Which Backup is Best?
I was recently tasked with figuring out which backup process would work best for an in-place upgrade to an application server with a local SQL instance. The client was looking for a restore process that could be implemented and completed quickly to bring that application server back to a pre-upgrade state in case something went south during the upgrade process. The application server needed to be upgraded several application versions to the latest release and then the SQL version needed to be upgraded. The client was running Microsoft Azure Recovery Services (MARS) files\folder level backups, using the Azure Backup Agent on this server because it worked nicely with the SQL chain of backups they had in place. SQL was scheduled for a full back up once at the beginning of the month and differential backups throughout the rest of the month. This setup worked fine for the client’s daily needs, however, the question for me was “how involved would a restore be if we needed to do one quickly during the weekend and have everything back to the way it was prior to the upgrade by Monday morning?”
Not having worked a lot with Azure backups, I enlisted the help of Microsoft tech support to talk through my scenario. I had to get up to speed on the latest and greatest Azure backup solutions and fast!
Although the client’s MARS backups were great for restoring files and folders fairly quickly, it was not an optimal solution for the kind of restore I would need in my current situation. Had I needed to restore the machine quickly, I would have needed to recreate a VM, install the Azure backup agent, restore the system state and then restore each drive in succession. This would take way too long for the VM to get back to a pre-upgrade state. I just didn’t have that kind of time.
Let’s take a closer look at what the benefits and limitations are for each and what is protected:
MARS backup
| Benefits | Limits | What is protected? |
|
|
|
IaaS-VM backup
| Benefits | Limits | What is protected? |
|
|
|
In my case, an Infrastructure as a Service Virtual Machine (IaaS-VM) portal backup would be my best option. I can go into Azure and go directly to the virtual machine to trigger this backup.
With this backup, all the managed disks attached to the virtual machine will be backed up and in the case of restoring this backup, Azure simply replaces the original disks attached to the virtual machine with the backed up disks. All the configuration, data, and programs that were present at the time of backup will all be there and available! My server would then be back to a pre-upgrade state in a matter of hours depending on how large the disks are, or how many disks I have to restore. This is the perfect solution for my scenario.
IaaS-VM backups allow you to do nightly backups, customize your retention policies, backups can be stored up to 10 years, and it allows for file or folder level restores by mounting the disks on any compatible computer.
Database backups are an essential part of any business continuity and a disaster recovery strategy because they protect your data from accidental corruption or deletion. As such we at Anexinet would love to have a conversation with you around your backup and recovery strategies, ensure you are doing everything you can to fully protect your organization, please don’t hesitate to contact us. Additionally, you can find other Azure posts and podcasts from Anexinet here. As well as a Cloud Strategy Kickstart that we offer, it is the perfect way to begin your Azure journey.
Share on 
