Schrödinger’s Backup states, “The condition of any backup is unknown until a restore is attempted.” The backup software said the job was successful – but was it really ? Was the job back up intact, restorable files, or was there an issue with the end user’s laptop that corrupted something two hours beforehand and the actual backup result was just a second copy of useless junk data ? The only way to tell for sure is to restore the data & utilize it and unfortunately this is often deemed to be too much work or too impactful to perform on a regular basis.
Opening the Restore Box
Testing backups on a scheduled basis is critical to the success of any business. Best Practice wisdom is that regular restore tests must be run to make sure the backup software is doing what it is supposed to. More critical server restores may only get tested every 6-12 months due to the impact but the nature of those servers ought to in turn demand a more rigorous restore test; do not restore just one or two files, restore the entire server, see if it boots up, and fire up your business-critical Application off the new copy. Extra-critical environments in turn should be restored from offsite copies at an entirely different Data Center. Does the server still run successfully? If not, why not ? Was the restore completed within the documented Restore Time Objective (RTO) to the correct Restore Point Objective (RPO) ? What can be changed going forward to ensure success ? These are the questions that should be answered by the business after every regular restore test.
One of the primary tenants of Science is that success has to be repeatable to be trusted – restores are the same way. Successful backups are frequently enforced by vendor SLAs but restores generally do not get the same treatment. Restores must be completed not just once but every single time the business asks to recover data; if the organization cannot expect a 100% success rate on restoration of critical data, then a deep-dive should be performed into why that is not currently possible and fixes put in place to cover future restores.
Show Your Work
While a successful restore is great something almost as important is knowing how to get there. If the primary backup administrator was to be unavailable, could anyone else complete the work ? If the entire team of administrators was unavailable due to a local disaster, would someone else at the company be able to complete the restore successfully ? With a completely documented restore process even when the regular administrators are unavailable the organization can move forward with critical restores.
The Cost of Doing Business
The ability to restore data has both Capital and Operational costs associated with it that Organizations strive to keep in check. While these restore costs are definitely non-trivial they are still cheaper than some of the alternatives – it might only take a few hours to redo a PowerPoint presentation for a Customer but what if the employee who did it originally is on leave now ? The main Oracle database being corrupted could halt all transactions at a Bank’s branches, or a water leak on the second floor could damage hardware for all the Application Servers in a particular first floor rack.
All Organizations need to plan for restores to happen, whether single file, single Server, or entire VM farms, and part of this planning is deploying enough gear to support the restores in a manner consistent with the needs of the Business. While a single tape drive for a DR site might be able to handle occasional Excel file restores it’s going to struggle to handle the demands an Oracle RAC cluster DR restore would place on it. Regardless of whether it is located at the primary site or a Recovery location the hardware must be capable of handling the demands that might be placed on it; purchasing the bare minimum will only harm the Business in the long run.
Resolving the Restore Paradox
The experts at VAST can help you perform regular restore tests of your critical data and know how to assist with sizing your recovery environment appropriately to the Organization’s needs. Whether you are looking for an initial implementation or just want a review of your existing procedures, VAST can help protect the data critical to your Business. There should never be any doubt about the restorability of your data or the ability of the Company’s personnel to perform the work.