TL;DR
JuiceFS is an open-source, cloud-native POSIX file system that stores data in object storage (for example S3) and keeps filesystem metadata in engines such as Redis, MySQL or TiKV. It aims to provide low-latency, strongly consistent, shared file access for big data, ML and container workloads, with compatibility layers for POSIX, Hadoop, Kubernetes and an S3 gateway.
What happened
JuiceFS is presented as a high-performance POSIX-compliant filesystem designed for cloud-native deployments. The project separates data and metadata: file contents are split into fixed-size blocks and persist in object storage, while metadata (file names, sizes, permissions, timestamps and directory structure) is stored in a pluggable metadata engine such as Redis, MySQL, SQLite or TiKV. The client coordinates reads/writes and exposes multiple interfaces including POSIX/FUSE, a Hadoop Java SDK, a Kubernetes CSI driver and an S3-compatible gateway. The project reports strong POSIX compatibility (all pjdfstest cases passed), close-to-open consistency, atomic metadata operations and support for locks, mmap, fallocate and extended attributes. Benchmarks included by the project claim higher throughput and metadata IOPS versus EFS and S3FS. JuiceFS is released under Apache License 2.0, collects anonymous core usage metrics by default (opt-out available), and documents a roadmap with optimizations and new features.
Why it matters
- Decoupling data and metadata lets users leverage scalable object stores while keeping filesystem semantics expected by applications.
- POSIX and Hadoop compatibility reduces application changes when moving workloads to cloud-backed storage.
- Strong consistency and shared access can simplify multi-client workflows for data processing and machine learning.
- Support for Kubernetes CSI and an S3 gateway helps integrate JuiceFS into containerized and object-compatible environments.
Key facts
- JuiceFS stores file data in object storage (examples: Amazon S3, GCS, Azure Blob, OSS, COS, MinIO, Ceph RGW, local disk) and metadata in engines like Redis, MySQL, SQLite or TiKV.
- Files are broken into Chunks, Slices and fixed-size Blocks (4 MiB default) before being persisted in object storage; metadata tracks these pieces.
- The project provides a JuiceFS client implementing POSIX, Hadoop SDK, Kubernetes CSI driver and an S3-compatible gateway.
- JuiceFS reports passing all 8,813 pjdfstest tests, claiming full POSIX compatibility and close-to-open consistency guarantees.
- Functional features include BSD (flock) and POSIX record (fcntl) locks, mmap support, fallocate with punch hole, xattr, encryption in transit and at rest, and optional compression (LZ4, Zstandard).
- Benchmarks published by the project suggest higher throughput and metadata IOPS compared with EFS and S3FS (details available in project docs).
- The project is open-source under Apache License 2.0 and maintains contribution, issue tracking and community channels (GitHub, Discussions, Slack).
- JuiceFS collects anonymous usage metrics by default but provides an option to disable reporting (example: juicefs mount –no-usage-report).
- A public repository shows community activity (stars and forks visible in the source repository).
What to watch next
- Roadmap items listed by the project: Gateway Optimization, Resumable Sync, Read-ahead Optimization, Optimization for Large-scale Scenarios and Snapshots.
- Adoption and integration updates for the Kubernetes CSI driver and Hadoop Java SDK as enterprise and cloud workloads evolve.
- not confirmed in the source
Quick glossary
- POSIX: A family of standards that define interoperable operating system interfaces, including filesystem semantics used by many applications.
- Object Storage: A storage architecture that manages data as objects in a flat address space, commonly used for cloud storage services like S3 or GCS.
- Metadata Engine: A database or key-value store used to keep filesystem metadata such as file names, sizes, permissions and directory structure.
- S3 Gateway: A component that exposes an S3-compatible API so object-compatible tools can access data stored through another system.
Reader FAQ
Under what license is JuiceFS released?
JuiceFS is open-source under the Apache License 2.0.
Can Redis Cluster be used as the metadata engine?
Yes; JuiceFS supports Redis Cluster but transactions must use keys in the same hash slot, so a filesystem can be limited to one hash slot.
Does JuiceFS collect usage data?
Yes, it collects anonymous core metrics by default; reporting can be disabled using the –no-usage-report option.
How does JuiceFS handle compatibility with object storage providers?
The project supports many object storage services and treats S3-compatible services as S3; for unsupported stores the guidance is to consult the supported list or report issues.
JuiceFS is a high-performance POSIX file system released under Apache License 2.0, particularly designed for the cloud-native environment. The data, stored via JuiceFS, will be persisted in Object Storage (e.g….
Sources
- JuiceFS is a distributed POSIX file system built on top of Redis and S3
- JuiceFS – Open Source Distributed POSIX File System for Cloud
- A distributed Posix file system built on top of Redis and S3
- JuiceFS S3 Gateway | JuiceFS Document Center
Related posts
- Zuck#: A PHP-inspired satirical language whose keywords satirize data harvesting
- Design and Implementation of Sprites: How Fly.io Built Disposable VMs
- TinyCity: SimCity-style micro city simulator for MicroPython devices