JuiceFS
juicefs.com
- Free plan
- Open source
- Windows
- macOS
- Linux
- API
- Self-hosted
What JuiceFS is
JuiceFS is an open-source, cloud-native distributed file system that stores file data in object storage and metadata in a separate metadata engine. It supports POSIX, HDFS, and S3 protocols, as well as WebDAV. Storage options include Amazon S3, Google Cloud Storage, Azure Blob Storage, MinIO, Ceph, WebDAV, and HDFS. Documented uses include big data, machine learning, Kubernetes, shared workspaces, and backup. Integration documentation covers Kubernetes CSI and Hadoop ecosystem components such as Spark, Presto, Hive, Flink, Trino, and StarRocks.
JuiceFS offers a self-hosted Community Edition, managed Cloud Service, and Enterprise Edition with on-premises, BYOC, and other deployment options. Its downloadable client supports Linux, macOS, and Windows; WSL 2 is recommended for Windows. Encryption in transit and at rest is supported, but at-rest encryption must be enabled when creating a file system and does not cover client-side cached data or file metadata. Community Edition is free; paid service starts at $0.02/mo. The Cloud Service comparison lists capacity of 100 million files, 10 PB, 100,000 IOPS, and 1,000 clients.
Who it suits
JuiceFS suits teams managing shared data for workloads such as big data, machine learning, or Kubernetes. It offers self-hosted, managed, and enterprise deployment options.
Good to know, and what to watch
- Supports POSIX, HDFS, S3, and WebDAV protocols.
- Integrates with Kubernetes CSI and Hadoop ecosystem tools.
- Supports encryption in transit and at rest.
- Community Edition is free.
- Cloud Service lists capacity up to 10 PB.
- At-rest encryption must be enabled when creating a file system.
- At-rest encryption does not cover cached data or file metadata.
- WSL 2 is recommended for Windows use.
Plans
Every plan and priceEnterprise Edition
Community Edition
Cloud Service
Free plan
Getting started
- Open the JuiceFS website
- Choose Community Edition for self-hosted use or a managed Cloud Service
- Install the downloadable client on Linux, macOS, or Windows
- Use WSL 2 when running the client on Windows
- Connect supported object storage and a metadata engine
- Enable at-rest encryption while creating the file system
Limits to check
Community Edition is free and self-hosted with community support through GitHub and Discord. Cloud Service lists capacity of 100 million files, 10 PB, 100,000 IOPS, and 1,000 clients; Enterprise pricing is custom.
The Club's review
JuiceFS is an open-source distributed file system for teams that need shared access to data across storage and compute environments. Its protocol breadth and deployment choices make it a strong candidate for data platforms and Kubernetes workloads, provided the team can account for its separate metadata layer and encryption limits.
Overview
JuiceFS keeps file contents in object storage and filesystem metadata in a separate metadata engine. It supports POSIX, HDFS, and S3, with WebDAV also among its storage protocols. This design can give different workloads a shared filesystem view, but it makes metadata infrastructure a distinct part of deployment planning.
The product is available as self-hosted software and as a managed cloud service. Juicedata was founded in 2017 and launched JuiceFS Cloud Service that October. The architecture and range of integrations suit organizations that want to connect storage and compute systems without tying file access to a single protocol.
Key features
Storage and protocol choices
Supported storage options include Amazon S3, Google Cloud Storage, Azure Blob Storage, MinIO, Ceph, WebDAV, and HDFS. POSIX, HDFS, and S3 support serves file-oriented applications as well as Hadoop workflows; S3-compatible access is also supported. The breadth is useful for mixed environments, though teams still need to choose and operate their storage and metadata components.
Data platform integrations
JuiceFS documents Kubernetes CSI integration and connections to Spark, Presto, Hive, Flink, Trino, and StarRocks. Those links make it a plausible shared storage layer for big data and machine learning workflows, as well as Kubernetes deployments. It also targets shared workspaces and backup, extending its use beyond analytics clusters.
Replication and protection
Data replication is supported, while erasure coding is not. That distinction matters when designing resilience: teams seeking replication can consider JuiceFS, while those requiring erasure coding will need another approach.
Encryption and compliance
JuiceFS supports encryption in transit and at rest, but at-rest encryption must be enabled when the filesystem is created and cannot be added later. It does not encrypt client-side cached data or file metadata. That makes creation-time security planning essential, and organizations that require those data paths to be encrypted should assess whether JuiceFS meets their controls. JuiceFS Cloud Service states that it is SOC 2 compliant.
Pricing
The Community Edition is free and self-hosted, with community support through GitHub and Discord. It fits teams prepared to operate their own deployment and rely on community channels. The free price removes a subscription cost, but the team takes on operating responsibilities.
Cloud Service uses pay-as-you-go billing with custom pricing. It is fully managed and lists capacity of 100 million files, 10 PB of data, 100,000 IOPS, and 1,000 clients. This is the fit for organizations that want managed operations and need capacity within those limits; the quoted starting figure of $0.02/mo does not explain a complete bill, so costs should be confirmed for the intended workload.
Enterprise Edition has custom pricing and supports on-premises, BYOC, and other deployment options, with customizable support. It fits organizations that need deployment flexibility or tailored assistance.
Platforms
The downloadable client supports Linux, macOS, and Windows, and JuiceFS also lists API and self-hosted availability. The site recommends WSL 2 for Windows use, so Windows users should plan around that guidance. Kubernetes support is documented through its CSI integration.
Who it's for
JuiceFS is best suited to teams running big data, machine learning, or Kubernetes workloads that need shared file access and already have a plan for object storage and metadata. Its protocol coverage also helps organizations connecting Hadoop ecosystem tools or multiple storage services. It is a weaker fit when deployment simplicity, encryption of cached data and metadata, or erasure coding is a firm requirement.
Pros and cons
Pros
- POSIX, HDFS, S3, and WebDAV support gives teams several ways to connect workloads.
- Storage options and integrations span major object stores, Kubernetes CSI, and multiple Hadoop ecosystem tools.
- A free self-hosted Community Edition and managed and enterprise deployment options cover different operating preferences.
- Replication is supported, and Cloud Service states SOC 2 compliance.
Cons
- File data and metadata live in separate systems, adding a distinct metadata component to deployment planning.
- At-rest encryption must be enabled at filesystem creation and leaves client-side cached data and metadata unencrypted.
- Erasure coding is not supported.
- Windows use comes with WSL 2 guidance, which adds a specific setup consideration.
Alternatives
SeaweedFS is another freemium option with a free plan; its community support has no guaranteed response time, so teams comparing community-led support should weigh that tradeoff.
BeeGFS is another freemium distributed file system with a free Community Edition for teams considering an alternative self-hosted option.
Tahoe-LAFS is free and self-hosted, with no hosted plan or paid tier stated; consider it when a free storage grid is the priority.
MooseFS offers a free Community plan with a single active master and single-parity erasure coding, making it worth considering where those characteristics suit the deployment.
Hadoop Distributed File System is a free, self-hosted Apache-licensed cluster option for teams focused on an open-source Hadoop deployment.
OrangeFS is a free open-source alternative with a downloadable release.
LizardFS is a paid option with a free plan and trial, and may suit teams comparing supported alternatives.
LucidLink is a paid alternative with a Starter plan at $7 per member/month, billed annually, with 10 members and 100 GB included per member, shared; it is relevant to teams comparing a per-member plan.
Browse more options in Distributed File System Software and Software-Defined Storage Software.
Verdict
Choose JuiceFS if you need a flexible shared filesystem for data platforms or Kubernetes and can plan for separate metadata infrastructure and creation-time encryption settings. Its protocol, storage, and integration range is the central reason to choose it; look elsewhere if encryption must cover caches and metadata or if you require erasure coding.
The verdict
JuiceFS combines object storage with a separate metadata engine and supports several common data protocols and integrations. Account for the limits of at-rest encryption and Windows guidance when planning deployment.
Questions people ask
Is JuiceFS free?
Community Edition costs nothing. Paid service starts at $0.02/mo, while Cloud Service pricing is pay-as-you-go and Enterprise pricing requires contact.
Which operating systems are supported?
The downloadable client supports Linux, macOS, and Windows. The site recommends WSL 2 for Windows use.
What storage can JuiceFS use?
Documented options include Amazon S3, Google Cloud Storage, Azure Blob Storage, MinIO, Ceph, WebDAV, and HDFS.
Does JuiceFS support Kubernetes?
Yes. JuiceFS documents Kubernetes CSI integration and lists Kubernetes among its use cases.
What encryption does it provide?
JuiceFS supports encryption in transit and at rest. At-rest encryption must be enabled at file-system creation and excludes client-side cached data and file metadata.
What the maker publishes
- API and SDKs
- The project exposes POSIX/FUSE and S3 Gateway interfaces plus Java/Hadoop and Python SDKs. source
- Architecture
- It stores file data in object storage and metadata separately in a metadata engine. source
- Build requirements
- Compiling Unix-like clients requires Go 1.20 or newer and GCC 5.4 or newer; Windows builds also require WinFsp. source
- Caching
- JuiceFS provides local, distributed and multi-level caching options for improving repeated-access performance. source
- Cloud SLA
- Cloud Service advertises 99.95% service availability and 99.99999999% data reliability. source
- Company
- Juicedata says it was founded in April 2017 and launched JuiceFS Cloud Service in October 2017. source
- Company funding
- Crunchbase lists Juicedata Inc. as the legal name, two funding rounds and seed backing led by China Growth Capital, with 11–50 employees. source
- Compliance
- JuiceFS Cloud Service states that it is SOC 2 compliant. source
- Compression
- The client supports LZ4 and Zstandard compression algorithms. source
- Consistency
- Committed file changes are immediately visible across mounted servers according to the documentation. source
- Documentation depth
- The documentation site includes introduction, architecture, deployment, benchmarking, administration, security, tutorials, reference, FAQ, development and release-note sections. source
- Encryption
- Documentation describes encryption in transit and client-side encryption at rest, while cached client data is not encrypted by default. source
- Encryption limitation
- Client-side cached data and file metadata are not encrypted by its at-rest encryption feature. source
- Founded
- 2017 source
- Hadoop SDK support
- The Hadoop Java SDK is compatible with Hadoop 2.x and Hadoop 3.x. source
- HDFS integration
- JuiceFS implements the HDFS API and integrates with Spark, Presto and Hive. source
- Independent catalog description
- Cloud Image Catalogue describes JuiceFS as open-source POSIX-compatible software storing file data in object storage and metadata in a transactional database. source
- Independent packaging
- Homebrew lists JuiceFS as a versioned formula described as a cloud-based distributed POSIX file system built on Redis and S3. source
- Integrations
- JuiceFS documents integration with Kubernetes CSI and Hadoop ecosystem components including Spark, Presto, Hive, Flink, Trino, and StarRocks. source
- Kubernetes
- A CSI Driver supports using JuiceFS as cloud-native Kubernetes storage. source
- License
- The Community Edition source code is released under Apache License 2.0. source
- Limit: benchmark variance
- A GitHub discussion reported 21.6 small-file writes per second versus 82.30 in the website benchmark until metadata placement and mount parameters were tuned. source
- Limit: macOS and Windows dependencies
- macOS FUSE mounts require macFUSE, while Windows mounts require WinFsp and use drive letters rather than directory mount points. source
- Limit: read performance
- An independent deployment report found JuiceFS reads significantly slower than NAS in one test scenario and required manual directory indexing for very large small-file sets. source
Checked against the maker's pages on 15 Oct 2026.