A Contributed Article by: Hervé Tardy, Vice President and General Manager, Distributed Power
Quality - Americas, Eaton Corporation
Keeping critical workloads operational
is a data center's top priority during a power outage. However, it's the
inconvenient truth that servers must sometimes be shut down in order to prevent
data corruption. Traditional data centers have long relied on a combination of
UPSs and power protection software to shut down servers during utility outages,
but server virtualization adds a degree of complication to defending data
during power failures. Live migration capabilities can move VMs dynamically
from one host server to another under a variety of conditions, but few such
products offer built-in functionality for migrating virtual workloads
automatically during outages, when the function is most often needed.
The good news is, technology continues
to evolve to solve such issues, and newly developed power management solutions
have the ability to significantly reduce the risk of data loss. By
automatically "showing up" to shut down servers in an orderly fashion when
disaster strikes, the latest solutions are essentially a knight in shining
armor protecting a virtualization environment's data.
First, let's take a closer look at the
challenge... when a virtualized data center loses power, technicians must shut
down not only their physical servers but also the virtual machines running on
those servers. They must also execute the many steps of that process in a
specific sequence, often in the face of stress and pressure. Virtual machines
must be shut down before physical ones, and core devices (such as domain controllers
and shared storage arrays), after the servers that depend on them.
There are three innovative
ways to overcome these challenges:
1. Download open source management code
Websites such as Network UPS Tools (www.networkupstools.org) have developed pre-written,
open source operating system codes that can shut down servers gracefully and in
the correct sequence during power failures. By downloading, installing and
customizing such code, data center managers can shut down servers in the proper
order when utility/server power becomes unavailable. This solution offers an
easy yet extremely powerful and highly configurable power management option to
organizations that use and customize Linux or other open source solutions, a
category that includes most operators of cloud computing data centers.
2. Deploy advanced power protection and management software
While power protection applications enable organizations to shut
down servers in an orderly manner during utility failures, most systems only
support physical devices. The latest and most sophisticated power protection
solutions, however, support virtual machines in addition to hosts such as
VMware ESX, Microsoft Hyper-V, Citrix XenServer and Red Hat KVM. This
technology, when added to a virtualization management system, can shut down
both physical and virtual servers in pre-defined sequences to minimize vulnerability
to data loss.
automatically moves virtual machines from host servers impacted by an
electrical outage to unaffected servers elsewhere within the server cluster to
preserve data integrity. Furthermore, when integrated with VMware's Site
Recovery Manager, Intelligent Power Manager software will initiate planned
migration moving server workloads from the protected site to a recovery site
hosted in a cloud environment.
When a virtualized
data center loses power, and there is no recovery site to move the workloads
to, Eaton's Intelligent Power Manager software will provide a graceful shutdown
of any or all servers and virtual machines running on them to minimize exposure
to data loss. Intelligent Power Manager's tight integration with Citrix and
VMware allows for graceful server shutdown, remotely,
without the need for any shutdown agent to be installed on either a host server
or virtual machine.
3. Add automated scripts to advanced power protection software
Many advanced power protection solutions enable users to create
scripts that automatically respond to specific alarms in a predefined manner.
Companies can use such scripts to augment their power protection system's built-in
functionality. For example, technicians could extend UPS battery runtime by
creating a script that automatically shuts down virtual machines running
non-critical workloads early in a power outage and then consolidates the
remaining virtual machines onto a smaller number of host servers.
The rate of innovation is amazing. Within
the span of a few years, server virtualization has progressed from a promising
new technology to data center mainstay. Although there have been challenges
along the way, virtualization-supporting software developments have armed IT and
facilities managers with potent new tools to maintain uptime and preserve
critical data. By harnessing the power of the latest power management solutions,
users can now take full advantage of server virtualization's rewards while
mitigating its risks, preserving the integrity of two very important things;
their business and their data.
For more information about Eaton's solutions
for virtualized environments, visit www.eaton.com/virtualization. To learn more about
Eaton's power quality products, visit www.powerquality.eaton.com.
About the Author
Hervé Tardy joined Eaton in
November 2007 as Vice President and General Manager of their Distributed Power
Solutions business unit, with responsibility over single-phase UPSs, software
and connectivity products. His responsibility has recently been expanded to
include the management of marketing and sales initiatives through the fast
growing IT channel in the Americas. Tardy is based in Raleigh, North Carolina.