Overview

MooseFS is a fault tolerant, highly available network distributed file system. It spreads data over thousands physical servers which are visible to the user as one resource. For standard file operations MooseFS acts as other Unix file systems:

Stores POSIX file attributes permissions, last access and modification times A hierarchical structure directory tree Supports special files block and character devices, pipes and sockets

Symbolic links file names pointing to target files, not necessarily on MooseFS Hard links different names of files which refer to the same data on MooseFS POSIX locks / flock locks MooseFS supports file locking

Access to the file system can be limited based on IP address and/or password.

Distinctive features

• High availability & reliability several copies of the data can be stored across separate computers
• Erasure Coding that allows to save a lot of raw disk space compared to Replication Goal
• Dynamic capacity expandable by attaching new computers and disks completely on the fly, with no service disruption.
• Transparent "trash bin" deleted files are retained for a configurable period of time
• Coherent snapshots of files are very fast and possible even while the file is being written/accessed
• "Project" quotas enable you to limit space usage per directory
• Tiering / Storage Classes give you a possibility to decide where to store the data and how to move it between various storage types or locations, thereby creating a Storage Policy for the cluster

Numbers

• Storage size is up to:
  2⁶⁴ = 16 ExbiBytes = 16 384 PebiBytes
• Single file size is up to:
  2⁵⁷ = 128 PebiBytes
• Number of files is up to:
  2³¹ = 2.1 bln
• Number of active clients is unlimited it depends on number of file descriptors in the system