GlusterFS

GlusterFS is a scalable, distributed, replicated filesystem in userspace. hastexo provides expert professional services around GlusterFS, including GlusterFS training and consulting services.

GlusterFS is a multiple-node, read-write accessible, optionally replicated and distributed filesystem implemented in FUSE (Filesystem in Userspace). Originally developed at Gluster, Inc., which was acquired by Red Hat in 2011, it now forms the core of the Red Hat Storage product line. The filesystem is fully open source, and ships with many Linux distributions.

Replication

GlusterFS supports automatic replication of files across multiple storage areas (called bricks) in an n-way capacity. This makes it possible to configure replicated volumes with 2, 3, or any number of replicas. Replication is synchronous and self-healing – when nodes temporarily drop off the cluster, their data automatically resynchronizes on the next file access. All replicas are of course writable.

Distribution

In GlusterFS distributed volumes, individual files are placed on bricks in a balanced fashion. This allows GlusterFS filesystems to scale out seamlessly. GlusterFS comes with a deterministic distribution algorithm which makes any central instance for file placement lookup unnecessary. This greatly enhances GlusterFS scalability.

Distribution can be combined with replication to ensure simultaneous scaleout and redundancy – an extremely useful configuration for High Availability solutions.

Striping

Striping allows GlusterFS to spread individual chunks of files over multiple bricks, and present them in a unified fashion. This, too, is a useful scalability feature.

Geo-replication

With asynchronous geographic replication, GlusterFS supports additional data redundancy in a backup datacenter or cloud hosted instances. Geo-replication is built on the rock-solid rsync algorithm and comes with built-in encryption via SSH tunnelling.

Geo-replication supports configurations with one master and multiple slaves, and even daisy-chained, cascading replication is supported.

Built-in NFS support

Besides its native FUSE client, GlusterFS comes with built-in support for NFS, making it a scalable drop-in replacement for the native Linux NFS server. Any TCP capable NFSv3 client can access NFS-exported GlusterFS volumes.

How we can help you with GlusterFS

Our team has expert knowledge in GlusterFS and its integration in high availability solutions based on the Pacemaker stack. We offer remote and on-site consultancy services around GlusterFS, and GlusterFS is also a key topic in our hastexo High Availability Expert training offering.

Got questions about hastexo Professional Services around GlusterFS?
Please drop us a note or Ask The Expert Now!™