FAQ

Ceph is an open source distributed storage system that delivers object, block, and file storage in one unified platform. It runs on commodity hardware, replicates data across the cluster, and heals itself when disks or nodes fail. Start with the introduction in the documentation.

The fastest path is cephadm: bootstrap a cluster on a single host, then expand it host by host. You need Python 3, systemd, a container runtime (Podman or Docker), time synchronization, and LVM2. Follow the cephadm installation guide or our getting started page.

Yes. Ceph powers production storage for organizations worldwide, from research institutions to cloud providers. The project ships a new stable release every year, supports each series for approximately 24 months, and publishes point releases every 4 to 6 weeks. See the release cadence documentation and our case studies.

Yes. The documentation recommends Rook, a graduated CNCF project, as the way to run Ceph in Kubernetes or to connect an existing cluster to Kubernetes, and the Ceph CSI driver provisions block and file volumes for workloads. See our cloud native page.

Ceph is designed for commodity hardware; there is no compatibility list to buy against. The documentation's guidance covers CPU, memory (plan for the OSD memory target, 4 GiB by default, per OSD), separate drives for the operating system and OSD data, and at least 10 Gb/s networking between hosts. Read the hardware recommendations.

For testing, a single monitor, manager, and OSD will run. For production, the documentation recommends at least three monitors for quorum and high availability, at least two managers, and at least three OSDs (or as many as there are copies of each object). See the getting started documentation.

Pools are replicated by default: simple, fast, and resilient. Erasure coding stores the same data with less raw capacity (profiles like 4+2 or 6+3 roughly double the usable space compared to three-way replication) at a performance cost that is most visible on hard drives and during recovery. The erasure coding documentation covers the tradeoffs.

Ceph delivers enterprise storage capabilities as free and open source software on hardware you choose, with no license fees and no lock-in, and it runs on premises, in private clouds, or alongside public cloud. See why Ceph for the full comparison.

Businesses, cloud providers, universities, and research institutions around the world. Browse the use cases and case studies, or explore the public telemetry dashboard for opt-in statistics from real clusters.

Code, documentation, testing, and bug reports are all welcome. The contribute page walks through the workflow, and the developer guide covers the details.

Most of Ceph is dual licensed under the LGPL version 2.1 or 3.0, as stated in the project's COPYING file. Some bundled components carry other free software licenses, such as BSD 3-clause, Boost, and Apache 2.0, and the documentation is licensed under CC BY-SA 3.0.