Protecting your Virtualized Server
Investment
Using the CA ARCserve® Family of Products with Hyper-V™
Technology—A Powerful Combination
This white paper is for informational purposes only. MICROSOFT AND CA MAKE NO WARRANTIES,
EXPRESS OR IMPLIED, IN THIS SUMMARY.
Microsoft, Hyper-V, Microsoft Exchange Server, Microsoft Hyper-V Server, Microsoft Office SharePoint
Server, SQL Server, System Center, Windows and Windows Server are trademarks of the Microsoft
group of companies.
Document published May 2010
Contents
Introduction................................................................................................................................................3
What is Hyper-V? ......................................................................................................................................3
What is the most important consideration in moving workloads and applications from a physical
environment to a virtual environment? ......................................................................................................3
®
What is the CA ARCserve Family of Products? ......................................................................................4
What is CA ARCserve Backup? ................................................................................................................4
What is CA ARCserve D2D?.....................................................................................................................6
What is CA ARCserve Replication? ..........................................................................................................7
What is CA ARCserve High Availability? ..................................................................................................8
Why companies are adopting Hyper-V for their virtualization infrastructure ...........................................10
Reliable Disaster Recovery Support ...................................................................................................10
Environmentally Responsible Technology ..........................................................................................11
Considerations in implementing virtualization .........................................................................................11
Determining Application Requirements ..............................................................................................11
Determining the Architecture of Your Virtualized Server Environment...............................................11
Determining Hardware Resource Requirements ................................................................................12
Processor Resources...................................................................................................................... 12
Memory Resources ......................................................................................................................... 12
Storage Resources ......................................................................................................................... 12
High-Performance Hardware Requirements................................................................................... 13
Determining Backup, Replication, and High Availability (HA) Requirements .....................................13
Backups That Use Native Applications ........................................................................................... 14
Guest-Level Backups ...................................................................................................................... 14
Hypervisor-Level Backups .............................................................................................................. 15
Determining Replication Requirements ..............................................................................................15
Determining High-Availability Requirements ......................................................................................16
High Availability within a Data Center ............................................................................................. 16
High Availability Across Data Centers ............................................................................................ 17
Implementing backups of Hyper-V environments using CA ARCserve Backup .....................................18
Full Weekly Backup with Daily Differential Backups ....................................................................18
Image-Level VM Restores and Granular Folder\File-Level Restores ..........................................19
Multiple Remote Office Backup .......................................................................................................19
Physical to Virtual Server Migration ...............................................................................................21
Implementing backups of Hyper-V environments using CA ARCserve D2D ..........................................21
Establishing a high-availability environment by using Hyper-V and CA ARCserve High Availability .....22
Basic High-Availability Scenario: Two Virtual Machines on a Single Physical Machine ..........22
Adopting Server Virtualization for Business Continuity and Disaster Recovery 1
Large-Scale High-Availability Scenario: Physical-to-Virtual Environment with Advanced Data
Recovery ............................................................................................................................................22
Conclusion...............................................................................................................................................23
Adopting Server Virtualization for Business Continuity and Disaster Recovery 2
Introduction
In today’s competitive business environment, your IT infrastructure must accommodate rapidly
changing business needs and unforeseen problems with minimal impact on day-to-day operations. But
this level of flexibility, scalability, and high availability can be expensive and difficult to deploy and
maintain, especially for small and midsize businesses. Businesses and organizations of all sizes need
control over their changing environment and having total protection, recovery and availability for their
systems, applications and data is critical. System downtime and data loss can cause irreparable harm.
Server virtualization is now a mainstream technology and virtual servers demand the same levels of
protection as physical servers.
About Microsoft Virtualization Solutions
Microsoft provides a complete suite of technologies to enable an integrated, end-to-end, virtualized
infrastructure. Using products that span the desktop to the data center, Microsoft technologies bring
capacities online in real-time, as needed; streamline and provision applications, services and data on-
demand; accelerate backup and recovery, and enhance availability to protect against system failure
and service interruptions. Microsoft's extensive partner ecosystem complements and extends the
Microsoft virtualization toolset with products for desktops, servers, applications, storage, and networks.
Together with our partners, we deliver the most robust, complete solutions for the virtualized
infrastructure. For more information about Microsoft Joint Virtualization Solutions, visit
http://www.microsoft.com/virtualization/partners.mspx.
Server virtualization technology helps significantly reduce overall IT capital expense and operational
™ ®
costs. Hyper-V virtualization technology, a feature of the Windows Server 2008 R2 operating
system, is an enterprise-class virtualization solution that can be quickly deployed in any business
environment. By using Hyper-V with the CA ARCserve Family of Products, you can build an IT
infrastructure with robust protection, recovery and availability capabilities to protect your systems,
applications and data while reducing total cost of ownership (TCO).
What is Hyper-V?
Hyper-V is the next-generation hypervisor-based server virtualization technology. Available as an
integral feature of Windows Server 2008 R2, it enables you to implement server virtualization with
ease. You can use Hyper-V to make the best use of your server hardware investments by
consolidating multiple server and application roles as separate virtual machines (VMs) running on a
single physical machine. With Hyper-V, you can also efficiently run multiple different operating
®
systems—Windows , Linux, and others—in parallel, on a single server, and fully leverage the power of
x64 computing.
And with more and more cores getting added to a CPU, the adoption of virtualization is going to
increase at a faster pace than ever before.
With Windows Server 2008 R2, Hyper-V adds additional functionality like improved memory
managemement and performance, greater scalability and Live Migration which provides the ability to
move virtual machines without application interruption to users.
What is the most important consideration in moving
workloads and applications from a physical environment to
a virtual environment?
Given the benefits of virtualization, when you want to move your workloads or applications into
virtualized infrastructure, some of the top considerations of your backup\storage administrator would
be ― How do I protect the data in the virtualized infrastructure? How do I make sure that there is
minimum impact on the existing backup processes? How do I make sure that my existing restores or
recoveries processes don’t change? How do I make sure I know the backups of my data in the virtual
infrastructure is backed up and restored in the most efficient way?‖ How do I limit my risk with more
virtual servers residing on a single physical server now?
These are some of the questions all storage\backup administrators will have in their mind as they see
more servers getting virtualized.
What is the CA ARCserve® Family of Products?
The CA ARCserve Family of Products gives you
control over your changing business by delivering total
protection, recovery and availability for your systems,
applications and data. With the expanded and
®
integrated offerings of CA ARCserve Backup, CA
® ®
ARCserve D2D, CA ARCserve Replication and CA
®
ARCserve High Availability, we continue to fulfill our
ore
promise to be ―M than Backup.‖ The ARCserve
Family of Products is the only solution that offers a full
range of products from Bare Metal Restore to fully
automated failover for High Availability You get a
complete strategy to manage your backup and
recovery responsibilities in both virtual and physical
environments.
What is CA ARCserve Backup?
CA ARCserve Backup offers enterprise-class data protection for distributed servers, databases, and
applications running in both physical and virtual environments, including Hyper-V. This high-
performance solution unites innovative data deduplication technology, powerful SRM reporting,
centralized management with infrastructure visualization and robust security measures — making it
one of the market’s most secure, reliable and efficient methods of protecting the information residing
on physical and virtual servers. More importantly, it has become an industry leader by delivering
enterprise-class data protection for a wide range of operating environments in a simple, easy-to-install
package..It includes the following critical features:
Data deduplication reduces disk storage for backups, increases recovery points, and reduces
recovery time objective (RTO)
Granular file and folder restores from image-level VM backups provide the ability to
perform a full VM restore or a granular file or folder restore from a single VM backup.
Powerful dashboard with storage resource management reporting provides a single-
pane-of- glass view through which administrators can monitor and manage storage resources
and backups. The dashboard displays statistics regarding recovery points of physical and
virtual machines, CPU, memory, disk, network cards, volume utilization, volume fragmentation,
and server OS and service pack levels that help ease management of the environment and
enable proactive problem avoidance. Performance Key Indicators allow thresholds to be set for
critical resources such as CPU, memory and disk space, and alerts when the threshold has
been reached.
Granular Restore of Active Directory objects from a System State backup without having to
reboot the AD Domain Controller. Single System state backup allows for both object level and
system level restore saving time.
Infrastructure Visualization provides an easy-to-read network diagram view of the entire
environment including all of the servers, storage and other devices showing how they relate to
each other and to the CA ARCserve Backup media server(s). Easy drill down on each
component offers status, backup methodology and other key information.
Granular Restores of SharePoint enables you to restore individual files, sites, and sub-sites
Adopting Server Virtualization for Business Continuity and Disaster Recovery 4
within a SharePoint environment for fast recovery. The entire environment can also be
restored if required and all types of restoration can be achieved using a single-pass full-farm
backup.
Security with password key management for encryption enables administrators to easily
secure data without the burden of memorizing passwords. CA ARCserve Backup has flexible
role-based administration and supports complete auditing capabilities to reduce administrative
efforts.
Extensive Microsoft application and OS support is available for a wide range of Microsoft
®
®
applications including Microsoft SQL Server , Microsoft Exchange Server, Microsoft Office
®
SharePoint Server, Microsoft Hyper-V Server, and all Windows, Unix, and Linux operating
systems.
All of these features, along with its affordable, flexible, and simple licensing structures, make CA
ARCserve Backup a value leader in the market delivering enterprise-class data protection for a wide
range of operating environments. Suitable for companies of all sizes, it offers leading-edge backup and
data protection technology that efficiently and cost-effectively helps reduce the costs of protecting
business-critical data and applications running on both virtual and physical servers. Certified across an
extensive range of industry platforms and applications, it provides compatibility and support for
heterogeneous environments.
Protection of your virtual server infrastructure
CA ARCserve Backup integrates with Microsoft Hyper-V Servers starting with complete integration
within the user interface—no scripting required! Some of its salient features are:
Auto discovery and protection of VMs. With the flexible nature of the virtual infrastructure,
virtual machines can appear and disappear dynamically, creating the potential for missed
backups if the backup administrator does not realize these changes have occurred. To avoid
this problem, CA ARCserve Backup has automatic discovery of VMs to help ensure that they
are all backed up regardless of how the environment changes.
Granular file\folder restores. One-step granular file recovery process allows restoration of
individual files and directories from an image-level backup of the VM directly from the backup
media. This requires just one full VM image backup; from that one backup it is possible to
restore either the full VM or granular file restores.
Fast and efficient restores. When a granular file\folder restore is required, CA ARCserve
Backup uses a patent-pending technology to restore the granular files\folders directly from the
backup media to the VM without staging. This ensures extremely fast and efficient restoration
of the data.
Flexibility and reduced storage requirements. They support mixed backups for weekly full
backup and daily incremental or differential backups.
Small backup window. Using multiple, controlled data streams, CA ARCserve Backup can
simultaneously back up data from multiple VMs across multiple Hyper-V servers to disk or to
tape, thus helping to reduce the backup window and improve backup efficiencies.
Adopting Server Virtualization for Business Continuity and Disaster Recovery 5
No change in processes. Backing up a virtual infrastructure is as simple as backing up any
physical machine. There are no custom scripts—it simply works out of the box.
Quick disaster recovery of physical servers to virtual servers. Achieve quick and efficient
disaster recovery into a virtual machine from the backups of your physical servers.
Migration. CA ARCserve Backup also aids in the migration of production applications from
physical to Windows Server 2008 Hyper-V virtual servers.
What is CA ARCserve D2D?
CA ARCserve D2D is a disk-based backup product designed to provide the perfect combination of fast,
set
simple and reliable protection and recovery for all of your business information. It offers a ― it and
forget it‖ solution that makes protecting and recovering your data a snap. The single snapshot backup
allows you to restore files, volumes, applications or the entire system to any server in both physical
and virtual environments. CA ARCserve D2D is the only product that has Bare Metal Restore to
2
dissimilar hardware and block-level infinite incremental snapshots (I Technology) in one complete
2
package. New I Technology means you only need to perform 1 full backup with incremental only from
then on saving time and storage space over other backup products, reducing backup windows, the
amount of data transferred over the network and load on the application and physical server the VM is
2
running on. I or infinite incremental backups refer to the function of performing an initial full backup
and incremental backups from then on eliminating the need for another full backup. CA ARCserve
D2D will also perform a physical to virtual bare metal recovery which can be used to easily migrate
your applications from physical to virtual servers.
CA ARCserve D2D delivers critical features including:
Block Level infinite Incremental Backup Like deduplication, this patent-pending technology
uses less storage space, performs backups faster and puts less of a load on your production
2 ™
servers. Disk-to-disk backups are based on CA’s block-level I Technology that has been
designed to intelligently manage backing up only blocks of data that have changed since the
last backup and presenting a consolidated point-in-time view of the protected volume for
multiple recovery types.
Single snapshot Backup with Four Restore Types Provides the ability to rapidly restore
files, volumes, databases, or the entire system to any server in both physical and virtual
environments from a single pass backup. One snapshot helps ensure recovery of the data in
the most efficient manner.
Bare Metal Recovery to Dissimilar Hardware Bare Metal Restore offers a lightning-fast way
to recover a crashed server to the same or dissimilar hardware. It simply paints the picture of
the original drive back onto the new drive and restarts the system – reducing a process that
can potentially take as long as 36 hours to just minutes.
Adopting Server Virtualization for Business Continuity and Disaster Recovery 6
The CA ARCserve D2D Management Screen
The CA ARCserve D2D management screen provides an easy-to-read overview of backup status, available
recovery points and user actions. It incorporates Dynamic Documentation and Web 2.0 technology to provide
an extremely easy to use interface that delivers the latest information to your desktop and allows for
management from any location using any Web client.
What is CA ARCserve Replication?
® ™
CA ARCserve Replication, formally known as CA XOsoft Replication, provides continuous data
protection along with an easy way to migrate data from one server to another and from one location to
another for disaster recovery. It is also used to aggregate data from remote offices to a central site for
backup and archiving. It is a solution that delivers reliable data protection for a wide range of operating
environments and includes the following critical features:
Host-based data replication across the LAN or WAN between production servers and replica
servers for data protection, availability and migration.
Continuous or Periodic Replication to help balance network resources with the desired level
of protection
Data Rewind for continuous data protection (CDP) complements any backup solution and
protects data between periodic backups for faster recovery time after accidental or malicious
loss or damage.
CA ARCserve Assured Recovery provides fully automated recovery testing to help ensure
recovery readiness, with no disruption to the production or DR environment. Optional VSS
snapshots may be performed after system validation for application-consistent backup.
Unified, web-based centralized management and no-reboot deployment and maintenance
makes CA ARCserve Replication a breeze to manage..
Extensive Microsoft application and OS support to protect a wide range of applications
including Microsoft Exchange, SQL Server, SharePoint, IIS, Dynamics CRM, Microsoft Hyper-
V Server, and most Windows operating systems. Oracle and Blackberry Enterprise Server are
also supported. Other applications may be protected by creating a simple script.
Adopting Server Virtualization for Business Continuity and Disaster Recovery 7
All of these features, combined with its affordable, flexible, and simple licensing structures, make CA
ARCserve Replication a value leader in the market delivering enterprise-class data protection tailored
for the SMB market. Suitable for companies of all sizes, it offers leading-edge replication and recovery
testing technologies that efficiently and cost-effectively help reduce the costs of protecting data on both
virtual and physical servers at the data center and remote offices, no matter where stored (DAS, NAS
and SAN). Certified across an extensive range of industry platforms and applications, it provides
compatibility and support for heterogeneous environments.
Continuous data protection for your virtual server environment
CA ARCserve Replication integrates easily with your Microsoft Hyper-V Servers. Some of its salient
features include:
Physical-to-physical (P2P), Physical-to-Virtual (P2V) and Virtual-to-Virtual (V2V)
replication. Provides the protection of data running on Hyper-V Servers and also allows the
use of Hyper-V Servers for replica or failover servers to help you reduce overall DR costs.
Flexible Hyper-V deployment options. Offers flexibility of replicating at the hypervisor level
or replicating individual guest operating systems/applications without the requirement of
installing CA ARCserve Replication on each of the guest operating systems. This enables
broader protection and simplifies CA ARCserve Replication deployment, improving IT
productivity and reducing risk.
System state and registry synchronization. Helps you protect and more quickly restore the
Windows Operating System through periodic, scheduled snapshots stored on the replica
server
Data migration. Enables migration from physical to virtual servers or from x86 to x86-64
platforms using a native 64-bit replication solution.
Storage independence. Allows you to pick and choose your storage vendors and devices on
both production and replica environments, to help keep DR costs low.
No reboot install and maintenance. Help eliminate disruption to the production environment.
Linux and UNIX replication. Provides data protection for organizations with heterogeneous
®
environments.
Integrated backup with CA ARCserve Backup. Allows combined replication and backup
managed in a single console.
Adopting Server Virtualization for Business Continuity and Disaster Recovery 8
What is CA ARCserve High Availability?
® ™
CA ARCserve High Availability, formally known as CA XOsoft High Availability, provides system,
application and data high availability along with continuous data protection to minimize unplanned and
planned system downtime that affects your sales, service, employee productivity and reputation. It is a
solution that delivers reliable system and data protection for a wide range of operating environments
and includes the following critical features:
Server Group Management enables failover of groups of servers, or server farms altogether
when protecting applications like SharePoint, Dynamics CRM, Websites and others that are
typically deployed across multiple servers. Used when the failover servers are located at a
remote site to help maintain application performance.
Host-based Data Replication across the LAN or WAN between production servers and
replica servers for data protection, availability and migration.
Data Rewind for continuous data protection (CDP) complements any backup solution and
protects data between periodic backups for faster recovery time after accidental or malicious
loss or damage.
CA ARCserve Assured Recovery provides fully automated recovery testing to help ensure
recovery readiness, with no disruption to the production or DR environment. Optional VSS
snapshots may be performed after system validation for application-consistent backup.
Automated and Push-button Failover includes real-time server monitoring and push-button
failback after server repair or replacement.
Unified, web-based self-protecting management console makes CA ARCserve High
Availability a breeze to manage.
Extensive Microsoft application and OS support to protect a wide range of applications
including Microsoft Exchange, SQL Server, SharePoint, IIS, Dynamics CRM, Microsoft Hyper-
V Server, and most Windows operating systems. Oracle and Blackberry Enterprise Server are
also supported. Other applications may be protected by creating a simple script.
All of these features, combined with its affordable, flexible, and simple licensing structures, make CA
ARCserve High Availability a critical component of every virtualized server environment. Suitable for
companies of all sizes where business uptime is critical, it offers leading-edge high availability; data
protection and recovery technologies that efficiently and cost-effectively help reduce the costs of
protecting both virtual and physical servers at the data center and remote offices. Certified across an
extensive range of industry platforms and applications, it provides compatibility and support for
heterogeneous environments.
Continuous availability of your virtual server environment
CA ARCserve High Availability integrates easily with your Microsoft Hyper-V environment. Some key
features include:
Full System High Availability provides replication and failover of the Windows operating
system, system state, application and data on physical and virtual servers to a Hyper-V server.
Physical-to-physical (P2P), Physical-to-Virtual (P2V) and Virtual-to-Virtual (V2V)
replication, failover and failback. Provides the protection of data running on Hyper-V
Servers and also allows the use of Hyper-V Servers for replica or failover servers to help you
reduce overall DR costs.
Flexible Hyper-V deployment options. Offers flexibility of replicating and failing over at the
hypervisor level or at the individual guest operating system/application level without the
requirement of installing CA ARCserve High Availability on each of the guest operating
systems. This enables broader protection and simplifies CA ARCserve High Availability
deployment, improving IT productivity and reducing risk.
Continuous or Periodic Replication to help balance network resources with the desired level
of protection
Storage independence. Allows you to pick and choose your storage vendors and devices on
both production and replica environments, to help keep DR costs low.
No reboot install and maintenance. Help eliminate disruption to the production environment.
Adopting Server Virtualization for Business Continuity and Disaster Recovery 9
Linux and UNIX high availability. Provides protection for organizations with heterogeneous
®
environments.
Integrated backup with CA ARCserve Backup. Allows for combined replication, backup and
high availability through a single management console.
Why companies are adopting Hyper-V for their
virtualization infrastructure
The following are some of the benefits of Hyper-V:
Integrated Management Tools
Integrated with System Center, monitoring and planning tools help businesses to see where
they can use virtualization to enhance their infrastructure.
Integration with management and monitoring tools make it possible for administrators to
monitor the status of all physical and virtual machines so that they can diagnose and address
maintenance issues before they become major problems.
Hyper-V integrates with industry-standard server management tools through the DMTF
standards.
Improved System Management Efficiency
Administrators can use Hyper-V tools to change many aspects of the virtual machine
configuration without shutting down the virtual machine.
Hyper-V can be monitored and management solutions can immediately alert administrators if
the performance of a Web server is reduced because of increased server workload.
Hyper-V can be configured to manage other Hyper-V Servers and their hosted virtual
machines.
Reliable Disaster Recovery Support
Hyper-V supports server clustering to balance system workload over several physical or virtual
machines and to minimize the potential impact of server failure.
Adopting Server Virtualization for Business Continuity and Disaster Recovery 10
Live Migration makes it possible for administrators to automatically or manually move virtual
machines between physical servers without noticeable downtime.
CA Recovery Management products are integrated with Hyper-V to provide a complete backup
copy of each VM/hypervisor and an off-site replica for failover in the event of a local disaster.
Hyper-V reduces the costs associated with replica servers used to provide HA for business-
critical applications.
Environmentally Responsible Technology
Server consolidation means a smaller impact on the power grid because server consolidation
can save power, rack space, storage, and cabling. This helps organizations that are
implementing green initiatives to develop smaller and more cost-effective data centers.
Windows Server 2008 R2 Hyper-V supports newest processor innovations that allow servers
to run more efficiently supporting the trend to green IT.
Considerations in implementing virtualization
As with any deployment of new server technology, introducing virtualization into an enterprise
computing environment is a complex task. The goal in deploying virtualization is to replicate the current
end-user environment while reducing costs and maintaining performance. In this section, we will
review some of the system requirements to evaluate before you plan your migration to a virtual
machine environment.
Determining Application Requirements
Before deploying your virtualization solution, it is a good idea to create an inventory of the end-user
applications that your organization relies on. This inventory should include the following information for
each application:
All software & hardware dependencies, including device drivers and maximum memory usage
The software version
The division or department that owns the application
The level of support provided for the application by the manufacturer
The Microsoft Assessment and Planning Tool (MAP) provides agentless inventory functionality. It is
*
available free of charge on the Microsoft Web site . With MAP, you can generate reports that can help
you to determine which systems and workloads are good candidates for virtualization.
Determining the Architecture of Your Virtualized Server Environment
There are three general approaches to deploying your virtual machine environment. The first is to
deploy to all data centers, servers, and workstation and portable computing platforms at once. The
main benefit of this approach is that virtualization tools and administrative practices are applied
consistently throughout the organization. This provides standardized administrative support that
reduces the complexity of IT maintenance tasks and their associated costs. The risk is that it is a much
larger deployment that has a higher likelihood of causing business disruptions.
The second approach is to deploy to hubs. Hubs are groups of virtual machines, users, and domains in
a central physical location that connect to satellite locations. Hub-based deployment starts at the hub
and extends to the satellite locations. This design also benefits from the standardization of
administrative support, but hubs can also be used as a test bed in deploying virtual machine
*
http://www.microsoft.com/downloads/details.aspx?FamilyId=67240B76-3148-4E49-943D-
4D9EA7F77730&displaylang=en
Adopting Server Virtualization for Business Continuity and Disaster Recovery 11
technology to a small number of servers before the full deployment. The disadvantage of this approach
is that the deployment is not performed across the entire organization at one time. Several smaller
deployments increase the amount of system downtime at the hub, which increases the risk of business
disruption.
The third approach is to gradually deploy to the entire environment, one location at a time. This
deployment method can be performed in environments with or without hubs. This is the least-costly of
the three approaches and provides the opportunity to test the new environment before deploying it to
the rest of the organization. However, decentralized deployment methods such as this have the
highest risk of business disruption, are the most complex to administer, and limit the benefits of
virtualization to only those areas of the organization that have been deployed.
Determining Hardware Resource Requirements
Virtualization technology aggregates processing loads of several workloads on server-computer
hardware and hence requires state-of-the-art server performance. Consider the size and design of
your organization’s virtual machine environment when making decisions about how much to invest in
newer, high-performance hardware resources. Newer servers typically have a set of new processor
technologies that will drive efficiency in respect to scalability and power consumption.
The following is the server hardware to which it is important to pay particular attention in this
assessment:
Processor architecture and configuration
Memory
Storage
High-performance hardware, such as storage area network (SAN) hardware
Network cards and infrastructure
Corporate security issues will often determine whether a virtual machine environment will be deployed
on a physical server. Take this into consideration as well when determining the hardware resources
that you will use in your virtual machine deployment.
Processor Resources
Each server virtualization product has its own requirements for processor resources. It is a good idea
to plan your investment in processor number and speed to match the requirements of the virtual
workloads running on the physical server. Add a buffer amount to maximize the capacity of your
system—tools such as MAP can help you estimate this. The risk in failing to do this accurately is that
one or more of the virtual machines that you host may exceed the processing capacity of the physical
server and degrade the performance of all the virtual machines.
Memory Resources
Most virtualization technologies make it possible for you to control the amount of memory that an
application can use. To determine the optimal amount of memory, test your application on the virtual
machine before deployment. Put greater than normal demands on the application. Here again, the
MAP tool can help you determine appropriate test workloads.
Adding a memory buffer to a virtual machine is not recommended because it will limit the number of
virtual machines you can host on a physical server. Your application inventory will be useful in
determining the memory requirements of your virtual machine environment.
Storage Resources
In determining the storage needs of your virtual machine environment, you will need to consider the
capacity and performance capability of your storage system and whether it will be direct-attached or
Adopting Server Virtualization for Business Continuity and Disaster Recovery 12
shared storage. Direct-attached storage is one or more disk storage devices that are uniquely
accessible by each host server. Shared storage is storage that is accessible by more than one host
server. Shared storage is usually implemented by a SAN, which is a networked group of disk storage
devices that the server system accesses by Fiber Channel (FC) or the Internet small computer system
interface (iSCSI) protocol, which makes high-speed SCSI storage access possible over conventional
network hardware. Direct-attached storage is generally cheaper and SAN storage is faster.
Determine which of these storage methods to use in your deployment based on your performance
requirements, cost considerations, your backup requirements, your virtual machine architecture, and
the type of data that you will be storing.
High-Performance Hardware Requirements
Support for high-performance hardware such as specialized storage devices like optical jukeboxes and
tape libraries may be limited or non-existent in certain virtualization environments. For example, Hyper-
V does not currently support host bus adapter (HBA) hardware mapped to a virtual machine that is not
the parent partition. Verify that the manufacturer of your virtualization technology supports your high-
performance hardware and test the interoperability of your hardware with your virtualization server
environment before deployment. Note that you can get better handling when specialized storage
devices are accessed by the virtualization host rather than by the virtual machine.
Determining Backup, Replication, and High Availability (HA) Requirements
The most common strategies that are used in organizations to protect data today fall into three
categories: backup, replication and high availability.
Backups, typically used as a core component of any disaster recovery solution, can be performed in a
number of ways. The most common method today involves using dedicated backup software like CA
ARCserve Backup and CA ARCserve D2D to perform periodic backup to disk or tape and archiving for
long-term storage or regulatory compliance. CA ARCserve Backup can be scheduled and implemented
using a combination of full, differential or incremental strategies as well as leveraging Microsoft VSS
Snapshot technology to ensure data integrity. Alternatively or as a complement to CA ARCserve
Backup, CA ARCserve D2D uses Infinite Incremental snapshots that perform an initial full backup and
subsequent incremental backups, eliminating the need for another full backup. Customers that already
have an installed backup system can use CA ARCserve D2D to improve their recovery time
exponentially.
Replication is used to migrate or copy data from one server/VM to another or from one site to another
for disaster recovery purposes. Replication software like CA ARCserve Replication may be used to
perform replication directly from each primary server or VM to other physical or virtual servers like
Hyper-V. Alternatively replication may be used after backup to quickly and easily copy backups to a
remote location, avoiding the cost and risks associated with physical media transport. And of course
replication software may be used to speed physical to virtual server migration too.
Today, businesses and organizations can’t afford downtime so many are moving to a high availability
solution. High availability software like CA ARCserve High Availability provides system, application and
data availability along with continuous data protection. The failover servers may reside onsite or offsite
for complete disaster recovery readiness.
It’s always a good practice to ensure that the backup images or replicated data is indeed restorable or
recoverable. This is not only useful for compliance purposes but also gives peace of mind that in the
event of a problem, you can restore, recover or failover to the data reliably and quickly. CA ARCserve
Backup, CA ARCserve Replication and CA ARCserve High Availability all include some form of
automated recovery testing, helping you conduct periodic Disaster Recovery testing without any
manual intervention.
Adopting Server Virtualization for Business Continuity and Disaster Recovery 13
You should consider a vendor like CA whose solutions can protect your data across the entire
spectrum of recovery points all the way from backing up data every day, using hourly replication, or
continuous replication, or implementing high availability of business-critical data.
There are three common data backup methods: backups that use native applications, guest-level
backups, and hypervisor-level backups.
Backups That Use Native Applications
It is possible to use custom scripts to back up applications like Microsoft SQL Server, Exchange
Server, or SharePoint Server. But the complexity of managing and maintaining the backup images on
disk or tape using custom-built scripts is costly, inefficient, and often prohibitive. More importantly,
debugging the details of why backups on certain applications failed and tracking all the different
recovery points for different data could easily become complicated and difficult to manage.
CA ARCserve Backup and CA ARCserve D2D provide application-aware backup and restore of
Microsoft applications. With CA ARCserve Backup, using a policy-based approach, a copy of the disk-
to-tape backup provides simple-to-use, point-in-time restores of applications and automatically restores
the full backup image, the differential backup image, and the transaction log backup images. The
administrative dashboard provides a single view of the historical trends of your application server
backup, for troubleshooting and storage resource management. With CA ARCserve D2D, backup to
disk couldn’t be faster or easier. After the first full backup, only subsequent changes are captured at
the block level and going forward, only incremental backups are required significantly shortening the
backup window. Overall, for backup and restoration of applications like SQL Server, Exchange Server,
SharePoint Server, and file servers, it is often more efficient and cost-effective to use a dedicated
backup solution with an integrated application agent like CA ARCserve Backup and CA ARCserve
D2D.
Guest-Level Backups
In this method, you install a third-party backup agent, such as CA ARCserve Backup and CA
ARCserve D2D, on each virtual machine that will be protected. You then use the backup agent to
establish a backup schedule for each server and identify and select the volumes, files and folders, and
databases, etc. that will be backed up. Other than the ability to provide centralized scheduling and
administration, there is no interaction between the backup processes of each virtual machine—each
virtual machine is treated like a physical server for backup purposes.
The advantage of this approach is that you can do granular file\folder restores. This is because you are
treating a virtual machine just like any other physical machine and doing the backups of files and
folders from within the VM. The disadvantage of this approach is that it would be slow to recover a full
VM. This specific problem can be addressed by doing hypervisor-level host-based backups.
Guest-level Backup using CA ARCserve Backup
Adopting Server Virtualization for Business Continuity and Disaster Recovery 14
Hypervisor-Level Backups
In this method, you deploy a third-party backup agent such as CA ARCserve Backup agent for virtual
machine on the Hyper-V host machine. You can then configure the backup server to do an online
backup of these VMs by taking a VSS snapshot of the VMs, thus ensuring the integrity and
consistency of the data inside each of the VMs.
Doing hypervisor-based backup has an advantage that the entire VM can be backed up. This helps in
the quick recovery of the entire virtual machine in case of a disaster or the need to clone a VM.
Moreover, using CA ARCserve Backup’s unique patent-pending capabilities, you can also achieve
granular restores of files and folders from the full image-level VM backup. What this means is that you
can do both types of restores from just one image-level backup of the VMs from the Hyper-V host, thus
increasing the operational efficiency of managing virtual machine backups and restores.
Hypervisor-level backup with CA ARCserve Backup
Implementing backups of Hyper-V environments using CA ARCserve
Refer to the section titled ―
Backup‖ later in this document for use scenarios corresponding to these backup methods.
Determining Replication Requirements
As we discussed above, replication refers to migrating or copying data between servers or locations for
disaster recovery. It complements any backup solution by providing data protection between periodic
backups. It is used to quickly and easily migrate or copy backups from the data center to a remote DR
location – even if it has been deduplicated. Replication is also used to aggregate data from remote
offices to a central location where centralized backup may be performed by trained IT staff in a more
secure and less costly manner. A replication solution should also include an automated, non-disruptive
testing capability to help ensure the data is available when needed.
CA ARCserve Replication provides LAN and WAN replication for onsite and offsite data protection and
migration and includes a data rewind feature for continuous data protection (CDP) as well as
automated, non-disruptive testing through CA ARCserve Assured Recovery which is included. For
Hyper-V servers, CA ARCserve Replication may be used to replicate the entire server at the hypervisor
level or replicate individual Guests/VMs depending on your specific needs.
Data Replication. In simple terms, data replication technology such as that provided by CA
ARCserve Replication, keeps a primary/production server synchronized with a replica/failover
server by copying the data and all subsequent block-level changes from one to the other, in
either an ongoing real time or scheduled manner. The replica/failover server can be located on
the LAN, but is typically housed at a remote location for disaster recovery purposes; the replica
server also maintains the same operating systems and applications as the production server.
Replication may be set to occur continuously, periodically or be performed manually at any
time to help balance network resources with the desired level of protection.
Adopting Server Virtualization for Business Continuity and Disaster Recovery 15
If data is lost or damaged on the primary server, an administrator may use the data rewind
feature on the replica server to quickly and easily recover data or even a complete database,
even while replication continues. Then, the data may be restored or copied back to the primary
server.
For large data sets or databases, CA ARCserve Replication provides Offline Synchronization
for Windows-based environments to shorten time-to-protection. You take a VSS Snapshot of
the primary server to removable media and then restore it to the replica server—instead of
synchronizing across the LAN or WAN which could take longer and use network resources.
Continuous Data Protection (CDP). CDP technology captures every change made to files in
a journal or log file and allows administrators to use a wizard-like GUI to quickly and easily
rewind to a known good point-in-time before a data corruption event. CDP provides full data
protection for periods of time between your backup jobs. CA ARCserve Replication includes a
Data Rewind feature for built-in CDP that is performed on the replica server to ensure that
there is no impact to the production server or users. CDP offers organizations a powerful data
protection solution to complement any backup solution. Microsoft provides a Hyper-V VSS
writer that can be used to protect virtual machines in a consistent state. If the virtual
machines are running Windows operating systems that support VSS and if Hyper-V Integration
Components are installed in the guest operating system, the Hyper-V VSS writer is capable of
communicating with the guest VSS framework to put guest applications in a consistent state
for backup. CA ARCserve Replication uses the VSS framework to detect the snapshot point
and inserts a bookmark to indicate the point where the data is consistent.
Determining High-Availability Requirements
High availability typically refers to maximizing uptime and availability of systems, applications and data.
It is used to overcome unplanned outages and data loss and should also be an integral part of any
disaster recovery strategy. Many organizations also use this solution to avoid business downtime
during server maintenance, upgrades and migration from physical to virtual servers. In most computing
environments, a two-tiered approach to assessing and deploying high-availability solutions is
necessary. The first tier includes ensuring high availability within the local physical location such as a
data center, headquarters or remote offices. But power failures, fires, floods and other man-made and
natural disasters at that location can result in a complete loss of the systems, applications and data
that is hosted at that site. To minimize these risks and add a higher level of disaster recovery
readiness, you should deploy an additional level of failover protection—using a secondary data center,
remote office or remote DR site, even if it is a third-party hosting facility—for replication, failover and
offsite data protection.
This section assesses the difference between the high-availability requirements within data centers
and the requirements between data centers.
High Availability within a Data Center
There are three basic approaches to implementing this tier of high availability. Each approach is
appropriate for a specific type of server or application, so you will likely use a combination of these
methods in your solution.
Network Load Balancing (NLB). This approach involves distributing the network traffic that is
being sent to an application across a group of copies of that application that are running on
different servers. If one server fails, the traffic that is sent to the application instance that is
running on that server is redirected to the copies that are running on the other servers. The
network traffic load is redistributed across the remaining application copies and servers by
using a load-balancing algorithm. This is particularly useful for stateless application instances.
Application-Specific Clustering. A server cluster is server software that is installed on
separate physical servers and is configured to operate as a single server. The server cluster
Adopting Server Virtualization for Business Continuity and Disaster Recovery 16
shares server configuration data, access to storage devices, and load-balancing operations.
When one of the physical or software servers fails, the remaining servers in the cluster
reassign to the operational servers the tasks that are performed by the defunct server. Some
mission-critical enterprise server applications—for example, Microsoft SQL Server and
Microsoft Exchange Server—can operate as a server cluster. These server applications are
said to be ―cluster aware.‖
Host Clustering. Virtual machine configurations that do not have cluster-aware software can
use host clustering to implement high availability. In this approach, several physical servers
that host virtual machine operate as a cluster. When one of the physical servers fails, the
operations that are performed by all of the virtual machines that are hosted on that server
computer are moved to the other physical servers in the cluster. The level of system availability
that host clustering provides is not as high as with application-specific clustering, because the
operating systems or applications deployed on the hosted virtual machines are not cluster
aware. With Hyper-V, a configuration with host clustering provides for planned and unplanned
downtime. During unplanned downtime, the virtual machines will be restarted on the target
node within the cluster, and during planned downtime the virtual machines will be transferred
statefully from one cluster node (physical server) to another node or alternatively the virtual
machines can be moved via Live migration to another server in the cluster by which downtime
of the VM is avoided all together.
High Availability Across Data Centers
Server virtualization technology makes it possible to run a cluster of virtual machines on one physical
server. This is a way to implement a high degree of system usage for relatively little cost. An optimal
high-availability solution is where the data and operations of a cluster of physical servers at the primary
site are replicated in an identically configured cluster of standby virtual machines at a remote backup
site. When the primary site becomes unavailable, data center operations are quickly transferred over a
wide area network (WAN) from the primary site to a standby site in a process called failover.
Network load-balancing and clustering technologies protect only the data that a server cluster
manages, so software that manages data protection between clusters is necessary. Also, not all
applications are cluster-aware and not all operating systems support clustering. CA ARCserve
Replication and CA ARCserve High Availability provide replication and high availability without the
need for clustering architectures which simplifies deployment, management and maintenance.
To accommodate this critical business continuity need, companies of all sizes may leverage two
powerful technologies. Data replication with automated and push-button failover provides onsite and
offsite data protection and continuous application availability (HA). Continuous replication combined
with a Data Rewind feature provides continuous data protection, often referred to as CDP, enables
organizations to protect data between their scheduled backups and instantaneously rewind their data
back to a known good point-in-time before a data corruption event from accidental or malicious
damage.
Replication and Failover. As we discussed in the Replication section above, in simple terms,
data replication technology such as that provided by CA ARCserve High Availability software,
keeps a primary/production server synchronized with a replica/failover server by copying the
data and all subsequent block-level changes from one to the other, in either an ongoing real
time or scheduled manner. The replica/failover server can be located on the LAN, but is
typically housed at a remote location for disaster recovery purposes, The replica server also
maintains the same operating systems and applications as the production server.
When using CA ARCserve High Availability with physical or virtual servers like Hyper-V,
administrators may use the Full System High Availability feature to replicate the operating
system, system state, application and data to an offline Hyper-V guest/ virtual machine. This
not only shortens time-to-protection, but may also reduce operating system and application
Adopting Server Virtualization for Business Continuity and Disaster Recovery 17
licensing costs since the replica server operating system and applications are considered
offline until failover occurs,
The system may be configured for automated failover and end-user redirection at first sign of
trouble, or, it may be configured for push-button failover where the administrator may research
the failure and then determine whether or not to failover and redirect users. The process is
transparent to users who typically continue to work as if nothing has happened. A key
component of any replication and high availability solution is an automated recovery testing
capability, especially one that may be scheduled for periodic testing and one that will not
impact the production environment or halt the replication process. Both CA ARCserve
Replication and CA ARCserve High Availability solutions include CA ARCserve Assured
Recovery that provides this must-have capability.
Continuous Data Protection (CDP). CDP technology captures every change made to files in
a journal or log file and allows administrators to use a wizard-like GUI to quickly and easily
rewind to a known good point-in-time before a data corruption event. CDP provides full data
protection for periods of time between your backup jobs. CA ARCserve High Availability
includes a Data Rewind feature for built-in CDP that is performed on the replica server to
ensure that there is no impact to the production server or users. CDP offers organizations a
powerful data protection solution to complement any backup solution. Microsoft provides a
Hyper-V VSS writer that can be used to protect virtual machines in a consistent state. If the
virtual machines are running Windows operating systems that support VSS and if Hyper-V
Integration Components are installed in the guest operating system, the Hyper-V VSS writer is
capable of communicating with the guest VSS framework to put guest applications in a
consistent state for backup. CA ARCserve High Availability uses the VSS framework to detect
the snapshot point and inserts a bookmark to indicate the point where the data is consistent.
Implementing backups of Hyper-V environments using CA
ARCserve Backup
Now let’s go through multiple scenarios on how you could back up the Hyper-V infrastructure using CA
ARCserve Backup.
Full Weekly Backup with Daily Differential Backups
A small organization is running Microsoft SQL Server, Exchange Server, and SharePoint Server on a
Hyper-V virtualized server. They need to perform a backup of each application as quickly and easily as
possible because IT staff is limited in number and experience.
By deploying the CA ARCserve Backup Agents for Microsoft SQL Server, Exchange Server, and
SharePoint Server in each of the VMs, application consistent backup can be quickly and easily
performed. The administrator can use the CA ARCserve Backup server console to submit Grandfather-
Father-Son (GFS) backup jobs, selecting each of the VMs for backup. The administrator may select
from the several backup options available such as multiplexing all of the data from multiple sources to
a single tape drive, multistreaming while backing up to tapes, or encrypting the data while backing up
to tape. Multiplexing and multistreaming help reduce the backup window, and encryption allows you to
ensure all data is encrypted on tape. For disk-to-disk-to-tape (D2D2T) backup, the administrator first
backs up to disk using data deduplication, and then specifies the policy to copy the data to tape. Since
data deduplication decreases disk space utilization, data can be retained on disk for longer periods
without adversely impacting storage resources.
The CA ARCserve Backup dashboard keeps track of the status of backups in progress, tracks
recovery points, and monitors disk and volume fragmentation. Data can be restored from either disk or
from tape. CA ARCserve Backup provides point-in-time restores and automatically chooses the
Adopting Server Virtualization for Business Continuity and Disaster Recovery 18
necessary full backup, differential backup, or transaction logs for restores.
If using encryption, CA ARCserve Backup remembers the passwords required to restore the data for
you.
Image-Level VM Restores and Granular Folder\File-Level Restores
In this scenario, an organization’s backup strategy requires the administrator to perform complete VM
restore for some VMs for Disaster Recovery, while being able to perform a granular file\folder restore
on other VMs.
To accomplish this goal, the administrator installs the CA ARCserve Backup agent for the VM on the
Microsoft Hyper-V Server. The agent will discover the VMs, publish them into the CA ARCserve
Backup server, and perform the backups. By using the agent deployment tool on the CA ARCserve
Backup server, the administrator can deploy the CA ARCserve Backup agent for VM on each of the
VMs of the Hyper-V Server. The agent is installed from the central location simultaneously to multiple
VMs. Please note that this agent is NOT used to move data—it is used to get a small piece of
metadata during backup, which is used later for granular restores. This agent can also be used when
you are performing granular files\folders restores.
Now the administrator can select the appropriate VMs and submit the backup job—all from a central
CA ARCserve Backup console. This ensures full Image VM restores and granular file\folder restores.
The CA ARCserve Backup Dashboard helps you by presenting statistics on backups, encryption,
recovery points, trending analysis of backups, and many more SRM style reports on disk, volume,
CPU, NIC, memory, and many other key data components.
Multiple Remote Office Backup
A midsize organization has multiple remote offices, each with different backup requirements. Each
office has multiple physical machines, some of which are under consideration for migration to
virtual machines to reduce costs. There is a mix of both physical and virtual machines at some of
the offices. The organization’s backup/restore requirements are typical of organizations of this size.
To reduce backup costs and support from employees at each office, the decision has been made
Adopting Server Virtualization for Business Continuity and Disaster Recovery 19
to perform backup to disk and eliminate all tape-based backup at each remote office in addition, for
disaster recovery purposes, a copy of all remote office data must be stored at the central data
center. Each remote office will use CA ARCserve D2D to protect all the VMs and physical servers
using backup to disk. Since CA ARCserve D2D uses infinite incremental backups to dick, backup
windows and storage requirements are significantly shortened.
CA ARCserve Replication software is then used to replicate the backup data on disk at the remote
offices to the central data center. In case data is lost at a remote office, CA ARCserve D2D will be
used for recovery. In case the local backup disk is damaged or the entire remote office location is
lost, the replicated data at the central data center can be restored to VMs and shipped to a new
remote office location.
Alternatively, you could replicate all the data at each remote office using CA ARCserve Replication
to a central site like the data center or DR site and then perform backup and deduplication off the
replica server. This helps overcome backup window constraints and eliminates the risk of having
non-IT staff perform backups out in the remote offices. The central data center not only works as a
ROBO DR site but by moving data to tapes it also provides long-term retention of data for
compliance.
Adopting Server Virtualization for Business Continuity and Disaster Recovery 20
Physical to Virtual Server Migration
An organization would like to quickly and easily perform physical to virtual server migration. CA
ARCserve Backup allows you to perform Bare Metal Recovery of a physical server onto a virtual
machine using the CA ARCserve Backup Disaster Recovery Option. Bare Metal Recovery (BMR) of a
physical machine to a virtual machine also enables far faster recovery (within an hour) of an entire
physical machine into a virtual machine providing minimal end-user downtime.
CA ARCserve High Availability may also be used to quickly and easily migrate from physical to Hyper-
V virtual servers by using the Full System High Availability feature. This feature replicates the entire
server including the operating system, system state, application and data to a Hyper-V virtual machine.
Then use the push-button failover feature to stand-up the new VM and automatically redirect users –
with NO DISRUPTION TO THE WORK ENVIRONMENT!
Implementing backups of Hyper-V environments using CA
ARCserve D2D
CA ARCserve D2D provides both host-level and virtual machine (VM) level protection for Microsoft
Hyper-V environments. The following scenarios describe several situations and options to protect the
Hyper-V servers using CA ARCserve D2D.
Protecting the Hyper-V host server and the Virtual Machines which are hosted on that server:
An organization would like to completely protect both the host server for Microsoft Hyper-V and the
individual virtual machines running on the host server. To accomplish this goal, the administrator would
install CA ARCserve D2D on the Hyper-V host server and run a backup. Once the Hyper-v host server
has been backed up, the entire machine is now completely protected. It is just that simple!
When a restore is needed, CA ARCserve D2D can restore the entire server or the individual virtual
machines to meet varying recovery requirements. The first option is to restore the VM to same location
- which can be easily accomplished by browsing recovery points or by selecting the specific files for the
VM in the CA ARCserve D2D restore user interface. When the restore job has finished, all the
administrator needs to do is to open the Hyper-V Manager and start the VM. Alternatively, the VM can
be restored to a different location on the same Hyper-V Host server or another Hyper-v host server
simply by selecting the "Restore to alternate location" option and selecting a different destination path.
Finally, the entire Microsoft Hyper-v host server can be restored to the same or different physical
server by using the Bare Metal Recovery option in CA ARCserve D2D.
Protecting just the Hyper-V Virtual Machines:
An organization would like to protect just the individual virtual machines running on the host server.
This might be a case where an individual department wants to add more frequent backups to a critical
server that has been virtualized. To accomplish this goal, an administrator could install CA ARCserve
D2D on the Hyper-V host server and select the volume where the VM files are located. Alternatively,
CA ARCserve D2D could be installed inside each Windows Virtual Machine. This option works as if the
VM was a physical server providing increased protection just for that VM and full restore of the entire
VM or granular restore of individual VM files.
Adopting Server Virtualization for Business Continuity and Disaster Recovery 21
Establishing a high-availability environment by using
Hyper-V and CA ARCserve High Availability
Before virtualization technology became generally affordable, physical server clustering was the
standard way to implement high availability. Because hardware server technology is usually
proprietary, these high-availability solutions were costly to implement, maintain, and scale. Deploying a
high-availability solution that uses virtualization technology involves converting many of the physical
servers to virtual machines. Newer and faster physical servers are often deployed to host the virtual
machines.
This process of server consolidation provides the same level of availability as physical server clusters,
with the following advantages:
Fewer physical servers mean reduced power costs and requirements for cooling systems and
floor space.
Virtual machines can be easily moved between physical servers without reconfiguration or
shutting them down. This process is known as quick migration.
After a system failure, virtual machines can be quickly restored from backups and restarted.
Administrators can restore manually or configure maintenance tools to automate the process.
The hardware configuration of the physical servers can be modified without changing the
configuration of the hosted virtual machines.
Processing loads can be quickly reallocated between virtual machines to respond to changing
use demands.
The following are two use-case scenarios where Hyper-V and CA ARCserve High Availability are used
to administer high-availability clusters in high-availability contexts. The scenarios illustrate how high-
availability clusters can be constructed from different configurations of physical and virtual machines.
Some of the advantages and disadvantages of each scenario are included.
Basic High-Availability Scenario: Two Virtual Machines on a Single Physical Machine
A small, budget-conscious, startup company can afford only one physical machine but requires
application high availability for SQL Server 2008 when run in a Microsoft Cluster Services (MSCS)
environment. The physical machine runs Windows Server 2008 and Hyper-V and hosts two virtual
machines running a clustered SQL Server 2008 and Failover Cluster. The virtual machines share a
single SQL Server 2008 database. The virtual machines are managed by Failover Cluster as a cluster
so that should the first virtual machine (active node) fail, the second virtual machine (passive node)
gains control of the shared resources and keeps the SQL Server available for applications and users.
While this scenario fulfills the minimum requirement for a high-availability cluster, it does not protect
against physical hardware failures or a site outage (or a software data corruption). The physical
machine is a single point of failure because when it fails, the entire high-availability configuration also
fails. Other advantages that clusters normally offer, such as planned physical server maintenance
outages, are not possible. This configuration is most useful for quickly testing different high-availability
cluster configurations using a minimum of hardware. The more complete solution would be to use CA
ARCserve High Availability to enable automatic failover to a remote server providing continuous
application availability.
Large-Scale High-Availability Scenario: Physical-to-Virtual Environment with Advanced
Data Recovery
In this scenario, the company has eliminated the physical server as a single point of failure by
purchasing four more physical servers and two SANs. All of the physical servers run Windows Server
2008, Failover Cluster, and CA ARCserve High Availability. The first two physical servers host SQL
Server 2008, the second two servers host Exchange Server 2007, and the fifth server runs Hyper-V.
Adopting Server Virtualization for Business Continuity and Disaster Recovery 22
The first four share access to the first SAN, where all of the server application data is stored. The four
physical servers at the main site are configured as primary servers, and the fifth is the standby server.
The standby server resides in a second corporate office and hosts four virtual machines—two running
SQL Server 2008 and the others running Exchange Server 2007. Each virtual machine runs Failover
Cluster and CA ARCserve High Availability. The second SAN is also deployed in this office as a
backup storage device, and all of the virtual machines share access to it. All of the physical servers
and SANs are connected by high-speed Ethernet over a WAN.
Here high-availability is maintained within the cluster as well as established across the WAN between
the physical servers and the virtual machines. If either the SQL Server or Exchange Server fails on a
primary server, the Failover Cluster switches server operations over to the SQL Server or Exchange
Server running on the other primary server. In the event that the entire cluster of primary servers fails,
CA ARCserve High Availability switches server operations to the corresponding virtual machine on the
standby server.
In addition, the company’s IT administrators use CA ARCserve Assured Recovery to periodically test
the disaster recovery capabilities of the applications on the standby server, to provide scheduled off-
host backups, and to implement continuous data protection for the data.
This approach provides several advantages to the basic high-availability scenario:
A high level of failover protection is provided for the two server applications that are running at
the main site.
Data redundancy between the two SANs is supported.
By using virtualization, the company does not need to purchase four more physical servers at
the second site as backups for the primary servers at the main site. This saves hardware and
maintenance costs.
This solution is more scalable. If an additional physical server or another server application is
deployed at the main site, another virtual machine can be hosted as a backup on the fifth
server.
Integrated CA ARCserve High Availability tools result in reduced system maintenance costs.
Assured recovery and continuous data protection help ensure the recoverability of your data.
Conclusion
Server virtualization reduces the cost of implementing recovery management solutions and improves
optimization of computing resources. However, if the deployment is not planned properly, much of the
efficiencies that are gained from server virtualization can be lost.
For this reason, as part of the plan, companies should consider reliable protection, recovery and
availability solutions that work well in both physical and virtual environments and that provide robust
system management tools. CA ARCserve Backup, CA ARCserve D2D, CA ARCserve Replication and
CA ARCserve High Availability products offer world-class technology for protecting file and application
servers that are running in physical and virtualized environments such as Hyper-V.
Adopting Server Virtualization for Business Continuity and Disaster Recovery 23