The need for highly availabile and redundant network attached storage systems has never been stronger than it is today. GlusterFS is an open source, distributed file system that provides a solution and makes any system administrators job a little easier. GlusterFS is built to handle petabytes of data as well as thousands of active clients. Most importantly, Red Hat Inc. stands behind GlusterFS and is a sponsor of their development community, so you can be sure you are getting a reliable and fully tested product.
How GlusterFS Works
GlusterFS uses TCP/IP or InfiniBand RDMA to aggregate disk and memory capacity of your network servers into one namespace, and allows you to treat it like a single, much more powerful server. This setup allows GlusterFS to function much like a RAID system, as each server attached to the system will have local copies of data. Since GlusterFS is strictly software based it also has the ability to store data on whatever file system your servers use such as ntfs, ext3, or xfs.
Directories and files stored in a GlusterFS cluster can be configured in the same way as local files and directories, with permissions based on owners, users, and groups. Theses files are accessed the same as any other file, so users feel like they are interacting with any other local file.
GlusterFS Performance
Since your disk space and hardware capacity are aggregated, there is no single point of failure and no performance bottleneck in a GlusterFS system. If one server is inaccessible, the single name space belonging to your GlusterFS node is still fully accessible, and since all data has been stored across your network’s servers, any directories and files are still available. GlusterFS also has the ability to support thousands of users without the need to purchase any additional, costly equipment.
Common Use Cases
Cloud computing, media storage and distribution, high performance computing, and file system backup are a few of the common use cases for GlusterFS. Since your storage capacity can be added to whenever you need, GlusterFS is ideal for use in systems ranging from only a few users to systems that support thousands of clients constantly adding data. Whether you are setting up a NAS system for your home or office, GlusterFS should be a strong competitor to fill your needs.
Basic Concepts
As with any other system, GlusterFS has its own unique set of terms and concepts. This small list of the most important and basic concepts will help you figure out what is going on behind the scenes:
- Brick: A directory or mountpoint, each ‘brick’ serves as a building block for a GlusterFS storage pool.
- Translator: Code that forms the bond between individual bits and the storage pool, the translator performs the actions users request.
- Volume: Bricks combine to form a volume, which can be thought of as a directory in a typical filesystem.
- Node: An individual server that is participating in the GlusterFS storage pool.
Conclusion
High availability, redundancy, and ease of use are some of the most important things in the world of information technology, where downtime can be catastrophic, files need to be accessed instantaneously at the click of a button, and users of all levels need to be supported. Building a network attached storage system that meets these qualifications can be a challenge no matter your experience level, and eats up valuable time needed for other essential tasks.
GlusterFS is something that every system administrator should be aware of , and it can help you meet these requirements in an easy to understand and reliable way. Lastly, as an open source platform, GlusterFS keeps your costs manageable and has an excellent community support network sponsored by Red Hat Inc. What more could you ask for?
