Distributed Data Center Virtualization using Windows Server 2008 Hyper-V and Failover Clustering beta release* *The

e clustered disk removal section will become obsolete once the solution ships in early August
Using Sanbolics Kayo FS File System and Microsoft Failover Clustering to enable flexible migration of active virtual machines across physical hosts running Windows Server 2008 with Hyper-V

Installation Guide

Step-by-Step Guide for clustering Hyper-V virtual machines with Sanbolics Kayo FS
Table of Contents

1 2 3

Software Components ...................................................................................................................... 2 Installation Requirements ................................................................................................................ 2 Steps for configuring a simple Hyper-V host clustering with Kayo FS ......................................... 3 3.1 3.2 Creation and configuration of the Hyper-V virtual machine: ................................................... 3 Failover Clustering Setup .......................................................................................................... 5

1 Software Components
The state and virtual local disk for a Hyper-V instance is contained within a VHD file which typically resides on the file system of the host server when used for single server partitioning. In order to create a distributed virtualization architecture in which a Hyper-V VM instance can be dynamically moved to any host server, all host servers must have concurrent access to the VHD file. Sanbolics Kayo FS clustered file system is designed to provide multiple Windows servers with concurrent access to shared data on a Fibre Channel or iSCSI SAN. By maintaining the .VHD files for all Windows Server 2008 with Hyper-V on SAN storage and installing Kayo FS file system on each host server, each physical host has concurrent access to all .VHD files. A Hyper-V VM instance can only be active on only one host at a time. Microsoft Failover Clustering is typically used for high availability applications to failover application to a standby server when the primary server fails. It can also be used as a tool to migrate the state of a Hyper-V VM instance across physical hosts while the application remains active. Microsoft Failover Clustering provides a central graphical interface for migrating virtual machines. Much of the complexity of typical failover configurations is removed when using Kayo FS clustered file system since there is no need to move a storage resource between servers-- all hosts have concurrent access to the storage resource, so only the application needs to be moved.

2 Installation Requirements
In order to successfully complete the clustering procedure, the following software and hardware requirements must be met (this paper describes a configuration with two physical hosts, but host environments of up to sixteen hosts can be configured as well): 1.1 OS requirements - Windows Server 2008 Enterprise or Datacenter edition 1.2 Software requirements Hyper-V feature, Failover Clustering feature and Kayo FS 1.3 Failover Clustering - fully functioning two-node cluster, configured with an appropriate quorum resource http://technet2.microsoft.com/windowsserver2008/en/library/13c0a922-6097-4f34-ac6418820094128b1033.mspx?mfr=true 1.4 Storage requirements access to shared storage by Fibre Channel, SCSI or iSCSI or other qualified block level access 1.5 Network requirements at least 2 network adapters dedicated to network communication 1.6 Hardware requirements Two identical server computers which have passed the Failover Clustering Hardware Validation [ http://technet2.microsoft.com/windowsserver2008/en/library/13c0a922-6097-4f34-ac6418820094128b1033.mspx?mfr=true ]

3 Steps for configuring a simple Hyper-V host clustering with Kayo FS

There are 2 Windows Server 2008 x64 physical hosts named Y-21-2k8x64 and Y-22-2k8x64. They have identical hardware and comply with Microsofts Failover Clustering Hardware validation and have a shared Kayo FS formatted volume with the same drive letter O: and a multitude of shared Physical Disks.

2.1

Creation and configuration of the Hyper-V virtual machine:

2.1.1 Open the Hyper-V Manager from the Administrative tools:

2.1.2 Start the New Virtual Machine Wizard:

2.1.3 follow the steps

select the appropriate network

and virtual hard disks

and click Finish.

If you have followed the above steps you will end up with a new Hyper-V virtual machine configured. You must install an OS on the Hyper-V virtual machine before continuing.

This concludes the first part of configuring the Hyper-V VM and now the 2 nodes are ready to be configured for failover clustering.

2.2

Failover Clustering Setup

The Failover Cluster configuration will be based on a Node Majority Quorum, which is used only for the demonstration purposes. For a list of available quorum types please refer to http://technet2.microsoft.com/windowsserver2008/en/library/13c0a922-6097-4f34-ac6418820094128b1033.mspx?mfr=true

2.2.1 Open the Failover Cluster Management and follow the steps below:

2.2.2 Enter the name for the cluster

2.2.3 validate the cluster configuration

2.2.4 configure the IP address for the cluster administration purposes

2.2.5 and create the cluster.

2.2.6 Now that the cluster has been created and needs additional tuning

2.2.7 By default, all the shared physical disks are clustered by the Failover Clustering and we have to remove them from the clustered storage since they are now being reserved for its usage*
*This is a required step in order to be able to use the shared disk independently from the Failover Cluster. It will become obsolete once the solution ships in early August

2.2.8 Once the Clustered storage is remove, we can start with the Virtual Machine clustering wizard. Make sure that the Virtual Machine is in stopped state before running the wizard:

2.2.9 Once the Virtual machine is clustered, there are a couple of things to do in order to ensure proper and fast operation of the Failover Cluster. The cluster disk must be removed from the Hyper-V service and then deleted from the Clustered storage*:
*This step will become obsolete once the solution ships in early August

2.2.10 When the Kayo FS formatted drive is removed from the Failover Clustering, it must be brought online from the 2 physical servers:

2.2.11 Now the Failover Clustering is properly configured for usage with Hyper-V VMs residing on a Kayo FS formatted volume and the clustered Hyper-V VM can be fast and easily migrated between the nodes:

You can add as many virtual machines into the Failover Clustering as you want by repeating the steps from 3.1 and 3.2.8 to 3.2.11. This will give you a reliable virtualized data center which can be managed fast and effectively:

Sanbolic Inc. 304 Pleasant Street, 2nd Floor Watertown, MA 02472 phone: 617 833 4249 fax: 617 926 2808 url: www.sanbolic.com email: sales@sanbolic.com