Blog

Don't Point Me in the Wrong Direction: Why Your Data Backups Shouldn't Be Playing Follow the Leader

Written by Vish Reddy | Sep 10, 2025 11:56:05 PM

In the digital world, your data is your most valuable asset. From precious family photos to critical business documents, the thought of losing it all to a hard drive crash or a malicious attack is enough to cause a cold sweat. This is where data backups come in as our digital safety net. But what if your safety net had a critical flaw, a flaw that renders it useless precisely when you need it most? This is the danger of relying on backups that use pointers or references. This flaw often hides behind benign-sounding technical terms, like "references" in the Atlassian data language, making it even more deceptive.

What Exactly is a Pointer/Reference?

In the realm of computer programming, a pointer is essentially a variable that doesn't hold actual data itself, but rather holds the memory address of where the data is stored. Think of it like this: instead of having a book in your hand, you have a library card that tells you exactly which shelf in the library holds that book. Pointers are incredibly useful for programmers. They allow for efficient memory management, faster data manipulation, and the creation of complex data structures. In essence, they are a powerful tool for working with data without having to constantly move large chunks of it around.

It's crucial for users of platforms like Jira and Confluence to understand that the term "reference" functions in precisely the same way as a pointer. It’s not a copy of your data; it's simply a signpost that directs you back to the original file. While this is an effective method for the live application to quickly access information, it's a deeply flawed foundation for a data recovery strategy.

The Allure and Peril of Pointers / References in Backups

Now, imagine applying this concept to data backups. On the surface, it might seem clever. Instead of creating a full copy of all your files, which can take up a lot of space and time and is costly for the vendor to create, a "backup" could be created that simply contains pointers or references to the original files. This backup would be small, quick to create, cost effective (perhaps the shards on which the backup jobs are running won't melt down sorry this is an inside joke :-), a product manager told me this once) and would seemingly reference all your important data.

This approach is especially tempting for busy IT and DevOps teams. They face constant pressure to reduce storage costs, minimize backup windows, and keep systems running at peak performance. A backup that completes in minutes instead of hours and consumes kilobytes of storage instead of gigabytes looks like a major win. It allows them to tick the "backups are running" box on their checklist without the heavy overhead of traditional methods.

Herein lies the critical flaw. This "pointer / reference backup" is not a true backup at all. It's merely a map to your original data.

The Pointer's/ References Achilles' Heel: A Useless Map to a Lost Treasure

The entire purpose of a backup is to be a separate, independent copy of your data that you can restore in case the original data is lost, corrupted, or otherwise inaccessible. A backup that consists only of pointers/references completely fails this fundamental requirement.

Consider our library analogy. You have a library card (the pointer) that tells you where to find a rare, one-of-a-kind book (your original data). Now, imagine the library burns down. Your library card is perfectly intact, but it's now utterly useless. It points to a location where the book used to be, but the book itself is gone forever.

This is precisely the scenario with pointer / reference -based backups. If your original hard drive fails, is wiped by a virus, or your files are deleted, the pointers in your backup now point to empty space or corrupted information. The backup file itself is fine, but the data it's supposed to protect is gone, and the pointers are nothing more than a ghost of what once was. In the world of data recovery, this is a nightmare scenario. You believe you have a secure backup, only to discover in your moment of need that you have a map to a treasure that no longer exists.

Let's apply this to modern business risks that affect Atlassian users every day:

  • Insider Attack: An disgruntled employee deletes attachments from your live Jira, Jira Service Management, Confluence or your Assets SaaS applications. You turn to your pointer-based backup, but the pointers simply lead you back to the now-deleted and inaccessible data. The map is accurate, but it leads to the place where the attachments are supposed to exist but unfortunately are not there any longer!, you have no path to recovery.
  • Accidental Mass Deletion: A well-intentioned script or a simple human error deletes a critical Jira or JSM work item or a Confluence page . The original attachments in the data are gone. The pointers in your backup now lead to a digital void. You cannot restore what no longer exists, and the backup's "map" only confirms the location of the empty space.

There are many more such scenarios which lead to data voids by following references or pointers. Here is what Atlassian calls out in its backup methodology - "You can’t download files and attachments because we store only references to the files and not the actual files. This reduces the time to take backups."

The Revyz Solution: A True Copy of Your Treasure

Understanding this critical flaw is the first step. The second is implementing a solution that guarantees data integrity. Revyz was built on the principle that a backup must be a true, independent, and restorable copy of your critical assets. We provide a real fortress for your Atlassian data, not just the blueprint.

With Revyz, you get:

  • Complete, Independent Copies: We don’t just reference your data; we create full, air-gapped backups of your Jira, JSM and Confluence environments. This means your backup data is stored in a separate, secure location, isolated from your live system and insulated from threats like insider threat, account takeovers, ransomware etc..
  • Granular, On-Demand Restore: Disasters aren't always system-wide. Sometimes you just need to restore a single deleted work item or a specific attachment. Revyz offers granular restore capabilities, giving you the power to recover exactly what you need with surgical precision, saving time and avoiding the risks of a full rollback.
  • Automated and Testable Security: Our system provides automated, regular backups for true peace of mind. Crucially, it also allows you to easily test your recovery process, turning the hope that your backups work into the certainty that they do.

The Golden Rule of Backups: True Copies Only

The takeaway is simple: a genuine backup must be a complete, self-contained copy of your data. Whether it's a full backup of your entire system, or an incremental backup that copies changes, the backup file must contain the actual data, not just directions to where the data once resided.

While pointers / references are a cornerstone of efficient programming, they have no place in a reliable data backup strategy. When it comes to protecting your digital life, don't settle for a map. Make sure you have a true copy of your treasure.

Protect your Atlassian data with a true backup solution. Learn more about Revyz today!