After years of using Windows, I assumed that the trade-off for its storage space feature was universal. This usually involves backing up everything to the desired drives, wiping the drives, creating a new pool, and then moving the files back to the drives.

I was surprised when I used mergerfs on Linux and was able to pull existing drives without touching any existing files. The process required no formatting, no migration, and no initialization. This has become the only way I connect drives.

My drives didn’t need to be replaced.

mergerfs pools that already exist – nothing is recreated.

When I first used Windows Storage Spaces, the part of the whole process that I hated the most was cleaning up my drives. This meant that I was forced to have a backup plan before using the storage spaces. Not everyone has the luxury of an extra drive large enough to back up data from all the drives they plan to add to the pool.

integration This is not required on Linux. It is a union file system based on FUSE. This is just a fancy way of saying that it shows you the drives you have installed in a folder. It does not replace the original drives; They maintain the existing file system. When I use it, my ext4, XFS, Btrfs, or NTFS drives mounted by the ntfs3 driver are the same before and after the process.

Also, it does not transfer files to drives during the process. It avoids copying or reformatting and doesn’t lock any data into a new pool format that would require merging before rereading. Because nothing about the drives changes, they remain readable even if you remove them later.

Mergerfs uses a placement policy (a simple rule for deciding where a file lands) to distribute new files across drives. I usually instruct it to back the drive with the most available space, but there are other options, and from personal experience, it’s not something I have to think about every day.

This feels like a much better process than Storage Spaces on Windows, where the first step is to create a new virtual disk and format it before moving any data inside. This time, instead of creating new storage, you’re just telling the existing drives to behave as a single unit.

It felt surprisingly easy.

Polling existing drives took less than a minute

On my home server, I run Linux Mint, and when I first decided to merge my files I was expecting some fuss. But it was surprisingly easy. After installing mergerfs, I wrote a little mount configuration to point to my current drives, and within a minute, they were all showing up as one folder.

The speed of the process was surprising and very different from Windows Storage Spaces, which would have taken me several minutes, as I would have had to first back up terabytes of media to an external drive, wipe everything, and copy everything back. Without doing much work or spending much time, I ended up with one folder instead of three.

I needed confirmation that it really worked, so I put my first files into the pool, and when I checked, they were spread across the physical drives where the placement policy directed, with the old files being placed on their original drives.

I have a mount order problem. Any drives that are not mounted before mergerfs starts do not appear in the pool. This is probably the problem if one of your drives is missing from the pool. However, it’s easy to fix: you only have to adjust when things load to boot by setting a strict mount order.

The main reason I would recommend the integration to anyone who needs to consolidate a stack of drives is that adding an additional drive requires only one configuration change. Nothing has to be rebuilt.

They solve different problems.

The better choice depends on what you are protecting.

It’s important to say that even though mergerfs and Storage Spaces seem like similar tools, they’re not always competing for the same task, so it’s hard to say that one is always the better option.

Redundancy is an interesting element that stands out on the table. Since mergerfs is not striping or mirroring any data, a failed drive means that the pooled namespace remains available, but the files on the failed drive are lost. With mirroring or parity enabled, storage spaces can survive a failed drive without data loss. This makes Storage Spaces the preferred option for protecting against failed drives without the need for additional third-party configuration.

You can connect with integration. SnapRAID Adding parity protection while maintaining the flexibility of mergerfs.

For the right setup, it’s hard to beat.

Mergerfs is not a tool that everyone should use. But I run a media server with Jellyfin, and I have a collection that keeps growing. Since it needs to be upgraded from time to time, it’s sometimes too much to demand a complete rebuild, so integration fits perfectly. If I were expecting RAID-like protection, perhaps because I keep important data, I’d combine the integration with SnapRAID.

Linux has an interesting reputation that honestly made me expect its design to be more complex. I was pleasantly surprised.