Blog

How To Test Your Server Backups And Data Protection Plans

Transcript:

If you aren’t testing your backups, your data isn’t protected. It’s as simple as that.

When it comes to backup and recovery, recovery is the only part that matters.

At Storagepipe, we’re frequently approached by companies after a major critical data loss incident. And the most common pattern we see is that – although these companies seemed to have adequate data protection in place – they never bothered to verify if these backups could actually be recovered when needed.

Even if your backup process was perfectly implemented on day 1, it may need to be revised as time goes on.

  • New servers may be added without notifying the backup administrator. This is especially true of virtual servers.
  • There may be new software or operating system upgrades which are incompatible with the current backup system.
  • Storage growth may slow down emergency recovery, or lead to unnecessarily high storage costs.

That’s why you need to test on a regular basis. At Storagepipe, we recommend testing your data recovery process at least once per year.

But how exactly do you test your backups?

The best way to test backups is by simulating real-world data loss scenarios, and recovering as if those scenarios were actually happening in real life. These disaster recovery drills should also be scheduled at random dates and times, so that you’re forced to constantly be ready for an unexpected drill.

In this video, we’ll walk you through a data loss scenario for a fictional pharmaceutical company called Acme.

Roger is the IT manager for Acme, where he’s in charge of over 200 laptops & desktops, and dozens of servers. Within their datacenter, they have Windows File Servers, Exchange Servers, SQL Servers with SharePoint, Linux Servers, iSeries servers and a number of other machines. This datacenter is also laid out as a mixture of virtual and physical machines, further complicating the data protection process.

  • At 3:00, Roger gets a call from his boss. The company is suing a former employee who shared trade secrets with a competitor, and they need to retrieve a series of emails dating back 5 years. This means that Roger will need to call the tape storage company and have them truck over about 150 archival tapes.
  • While he’s waiting for these old tapes to get trucked over, an employee walks into the server room says that they’d unknowingly saved over an important document. The mistake happened 2 weeks ago, and they didn’t notice until just now. Roger has a point-in-time copy of this file, but it’s stored at the off-site storage facility. So he calls up the tape storage company and asks them to courier over that backup tape. Unfortunately, this version is a few days older than the last recorded changes, so the employee will need to spend another 5 hours re-creating the old file.
  • Unbeknownst to Roger, this employee had rested their water bottle on top of a server and forgotten it there. While turning around, Roger accidentally knocked over the water bottle, taking out the SharePoint server for the entire company. Now there are 100 employees that can’t do their jobs. This wasn’t his fault, but he’s the one who’s going to get blamed.  This server needs to be brought back online post-haste. His job depends on it. Although he’s protected the SharePoint server with a solution that allows bare-metal recovery, he needs to order a new physical server with identical hardware.
  • That’s when he gets a call from a one of the sales reps, saying that his laptop was stolen at the airport, and he’s got an extremely important client meeting in 4 hours. The sales rep need all of this hard drive data back, and sent to China before the meeting starts. Roger already knows this person well, because he’s not very smart with computers and makes frequent calls in to tech support.
  • Out of curiosity, Roger loads his web browser to look up the sales rep’s location on Google maps. But as soon as the browser opens, he sees that the Acme’s busy web site has been hacked. Their e-commerce database has been flooded with erroneous data, and customer accounts have been deleted. The company has lost about 4 hours worth of customer transactions, and can no longer generate revenue due to the site outage.

Now let’s think about the implications of this scenario in more detail.

  • When the archival tapes finally arrive from the storage facility, how long would it take Roger to sift through 150 old backup tapes – one by one – and find a few critical emails dating back several years?
  • What would happen if those emails contained attachments which could only be opened using obsolete software that was no longer available?
  • What if – instead of just destroying a single server – the spilled water started an electrical fire that took out the whole datacenter?
  • When it comes to critical systems, what kind of Recovery Point Objectives and Recovery Time Objectives need to be met? In other words, how much downtime and data loss can you tolerate?
  • How can you maintain accurate laptop backups in a company that has over 200 travelling laptop users?
  • What if Roger had been away on vacation when all of this happened? Would he be able to guide a non-technical employee through the recovery process over the phone?

As you can see, testing your backups is a bit more involved than just recovering yesterday’s files or performing a full recovery to another location. You need a comprehensive set of tools that will cover you in a broad range of scenarios, while also ensuring you stay within budget.

You need to individually analyze every internal system and ask yourself:

  • How much of this data do I need to retain?
  • How fast do I need to recover?
  • How much data can I afford to lose?
  • When restoring very old data, will I run into compatibility problems?
  • Am I just protecting data, or do I also need to protect hardware?
  • If I were to leave the company, would anyone else know how to recover this data?
  • If I can’t bring systems back online immediately, what it the Plan B during the downtime?
  • How much will the recovery process cost?

The best advice we can give you is to keep your backups simple. This will make testing and – more importantly – recovery much easier. And it will allow you to adapt more easily to growth and change as time goes on.

Rather than creating a mish-mash of different backup solutions, try to find a single provider who can back up your Email Servers, Databases, Flat Files and Laptops, while also providing ediscovery, archiving, mirroring and bare-metal restore capability.

If you’d like more information on protecting your datacenter, please visit http://storagepipe.com. We’d love to answer any server backup testing questions you may have.

Leave a Reply