01 Sep / 2016
Author: Oliver Gehrmann Tags: There is no tags Comments: 0

Bookmark and Share

Windows Server 2012 R2 Failover Clustering with virtual iSCSI Storage

Well, I won’t start with a big opener because this is going to be a damn long contribution. Failover Cluster, huh? Pretty boring. Aren´t there 25973534563 other guides about iSCSI and Failover Clusters out there already?
Probably, but I didn’t found a single one with a decent explanation about the single steps from A to Z. The reason why I am writing about this is because we had to install and configure a SQL Failover Cluster on a single Windows Server 2012 R2 Hyper-V Host. And there is no other way of creating a Cluster without having an Hyper-V Failover Cluster underneath your hands.

So here we go.

The guide consists of four parts:

  • Part 1: Installation of the iSCSI Target
  • Part 2: iSCSI Initiator Configuration for the Nodes
  • Part 3: Node Storage Configuration
  • Part 4: Failover Cluster Configuration

Be sure to add enough virtual disks to your File Server before starting this process because you need disk space for the iSCSI targets.

Part 1

Let us start with the iSCSI Target Installation. Open Server Manager and navigate to “Add roles and features”.

Install the Fileserver and iSCSI Roles on your Target Server.

In detail:

  • File and Storage Services > File and iSCSI Services > File Server
  • File and Storage Services > File and iSCSI Services > iSCSI Target Server
  • File and Storage Services > File and iSCSI Services > iSCSI Target Storage Provider (VDS and VSS hardware providers)

Open the Server Manager and navigate to File and Storage Services -> iSCSI

Set the iSCSI Virtual Disc Location

Set the iSCSI Disk Name

iSCSI Disk Size. If you have a high level of disk activity e.g. SQL you should take a fixed size. Otherwise select Dynamically.

Create a new iSCSI Target

Set a target name

Add the Access Servers (Nodes) via ID or IP

Access Server Overview

Configuration Overview

Part 2

Now we’ll connect the Nodes with the target server over the iSCSI Initiator. You have to do the following steps on all Nodes!

Open the iSCSI Initiator on the Nodes

Type in the name of your target server and press quick connect. The Node will automatically connect to the target server

After the successful connection auto configure your disks in the ‘Volume and Device’ tab.

Part 3

Ok, preparation nearly done. We’ll configure the connected disks now. You have to do the following steps on all nodes.

Once connected, the iSCSI Disks will appear offline. Diskpart will fix this!
Start a new administrative command prompt and execute the following commands:

  • Diskpart
  • List disk
  • Select disc (iSCSI Disk which should appear offline)
  • Online disk
  • Attributes disk clear readonly
  • List partition (There should’t be any partitions now otherwise use the command ‘clean’)
  • Create partition primary
  • List partition
  • Select partition
  • Format fs=ntfs quick label= (DiskLabel is the label you want assign to the disks)
  • List volume
  • Select volume
  • Assign letter= (If you don’t select a letter, the next available letter will be used)

Repeat this steps for all new assigned iSCSI Disks on all Nodes. Afterwards verify with ‘list volumes’ that all of the volumes are formatted.

Part 4

Final Step! Let’s have some failover fun.
It’s important to configure the network settings for all involved servers correctly, otherwise the validation test will fail or report warnings will come up. The best practices network configuration for a basic cluster with two nodes and one file server look like this:

I have done it with virtual HyperV Switches. Each Node has his own iSCSI network connection to the target server.
Furthermore, the Nodes communicate among themselves over a heartbeat network.

Open the Server Manager and install the Failover Clustering Feature for each Node.

Start the Failover Cluster Manager. It doesn’t matter on which Node

Start the Cluster Validation Wizard. The following validation will check if all prerequisites for the cluster are met.

Select all Cluster Nodes

Run all Tests

There should be no errors or warnings, otherwise you will get trouble while creating the cluster. The Cluster Creation Wizard will be started after you close the Validation Report

Set a Cluster Name and IP Adress

Untick the option to add all storage, otherwise all available local disks will be added to the cluster!

The Cluster is now created and ready to use.

Have fun with your new Failover Cluster and don’t hesitate to contact us for any questions.

Blog - Kommentar verfassen

Please notice, that all fields markes by * have to get filled out in order to be able to publish the comment.

Security check *