Virtualization Strategies for the CIO
When implemented correctly, high-availability servers and systems add much-needed redundancy and data protection to a virtual environment.
The importance of high-availability technologies has
Requires Free Membership to View
increased with the popularity of server
virtualization. As organizations virtualize more workloads and increase
consolidation ratios, high-availability servers and systems protect virtual machines (VMs)
against extended periods of downtime.
There are a variety of ways to architect high availability into a virtual infrastructure, including failover clustering, load balancing and fault-tolerance technologies. When evaluating these options, IT pros must also account for their existing hardware, software and budgetary constraints.
The following resources will show you how high-availability servers and systems function in a virtual infrastructure, how to choose the best options for your data center as well as helpful hints and best practices for deploying efficient high-availability systems.
TABLE OF CONTENTS
IMPLEMENTING HA | THE
DIFFICULTIES OF HA | HA TOOLS AND STRATEGIES
IMPLEMENTING VIRTUAL HIGH-AVAILABILITY SYSTEMS
An efficient data center needs a disaster-recovery plan and high-availability servers and systems tailored to its specific needs. If your data center relies on virtualization, you’ll have to carefully architect the hardware and software to ensure the availability of VMs during a failover. As such, you’ll need a strong understanding of the requirements, downtime allowances for workloads and the costs associated with high-availability servers and systems.
Virtualization high availability checklist
Virtualization high availability can greatly improve the efficiency and reliability of your data
center. Without proper planning, however, high-availability servers and systems can cause serious
headaches. Review this checklist to ensure your VMs, servers, mirroring software, failover
technologies and backup regimen are conducive to a successful virtualization
high availability deployment.
Achieving high availability in a virtualized environment
Unfortunately, there is no one-size-fits-all option for achieving virtualization high availability.
Some high-availability technologies can even outstrip the capabilities of virtualization
platforms. You can, however, still increase data center availability by learning how to use
high-availability
methods to customize deployments.
Data center planning considerations for high availability
Virtualization improves data center flexibility, but carelessly implementing high-availability
servers and systems can lead to unexpected downtime and resource shortages. As such, proper
data center planning is the key to a successful HA deployment. You need to consider a variety
of factors, including the best method for protecting legacy applications, existing data center
hardware and resources as well as which high-availability method will meet your needs and budgetary
constraints.
Ensuring high availability in the data center
Many of today’s virtual data centers use high-availability technologies to hedge against downtime.
But using high availability in your virtual environment does not always provide the best possible
protection for workloads, as it’s still possible for Web servers and databases to go down.
Fortunately, there are still ways to decrease downtime with OS-level
high availability.
THE DIFFICULTIES OF HIGH AVAILABILITY SERVERS AND SYSTEMS
The benefits of high availability are obvious, but they may obscure some of the difficulties that may arise with the technology. Careless or inexperienced virtualization architects can easily make mistakes that will not be apparent until a failover occurs. But understanding the difficulties and pain points may help you avoid potential problems.
Virtualization high-availability pitfalls
High-availability software is easy to install, and many users understand the value it brings to a
data center. But it doesn’t provide flawless protection for virtual machines and services. So it
would be wise for IT pros to understand the flaws
and limitations associated with high availability, so they won’t be caught off guard if a
problem arises.
Five situations where virtualization high availability is overkill
Many data centers adopt high-availability technologies to complement disaster-recovery plans. But
high-availability servers and systems are expensive and may be more of a luxury than a necessity.
Here are five examples where virtualization
high availability is overkill.
Server cluster high-availability gotchas: DNS and failover
High availability is still one of the lesser-understood components of clustering. Many
organizations forge ahead with high-availability servers and systems without properly preparing for
the potential problems. Understanding how how
DNS and the failover order affects high availability, for example, will help ensure the success
of your server cluster.
VIRTUALIZATION HIGH-AVAILABILITY TOOLS AND STRATEGIES
Once your data center decides to apply virtualization high availability, the challenge then becomes how to implement the best possible method. Sorting through the dizzying array of high-availability tools and software can be a daunting task. The following tips describe the available options and offer advice for assembling high-availability servers and systems in a virtual environment.
High-availability architecture: Redundancy vs. abstraction
Virtualization has caused many organizations to rethink high-availability
architecture and strategies. Unless an IT shop has a virtualization-specific strategy for high
availability, even the failure of a single server can be catastrophic. It’s important to implement
resiliency in even the smallest of virtual data centers through redundant hardware or
abstraction.
High-availability clustering in virtual data centers
In practice, traditional high-availability
clustering will often work equally well for virtualized servers, but there are some subtle
differences. With virtualization, you have more flexibility in the choice of servers, and each VM
is protected through a periodic or continuous VM snapshot process. To reap the benefits of
virtualized high-availability clustering, you must understand the differences and how to perform
routine maintenance.
Reserving cluster resources to boost high availability
Every infrastructure supporting high availability needs a cluster reserve, which is a predetermined
set of resources set aside for a failover. And a lack of resources can compromise a
high-availability cluster during a failover. It may be difficult for smaller data centers to
justify setting aside an entire server for high-availability failovers. Fortunately, today’s
hypervisors allow you to reserve any percentage of cluster
resources as spare capacity.
Data center high-availability technology: A holistic approach
High-availability
technology can foster a false sense of security. Smart IT shops will have high-availability
plans that extend beyond virtualized servers to encompass networks and storage. To prevent
undesirable results, like downtime from the loss of a single, physical infrastructure component, be
sure to regularly test your high-availability servers and systems.
Testing high-availability software options from vendors
High-availability
software can vary significantly. Before making an expensive purchase, carefully explore and
test these tools, because they will ultimately affect how you manage your high-availability
architecture. If you haven’t deployed a virtualization platform, you may want to compare
hypervisors to determine which best fits the high-availability architecture you already have in
place.
High-availability solutions in nonvirtualized environments
Many businesses face astronomical costs for each hour of of lost network access or user
productivity. To avoid incurring such costs, IT pros managing both nonvirtualized and virtualized
environments turn to high-availability
strategies. If your organization plans to implement virtualization high availability in
the data center, be sure to understand the basics of high availability in nonvirtualized
environments first.
How a virtual server cluster brings high availability to test and dev
Ensuring
high availability in test and development environments can prevent the loss of revenue and
reduce risks when moving your environment to production. A virtual server cluster increases high
availability by allowing easier VM movement and management. Nonetheless, not every test and
development environment is a good candidate for a virtual server cluster.
High-availability technologies and trends
To manage an efficient data center, you must pay attention to the latest developments in
virtualization platforms and disaster recovery strategies, as well as high-availability
technologies and trends. Scott Gorcester, founder and president of Moose Logic, offers helpful
hints in this expert Q&A.
This was first published in April 2012
Join the conversationComment
Share
Comments
Results
Contribute to the conversation