WavebreakmediaMicro - Fotolia
Imagine you're driving down the road and, suddenly, you hear an ominous sound coming from the rear of your car: thumpa-ta, thumpa-ta, thumpa-ta. As the car becomes increasingly difficult to handle, you begin to understand what has happened: You have a flat tire.
No problem. Just find a safe spot, and pull the emergency spare out of the trunk. Uh-oh, the spare is flat, too.
A similar crisis faces an untold number of storage administrators every day. Due to an oversight, error or a prime storage media failure, a need suddenly arises to access a particular set of files stored on backup media. But the backup data is missing, outdated or defective. Like an unlucky driver, the storage administrator now faces a predicament that could have been easily avoided with some advance planning, in the form of testing backups.
Here's what you need to do.
1. Understand the seriousness of regular backup testing. Just as it's important to test a spare tire to ensure it will work when it's needed most, you need to be testing backups, said Girish Dadge, product management director for Sungard Availability Services. "Testing your backups also gives you a chance to assure yourself that your backup policies and schedules work properly," he added.
2. Create a documented backup testing plan. Familiarity with a documented test plan ensures that employees have both the skills and experience necessary to successfully perform data recovery and provides confidence to the organization, observed Eamonn Fitzmaurice, worldwide data protection lead at IT services firm HPE Pointnext.
3. Make testing backups a routine. To assure the validity and integrity of any backup, it's essential to carry out regular restoration tests. "It is not unusual to find organizations that have systems that are inadvertently not being protected via a backup schedule," Fitzmaurice explained. Routine and comprehensive backup testing is a strategy that can highlight anomalies so that corrective action can be taken.
4. Take a holistic approach. Organizations need to understand their data layout and why they are doing backups. They then need to develop a test backup plan to meet their desired objectives.
Every organization has different backup objectives. "For example, the banking industry needs backups for compliance, audit and legal," Dadge said. "Healthcare organizations have personal data, so they need to focus on security, retention and legal requirements." All restore and recovery testing should include data, application and system state testing, Dadge recommended.
5. Test frequently according to regular schedules. Ideally, a test should be conducted after every backup completes to ensure data can be successfully secured and recovered. However, this often isn't practical due to a lack of available resources or time constraints. "Each organization should, at a minimum, commit to a regular schedule of weekly and/or monthly restores of systems, applications and individual files with checks to ensure the data is valid and accessible as intended," stated Marty Puranik, CEO of Atlantic.Net, a cloud hosting provider. "This will also provide your organization with a realistic time frame for recovery when disaster strikes."
Not all data is created equal, a fact that should impact frequency of testing backups. "Some data is more important than others," noted Atif Malik, a director in KPMG's CIO Advisory unit. For instance, Sarbanes-Oxley Act compliance and regulator data might be considered more important than marketing data. "Controls should be in place to mitigate risks based on the importance of that data," Malik advised.
6. Take full advantage of automation. Automation should play a key role in any backup testing strategy. "Organizations should strive to automate as much of their backup testing as possible to ensure consistency and data validity and to reduce the burden on staff tasked with testing backups," Puranik suggested. "Test restoring full systems to virtual machines, applications, databases and individual files," he added.
7. Ensure that the backup test covers all bases. If the backup test doesn't actually test the entire workload being restored, it can't be considered a real test. "Many organizations will simply restore one or two files from archive and consider that a success," noted Chris Wahl, chief technologist at cloud data management provider Rubrik. "This workflow has no relationship to the reality of restoring complex applications and should be avoided when considering a real backup test."
8. Make testing backups an integral part of internal app development and deployment. Backup testing should be in the front of everyone's mind when developing and introducing new applications into the organization. "The most successful enterprise data management strategies involve knowing how and when to perform backup validation tests before allowing data to move into a production workload," Wahl explained.
9. Ensure backup accuracy. When data is recovered, storage administrators and database administrators can perform an initial "sanity check" on the data. "However, the end users of the business applications are often best positioned to highlight if the data restored is accurate and consistent," Fitzmaurice observed.
10. Have redundant backups. Never back up to only one tape or set of tapes. "If you do use tapes, replace them on a regular basis," recommended Brian Engert, senior application developer at customer software developer Soliant Consulting.