- Pip install Boto3 - Thu, Mar 24 2022
- Install Boto3 (AWS SDK for Python) in Visual Studio Code (VS Code) on Windows - Wed, Feb 23 2022
- Automatically mount an NVMe EBS volume in an EC2 Linux instance using fstab - Mon, Feb 21 2022
Data backups are one of the most important tasks for any IT department. In my view, they are even more important than security. It can get nasty if a virus knocks down your network for a few hours. However, if you lose essential data, it could knock down your whole company forever. If you need some more motivation to read on, check out these statistics according to which 60% of companies that have lost their data will shut down within 6 months.
The problem is that admins are often busy with keeping the network running or meeting project deadlines. Backup management is often the first task that is neglected when admins are under time pressure because daily operations are usually not affected by a faulty backup strategy. But this is only true until a backup is really needed.
I have quite some experience with backup strategies and different kinds of backup software. While it is impossible to cover all relevant topics in two blog posts, I've outlined my four golden backup strategy rules. You'll notice that some of them are perhaps a bit uncommon.
Don't schedule backups ^
If you open a book about data backup, the first thing you usually read is that it is essential to not run backups manually and, therefore, it is essential that you schedule backup tasks. While the antecedent in this sentence is true, the conclusion is wrong.
It is true that backups should always run automatically without the need for human intervention. However, if you are still working with a backup strategy where terms such as full, incremental, or differential backups appear, then you missed an important technological development of the last years. I am talking about Continuous Data Protection (CDP).
With CDP, you don't schedule backup because the data is secured at the moment it is created. The important difference from conventional backups is that you don't lose the data that has been created between the last backup and the time when you need to restore the data.
While it might have been sufficient to run backups once per day, ten years ago, it is, in my opinion, grossly negligent to continue to pursue such a backup strategy. Most businesses now depend a lot more on digitally stored data, and any kind of data loss is no longer acceptable. A backup strategy that allows for possible data loss is what I call a "data loss strategy."
I think, CDP is now ready for prime time. The teething troubles of the first years have been solved. Thus I see no reason why anyone would still need to schedule backups.
Don't configure backup jobs ^
Backup software that relies on scheduled backups has another important downside. Depending on the size of your organization, managing backup jobs can be quite complicated. Backup solutions of major vendors have become quite bloated over the years with myriad features that appear to be quite useful at first but, if not properly configured, only increase the likelihood that restores will fail.
While I am in general a fan of bloated, feature-rich software, backup software is an exception. The difference from other software is that you can usually easily verify whether your setup works properly or not. However, with backups you often only realize that something must have been wrong with your configuration once you try to restore data.
Of course, you also have to configure CDP software. However, the configuration of CDP software is much simpler than that of conventional backup solutions because you essentially only have to configure which servers or clients you want to back up and some general settings such as retention periods. This leaves less room for mistakes and increases the probability that restores really work when you need them.
Some vendors of conventional backup software have added CDP features because they don't want to miss the trend. However, they still need to sell the legacy code of their old software; since CDP and scheduled backups don't really fit together well, things often get even more complicated if you use a conventional backup tool for CDP. Hence, I recommend moving to a CDP specialist and saying goodbye to your beloved backup software that you have been using for the last ten years or so.
Subscribe to 4sysops newsletter!
In my next post I will discuss offsite backups and disaster recovery.