|
Crashes happen. So do natural disasters and computer viruses. Hardware and software sometimes fail as well. People make mistakes. Things get stolen. Systems go down.
System failures, regardless of their cause, can be an organization’s worst nightmare, particularly since 85 percent of systems are not backed up. What’s more, between 17 and 40 percent of tape restores actually fail, according to recent analyst reports.
That’s not such a big deal — unless the system is running mission-critical applications or houses timely, vital data without which the business cannot succeed and move forward. Of course, most systems are integral to the business, and their uptime is essential for ensuring business continuity.
Faced with a hardware failure, organizations traditionally have had little choice but to repair or buy new hardware and then reinstall everything from scratch, the old-fashioned way. That includes installing the operating system, applications, low and high-level device drivers, service packs, patches, updates, hot fixes, and configuration changes — then applying journal changes and testing before going live. This can take days or even weeks.
If reinstalling everything were only a one-time process, it wouldn’t be so bad. If duplicate, unused systems were standing by just waiting to be used (and they likely are not since they are far too expensive and typically include slightly different components), it wouldn’t be so bad.
But all too often, bare metal recoveries are needed more than once. And only dissimilar hardware is available.
Meanwhile, every second counts.
Clearly, organizations must find a better way — one that is rapid, simple, and reliable and overcomes the limitations of traditional, manual bare metal recovery to dissimilar hardware platforms.
Getting Down to Bits
The key to meeting the dissimilar hardware restore challenge is to break through the hardware abstraction layer (HAL) and overcome the differences in mass storage controllers. Doing so not only allows organizations to rapidly recover entire systems to dissimilar hardware platforms but also to virtual environments.
These capabilities move bare metal restore beyond the limitations of traditional automated system recovery for Windows systems. With each operating system designed for a specific hardware device, its configuration is slightly different from the next. And automated Windows system recovery cannot account for the dissimilar hardware components of a new computer system.
But it is the Windows HAL, kernel, and storage controllers that further complicate matters. These three elements must be correctly assigned to hardware in order for a Windows system to boot.
In contrast, system recovery tools that can create file images at the sector level, a layer below the operating system, are able to avoid any file-level locks associated with automated Windows system recovery processes.
Best of all, time-consuming layered restorations become unnecessary. With hot image hardware-independent restores, any hardware incompatibilities that might have impeded a manual restoration are no longer an obstacle.
How It Works
Hot image hardware-independent restores enable organizations to take base and incremental recovery points — entire system images — on a scheduled basis without having to bring the system down. Base, or full, system recovery points typically are run during non-production hours or during periods of lower system resource utilization. Incremental recovery points are typically scheduled to run during production times. Because they’re scheduled, no manual intervention is required. Instead, administrators simply set it and forget it.
Once clients are installed on servers, desktops, and laptops and recovery points are established, administrators then store these recovery points to media. A growing number of hardware-independent restore solutions support an extensive variety of media, from NAS and SAN devices to SCSI, SATA, and ATA as well as flash memory, CD, and DVD, and USB and Firewire devices.
The most advanced hardware-independent restore tools streamline the process even more by obviating the need to restore multiple incremental recovery points to return a system to a particular point in time. These tools provide all the storage, HAL, and kernel drivers that Windows operating systems use when performing a new installation. They also offer a large number of drivers that are not part of the standard Windows installation media but are popular third-party drivers. What’s more, if these tools do not have a needed driver as they attempt to do a restore, they will stop the restore process and prompt the administrator, giving him time to inject the driver and avoid the dreaded blue screen that might otherwise result.
With the necessary drivers loaded, administrators then select the desired recovery point and the destination. The recovery then restores the entire system to unallocated space on the selected device, regardless of hardware platform. Storage, HAL, kernel, and other critical drivers for the newly restored system are automatically updated. From there, the administrator boots for the first time on the new hardware, Windows plug-and-play runs, and the administrator can log in and check for non-critical components that plug-and-play did not detect. Minutes later, the new system is up and running.
But these tools are not limited to restoring only to physical environments. They also work in virtual environments. With an increasing number of businesses leveraging virtualization to centralize and consolidate servers and to facilitate disaster recovery, organizations must be able to also provide rapid conversion of recovery point files to a virtual system. This requires support for conversion of physical systems to virtual systems as well as virtual systems to physical systems.
Hardware repurposing is also a cost-effective solution for organizations. After all, virtually every environment includes servers that are underutilized as well as others that are overutilized. To make better use of existing resources, organizations rely on hardware repurposing.
Hardware-independent restore tools can ensure that the installation time for migrating from one platform to the next is as rapid as possible. While manual reconfiguration of a server is a multi-step process that involves as many as 90 steps and several hours, these newer hardware-independent restore tools accomplish the same — but with just four steps and a few minutes. Moreover, the steps do not have to be journaled and can be replicated the same every time with no special training.
Perhaps one of the more vital benefits of hardware-independent restore tools is the ability to meet stringent recovery time objectives (RTOs) that do not require immediate failover. For computers that cannot justify the high cost of clustering or mirror sites but must be recoverable quickly, this offers a much-needed recovery solution. With shorter recovery time objectives becoming the norm, organizations must increase the disaster tolerance of the system. Many of these systems are considered of medium criticality — that is, systems and replica sites must be available in minutes or hours, hardware does not need to be similar, and virtual systems can be used.
Today’s advanced hot image hardware-independent restore tools enable administrators to achieve medium criticality objectives while keeping costs low by factoring in the manpower that would have been required for manual installation.
These new tools are changing the way organizations perform a wide range of IT tasks. Bare metal system recovery, restoration to virtual environments, hardware migration, and more are made flexible, reliable and simple. Organizations can even use these tools to meet ambitious recovery time objectives and service level agreements.
Fast, reliable Windows system recovery, reduced administrative costs, and maximization of current IT investments all point to hot image hardware-independent restore as a tool whose time has truly come.
About the Author:
Brian Wistisen is a senior product manager, Data and Security Management Group, Symantec Corporation.
Go Back
|
