GlusterFS

GlusterFS is a scalable distributed file system that aggregates storage resources from multiple servers into a single global namespace, providing POSIX-compliant access to data across machines.

Each storage server exports one or more directories as bricks; bricks are combined into volumes. A trusted storage pool, formed by the participating servers, is managed by the GlusterD daemon on each node. A GlusterFS client mounts the volume, typically via a FUSE-based mount, and presents a POSIX file system to applications.

GlusterFS supports several volume types to tailor redundancy and performance: distributed volumes route data across bricks without replication; replicated volumes store exact copies on multiple bricks for fault tolerance; striped volumes split files across bricks for performance; dispersed volumes use erasure coding for redundancy with lower storage overhead, and combinations such as distributed-replicated volumes mix features.

Key features include data redundancy, optional self-healing, geographic replication for multi-site disaster recovery, and online rebalancing and resync. GlusterFS emphasizes commodity hardware, scalability to large deployments, and integration with cloud and virtualization environments. It can be used with container platforms such as Kubernetes and OpenShift via native persistent volumes.

GlusterFS originated as an open-source project developed by Gluster (later acquired by Red Hat in 2011) and became part of Red Hat Storage. It remains community-driven and is documented at the Gluster project site. It is used in HPC, media streaming, backups, and cloud storage deployments that require scalable, distributed file services.