Backup and Recovery Best Practice
IT’S ALL ABOUT FAST, EASY AND RELIABLE RECOVERY
Imagine an emergency situation where your business is confronted with a serious problem that may ultimately result in permanent loss of data and/or unexpected downtime.
The way to prevent these situations from occurring is to develop a plan and ensure it is functional.
When considering the severity of outages, it’s helpful to compare essential services into a list of priorities.
Establish priorities for services like: Internet access, Email, CRM, Files & Printers, Shipping, Inventory etc.
Ask yourself: How much total downtime is acceptable? How much data is an acceptable loss? How long can we operate without said service(s)?
The outcome will direct the following:
- Recovery Point Objective (How frequently do we need to back up?)
- Recovery Time Objective (How long can it be offline?)
- Total Cost of Ownership (Can we afford it?)
24hrs -> 8Hrs -> 4Hrs -> 1Hr -> 15mins -> Continuous
| Downtime & Data Loss Costs (RPO): | High <—————————> Low |
| Backup Job Duration (RTO): | Slow <—————————> Fast |
| Recovery Time (RTO): | Slow <—————————> Fast |
| Complexity & Costs (TCO): | Low <—————————> High |
“As Easy as 3-2-1 and 1-2-3”
~infamous backup Sherpa
Always keep 3 Copies of your Data, 2 stored locally, 1 stored remotely. The 3-2-1 Rule also applies to Recovery Points in Time; 1 Year, 2 Months, 3 Weeks.
Different Backups for Different Data Server Images, Databases, Applications and File Data all have different needs to consider for a fast, easy and reliable recovery.
Databases and Enterprise Apps may require special licensing and separate backup jobs.
Compliance & Archiving
Regulatory compliance for things like PCI-DSS or PIPEDA may mandate further regulations over and above our 3-2-1/1-2-3 rules above.
The ability to restore applications and systems at points in time for 3rd party compliance or independent auditing requires extra consideration typically in the form of additional archive copies.
Data Growth, Expansions & Acquisitions: More files, emails, and apps are being used than ever before and data itself gets bigger as file sizes increase year over year. A rule of thumb is to factor 15-24% year over year data growth and inflation.
“Generally, every 5 years business data grows 2.5-3.5x its original size”
Expanding business and acquisitions will contribute to explosive data growth, which can completely undo existing data backup & recovery plans. Be sure to consider any possible explosive growth needs when planning for data growth.
Overheads & Incidentals: One final consideration is to plan for data overheads. Things like 20% free on system and data volumes, incompressible data, GiB to GB conversions or incidental changes can inflate any well-designed backup & recovery plan. Adding a 20-30% overhead is a great start to factor into your plan.
Disaster Recovery: Not just data protection, its ensuring uptime overall Every recovery plan needs to include remediation for network service failures. Redundant Internet Connections, Routers, Switches, and Spare Devices all contribute to ensuring the most possible uptime.
System Monitoring & Alerting Notifications: The worst way to find out something failed is for your clients to make you aware. Ensure reliable monitoring & alerting is configured for all your essential services, including backup job successes as well as failures.
Practice Restoring & Disaster Recovery: Trusting successful backup notices without periodically testing restores will inevitably lead to failure as all experienced IT technicians can attest to. “A good backup is only as good as its last successful restore” Be sure to test restores of files, systems and archives periodically to ensure your plan will recover successfully.
How to size a backup solution
Start by calculating the total capacity currently used. This involves reviewing each server or system and its used capacities. In this example we’ll use 1TB or 1074GiB.
First, I need to consider data growth. With historical data, one could compare factual data growth from historical backup jobs and their growth year over year. In this example, I’ll use an average of 19.5% year over year growth estimate (1.625% monthly).
Second, I need to consider growth for my target refresh cycle 1, 3, 5 years (TCO). (Don’t forget to consider compounding data at rest for the end of the intended period)
Based on a 5 Year plan, with a planned replacement or review I can expect the following growth:
YR 1: 1283.43GB YR 4: 2190.16GB YR 2: 1533.70GB YR 5: 2617.24GB YR 3: 1832.77GB Total: Est 2.44x capacity growth over 5 years for data inflation.
Third, I need to factor full and incremental copies (Based on 3 Copies, 2 locally, 1 Remote and 1 annual full, 2 monthly full copies and 3 weekly incremental copies)
At the end of the 1st year of growth, my 2nd local copy will be around 1283.43GB.
Estimating my size requirements: 1 Annual Full: 1283.43GB 2 Monthly Full: 2566.86GB 3 Weeks Incremental: 13.86GB (1283.43-1074=209.43GB annual growth. 288GB / 365=574MB/Day. 789MB*21= 12.05GB for 3 weeks growth, add 15% overhead=13.86GiB)
Total copies: 3864.15GB Adding Overhead of 22% = 4714.26GB
*Advanced backup software will leverage technologies like deduplication and compression to reduce this capacity, however it can vary wildly based on file types and number of copies. For this example, we’ll use 45% reduction in capacity. 4714.26GB-45%=2592.84GB
At the end of year 1, using our projected inflation and backup retention policies, we end with 2592.84GB or 2.4x our initial dataset.
At the end of year 5, we end with 4334GB or 4x our initial dataset.
Summary Finalizing this example starting with 1TB of data: The primary server should be as big as the TCO year’s data inflation, in this case min. 2.6TB. The backup server should be min. 4.7TB (to hold all the backup data pre-deduplication). The offsite/cloud/archive size should be min. 4.3TB.
“It’s as easy as 3-2-1 and 1-2-3” “A good backup is only as good as its last successful restore”