DNF5: A Worthwhile Speed Boost for Fedora
I’ve used Fedora as my primary workstation for over two years and have been extremely satisfied with the distro’s update speed. However, one thing has always bothered me: the startup speed of the traditional dnf command. Even after tweaking max_parallel_downloads or switching to the best mirrors, waiting for metadata synchronization remains frustrating.
About six months ago, I decided to switch completely to DNF5 for my daily tasks. DNF5 isn’t just “old wine in a new bottle”; it’s a genuine architectural revolution. Instead of relying on the slower Python, DNF5 has been completely rewritten in C++. The result? Operations are nearly instantaneous, and RAM consumption is noticeably slimmer.
Quick Start: Experience DNF5 in a Few Commands
If you’re on Fedora 39, 40, or have recently upgraded to version 41 (where DNF5 is set to become the default), trying it out is incredibly easy. You don’t need to uninstall the old DNF, as both versions can coexist peacefully on the same system.
1. Installing DNF5
sudo dnf install dnf52. Checking the Version
dnf5 --version3. Try Installing any Package
Try installing htop to see the lightning-fast metadata loading speed:
sudo dnf5 install htopThe most obvious difference is that the transaction summary appears immediately after you press Enter. No more waiting 10-15 seconds for “Determining fastest mirrors” to finish. With DNF5, this takes only about 2-3 seconds.
Why Does DNF5 Outperform the Old DNF?
To understand why DNF5 is so powerful, we need to dissect its underlying structure. DNF4 (the current version) is a hybrid of Python and legacy C libraries. While Python offers flexibility for rapid development, it becomes a major performance bottleneck when processing tens of megabytes of metadata during every update.
Unified Library Architecture
In the past, DNF, PackageKit, and Microdnf each used their own libraries, making it impossible to share a cache. You might have just run dnf update, only for GNOME Software to start downloading metadata from scratch. DNF5 resolves this inefficiency by consolidating everything into a single libdnf5 library. All package management tools now share the same database and cache, significantly saving disk space and bandwidth.
Saying Goodbye to Python
Moving to C++ allows DNF5 to completely eliminate the overhead of the Python interpreter. This difference is stark on low-spec machines or in containers. Testing on my machine (16GB RAM), the dnf5 upgrade command consumes only about 80-100MB of RAM, whereas DNF v4 frequently climbs to the 250-300MB mark.
Advanced DNF5 Commands for Sysadmins
Fundamentally, DNF5 retains a familiar syntax, so there’s no learning curve. However, there are several valuable improvements that I believe you’ll find useful.
Better History Management
The history command in DNF5 is much more intuitive and detailed. It’s a lifesaver whenever I accidentally install a bunch of libraries that cause conflicts in my dev environment:
# List recent transactions
sudo dnf5 history
# Inspect a specific transaction (e.g., ID 5)
sudo dnf5 history info 5
# Rollback to a previous state
sudo dnf5 history undo 5Handling Advisories (Security Notifications)
For server administrators, filtering security patches is a top priority. DNF5 provides excellent support for filtering by severity:
# List all security advisories
sudo dnf5 advisory list
# Only install 'Critical' severity patches
sudo dnf5 upgrade --advisory-severity=CriticalReal-World Experience: Points to Note
After using DNF5 as my primary tool for six months, here are a few observations:
- Plugins are a work in progress: DNF4 has a massive plugin ecosystem. DNF5 is still in the process of porting these over. If you rely on specific internal repo management tools, double-check compatibility first.
- Use Aliases: To type faster, I often add
alias d5='sudo dnf5'to my.bashrcfile. I recommend against immediately renamingdnftodnf5to avoid breaking legacy system scripts. - The Cache Story: DNF5 stores its cache at
/var/cache/libdnf5/, which is entirely separate from DNF4. Therefore, it will still need time to download data on the first run—don’t judge its speed based on that!
Closing Thoughts for Fellow Fedora Users
DNF5 is a much-needed boost for the Fedora ecosystem, effectively solving the speed issues that have been a headache for so long. If you’re running Fedora, don’t hesitate to give dnf5 a try today. Honestly, once you’re used to the smoothness of DNF5, going back to DNF v4 feels quite sluggish.
In the near future, as Fedora officially transitions fully to DNF5, we’ll see even more exciting features, especially deeper integration with Podman. Have you encountered any issues during installation or have any DNF5 tips? Share them in the comments below!
