fenio/modern-fs-benchmark — explained in plain English
Analysis updated 2026-05-18
Compare how btrfs, ZFS, and bcachefs perform under snapshots, compression, and degraded disk scenarios.
Check whether a filesystem detects and repairs silently corrupted data.
See how write performance changes as snapshots accumulate over time.
Evaluate a filesystem's rebuild behavior after simulating a failed storage device.
| fenio/modern-fs-benchmark | 16nic/comfyui-agnes-ai | 6c696e68/gpt_signup_hybrid | |
|---|---|---|---|
| Stars | 19 | 19 | 19 |
| Language | Python | Python | Python |
| Setup difficulty | hard | moderate | hard |
| Complexity | 5/5 | 2/5 | 4/5 |
| Audience | ops devops | vibe coder | developer |
Figures from each repo's GitHub metadata at analysis time.
Running the full benchmark matrix requires multiple disks, several out-of-tree kernel modules, and a CI environment.
Modern FS Benchmark is a continuous testing project that measures how well different advanced filesystems handle real world storage tasks, not just raw speed. A filesystem is the system that decides how data gets organized and stored on a disk, and this project focuses on a newer style called copy on write, used by systems named btrfs, ZFS, and bcachefs, comparing them against the more traditional ext4 and xfs filesystems that most computers use by default. Rather than running a single simple speed test, this project checks many specific, realistic scenarios: how a filesystem behaves when it needs to store many small files, how quickly it can create and delete a large number of files in one folder, what happens to write speed as more and more snapshots, which are saved point in time copies of the data, pile up over time, how well compression works, and how a filesystem performs when one of its storage devices fails and needs to be rebuilt. It also deliberately corrupts data behind a filesystem's back to see whether that filesystem notices the damage and can repair it, or whether it silently serves the broken data without warning anyone. The project runs these tests automatically and regularly, then publishes the results as an online dashboard with charts that can be filtered and compared across different filesystems and storage layouts. Every test result also records the exact software versions used, which matters because some of these filesystems, like ZFS and bcachefs, are not built directly into the standard Linux kernel and need to be added separately. The author's central point is that the real question for long term data storage is not which filesystem is fastest, but which one can be trusted to tell the truth about your data, catching and fixing silent corruption rather than serving damaged files without any warning. This is a specialized tool for people who manage servers, storage systems, or personal file archives and want evidence based comparisons before choosing a filesystem, rather than a tool meant for typical end users.
A continuous benchmark suite that tests how btrfs, ZFS, and bcachefs handle snapshots, compression, device failure, and silent data corruption, compared to ext4 and xfs.
Mainly Python. The stack also includes Python, btrfs, ZFS.
Setup difficulty is rated hard, with roughly 1day+ to a first successful run.
Mainly ops devops.
This repo across BitVibe Labs
double-check against the repo, no cap.