependency. And why is that required. It is because in order to bring a resource online (like dfs) its name should be available. So we need to have at least one dependency (Which should be network name)
Lets go into the Resource Name Advanced Properties
Add the Network Name as the dependency
Done
The Resource Group has been created now however this is offline
Right click on the resource group and select the option "Bring Online"
Online Pending
And finally, all the resources (IP Address, Network Name and the File Share is Online now on Node2)
Now on a Windows 2008 R2 Server, open the DFS Snap-In
Wait, you said on a different server which is not a Node of a cluster?
Yes, I said that because this snap-in will just "Connect" and not host the DFS
Add Namespace to Display
Select Node2 and click on Show Namespaces
See, The "Network Name" (DFSNETNAME) is getting displayed
And its done
But you need to understand this. This server is NOT hosting this DFS, its getting hosted on Node2 (which is the active node for this resource right now and then once the failover happens this resource will switch over to Node1)
Lets access this path
Now, where is this path taking us to ? Node 2 ?
The answer is No, Its taking us to SAN as the data is NOT stored on Node2 its stored on the SAN which we have created earlier and which was Presented to Node2
Lets create a folder here
And you can verify that the data is still on the S drive on Node1 which was earlier created on Node2. The data is on S drive which happens to be on SAN and is not local to Node1 or Node2. So you will be able to view the data on both the server which so ever server will be active as the drive will be active only on one server at a time
Now lets move the group back to Node2
Done
And again, we still have the data on the drive, accessible from DFS
