Next Article +

Avoid Disaster in Disaster Recovery

PlateSpin Forge and PlateSpin Protect Span the DR Gap for Both Windows and Linux, Including SUSE Linux Enterprise from Novell

Written by Jason Dea

When you hear "disaster recovery", if you picture expensive like-for-like infrastructures replicating your most essential servers, perhaps it's time to rethink disaster recovery (DR). If you envision cumbersome backup servers running all night followed by tape cartridges traveling to a storage vault, again, it's time to reconsider your DR strategy.

If you are a Linux administrator and see yourself orchestrating your DR solution with scores of commands, it's time you find a more efficient way to manage DR in today's environment of shrinking resources and growing IT needs.

In each of these scenarios a gap exists between the cost and the effectiveness of disaster recovery. You get either cumbersome and affordable or expensive and effective—the choices administrators have been facing for years.

Novell designed PlateSpin Forge and PlateSpin Protect from the ground up to span this gap, and Windows administrators have been using them for some time. Novell has recently released the latest updates to these products to now protect environments running Linux and mixed Linux-and-Windows environments.

If you see yourself orchestrating your DR solution with scores of commands, it's time you find a more efficient way to manage DR.

Spanning the Gap

A high-availability, or DR-by-duplication, strategy focuses on protecting workloads using local clusters, duplicate host sites or other like-for-like physical infrastructures. These strategies provide very low recovery time objective (RTO) and recovery point objective (RPO). These are the time it takes to move from a failure back to full production and the amount of data you lose in the process of recovery, respectively. However, even though RTO and RPO can be very low, these solutions can be very costly and extremely complex to configure and manage.

Traditional backup, on the other hand, eliminates the need for duplicate systems by storing backup data on tape or imaging devices, but acquiring new hardware, configuring it to the state of the original and restoring the data from the backup archive can take a long time, reducing effectiveness and increasing administrator aggravation.

The PlateSpin Forge and PlateSpin Protect products from Novell span the gap between cost and effectiveness with an entirely different way to think about DR.

The concept is quite simple, and the solutions are really equally easy to administer. PlateSpin Forge and PlateSpin Protect create a virtual machine (VM) copy of protected physical or virtual workloads and update the virtual machines at intervals you specify from 60 minutes to weekly or longer. When a server fails, you receive a notification and switch the users to this backup VM, almost immediately. The VM resumes the application load using its most recent update. And once the failed server is operational, you can either restore the VM from scratch to the server or do a reverse replication of just the changes since the failure occurred. The reverse replication would be appropriate when the original server can be repaired. If the original server configuration cannot be obtained, PlateSpin Forge and PlateSpin Protect have the flexibility to restore back to any x86 server from any vendor.

PlateSpin Forge versus PlateSpin Protect

Before we go further, you should understand that PlateSpin Forge and PlateSpin Protect are like two versions of the same product. They provide the same capabilities, use the same administration console and are managed the same way.

PlateSpin Forge is an all-in-one appliance that includes the storage, replication software, remote management interface and hypervisor. You just plug the appliance into your network, configure it, and it begins to protect up to 25 workloads. Novell designed PlateSpin Forge for medium-size enterprises and branch and field offices of larger enterprises.

PlateSpin Protect offers more flexibility in larger, more complex networks. It includes just the software—the replication software and management interface. You install PlateSpin Protect on your own virtual infrastructure connected to your own storage system. You can purchase as many workload licenses as you want, giving you maximum flexibility.

With both products, you manage the entire DR system, regardless of the number of licenses and DR appliances, through a single graphical interface. You can also integrate these products into your existing DR infrastructure with an SDK that lets you configure a custom solution.

How You Manage Disaster Recovery

Now that you have a basic idea of what PlateSpin Forge and PlateSpin Protect are and how they might fit into your DR plans, you're probably most interested in what you have to do as an administrator. Let's take a look at how you monitor workloads, configure workloads, execute a failover, perform a failback, and conduct testing. We'll also look briefly at some of the reporting capabilities.

The PlateSpin Forge and PlateSpin Protect products from Novell span the gap between cost and effectiveness with an entirely different way to think about DR.

Monitoring Workloads

The PlateSpin Forge and PlateSpin Protect Web-based console begins with the Dashboard. (See Figure 1.) Other than the name in the top left corner, the management consoles are nearly identical for both products.

The Dashboard gives you an overall view of the workloads that are being protected: how many workloads are protected, whether any have failed and how many workloads are not protected.

Simple icons in the Workloads Summary show you the status of each workload. A green light means the workload is running normally. An orange light means there is an error in the configuration; perhaps you forgot to configure some part of the protection. A red light means the workload has failed.

The License Summary shows how many licenses are being used.

The Storage pie chart shows the status of the available storage. PlateSpin Forge includes storage in the appliance, but you can also attach external, SAN or NAS storage.

On the right side of the screen, you can see a log of the past, current and upcoming events.

Simple icons in the Workloads Summary show you the status of each workload.

Configuring Workloads

To configure a workload, you just need to tell PlateSpin Forge or PlateSpin Protect where the target workload is on the network by providing a host name or IP address. (See Figure 2.) You also provide the administrator credentials so the product can interact with the workload as needed. Then you identify the schedule for incremental replication based on how often the data changes and how much network bandwidth you want to use for replication.

One nice feature Novell added to these products is Smart Replication, which slices large batch replications into small portions that can be processed quickly without overloading the network.

A number of other configuration settings include whether you replicate by file or data block, how often you want to replicate, how often you want the devices to ping the workloads to ensure they are still functioning, whether to encrypt transfer data, the volumes that are protected and so on.

These are all menu driven and easy to follow. You can also set these configurations individually for each workload, or you can configure sets of workloads using templates.

Managing DR by Workload

Novell designed these DR products to take the guesswork out of managing them. For example, in the Workloads tab, where you will spend the bulk of your day-to-day management, you can see that the console activates only the commands that are applicable to each workload. (See Figure 3.) If you check on a workload that has failed, the Prepare for Failover and Run Failover commands become active.

This window also provides the other information you need about each workload, such as whether it's online, its protection tier, its replication status and so on.

For a description of each command, see Workload Commands in the Additional Resources.

When a Workload Fails

If a workload or a server fails, you have two options: conduct an immediate failover, wherein the VM will boot up and begin serving the applicable users from the last snapshot, or start with the Prepare for Failover command. If there is a chance the failover is a false alarm, say someone simply tripped over a network cable, you should start with Prepare for Failover. With this command, the DR appliance stages the failover by configuring the virtual machine to take over for the failed workload but leaves it in a paused state. You can then choose whether to complete the failover by selecting the Failover or Cancel Failover commands.

Novell designed these DR products to take the guesswork out of managing them.

Restoring the Production Server

This is where PlateSpin Forge and PlateSpin Protect really shine. Traditionally, once you've repaired or replaced a failed server, reinstalling the OS, applications and data can be a nightmare. However, with these products, all these steps are a one-click process. They can build the new workload from scratch, or in the case of a repaired machine, you can chose to restore just the changes that have occurred since the workload failed. This includes restoring plug-and-play drivers that have been updated or that are different because you're using new equipment. PlateSpin Forge or PlateSpin Protect automatically install any necessary up-to-date drivers for Windows and Linux.

Testing Your DR Environment

If your DR system failed, now that would be a disaster. Testing is an important part of every DR strategy, and in the past this often required building an expensive and cumbersome parallel network just for testing. These DR products from Novell provide a Test Failover command that takes a snapshot of the VM and boots the snapshot to a virtual environment that you have defined as the test network. You don't boot the backup VM itself, and you don't affect the production network, so you can run this test anytime that is convenient for you.

Tracking DR Performance

PlateSpin Forge and PlateSpin Protect keep a history of all DR activities and events and measures performance such as how long it takes to run a job. They provide this data in several reports that you can view online, save as PDFs or export to applications such as spreadsheets. (See Figure 4.)

An Easier, Affordable DR Solution

When one of your workloads fails, you no longer need to go through a cumbersome several-step process from retrieving a flat-file archive, sourcing a new server or repairing the existing server, redeploying the operating system and other software, and rebuilding the data files just to get your services back up and running. And you don't need to employ an expensive duplicate system. With PlateSpin Forge or PlateSpin Protect, you have a backup of the workload ready to move into place almost immediately, and it restores back to your production servers just as smoothly, with very little disruption to the users on your network.

To learn more visit PlateSpin Protect and PlateSpin Forge.

Subscribe // Free monthly magazine

Or: Twitter | Facebook

Next Article +

Novell Connection Magazine
Special Editions

© 2011 Novell, Inc. All Rights Reserved.