From Linux NFS

(Difference between revisions)
Jump to: navigation, search
(Clean up)
(Junction Resolution)
Line 76: Line 76:
=== Junction Resolution ===
=== Junction Resolution ===
The process where a fileserver converts the contents of a junction to data that can be used to respond to a file-access client during a referral event.
The process whereby a fileserver converts the contents of a junction into a referral for a file-access client.
=== Location ===
=== Location ===

Revision as of 15:56, 27 November 2012


Project: fedfs-utils

[ Project Home | News | Downloads | Docs | Mailing Lists | Source Control | Issues ]


This article contains a glossary of terms related to FedFS.


Partially derived from draft-ietf-nfsv4-federated-fs-protocol-13.

Administrative Client

A network host that performs domain administrative tasks remotely using the NSDB and ADMIN protocols.


A user with the necessary authority to initiate administrative tasks on one or more servers in a FedFS domain.

Domain Root Directory

A domain root directory is the top-level directory of a FedFS domain.


A collection of independently administered fileservers that are linked together by a common namespace. Also known as a FedFS Domain.


FedFS is short for Federated File System. It's a set of administrative protocols and techniques for creating a file name space that can cross multiple shares on multiple fileservers, and is consistent no matter which client it is accessed from.

FedFS Domain

A FedFS domain is a file name space that can cross multiple shares on multiple fileservers. A FedFS domain is typically a single administrative entity, and has a name that is similar to a DNS domain name.

File-access client

A file-access client accessed data stored on fileservers via a standard file-access protocol such as SMB or NFS.

File-access protocol

A network filesystem access protocol such as NFSv4 or CIFS.


A fileserver provides access to file storage via a standard file-access protocol such as NFS or SMB.


A fileset is a collection of files and directories that are considered as a single administrative unit in a FedFS domain. Each fileset may reside at a single location, or it may be replicated to several locations. All files within a fileset are descendants of one directory. Filesets do not span filesystems.

Globally Useful Name

A Globally Useful Name is a pathname in the FedFS domain namespace which is the same no matter which client is used to access the file.


A junction is a link between two shared file systems. These two shared file systems might not reside on the same server. Junctions tie together separate shared file systems into a single FedFS domain namespace.

A junction's pathname is the path in a fileserver's local namespace where the junction resides. A junction's target is a list of one or more locations. Each location in this list is a replica of the same file system data.

Currently there are two types of junctions:

  • An NFS basic junction stores location information directly in each junction
  • A FedFS junction stores location information on an LDAP server so it can be shared with other fileservers

Junction Resolution

The process whereby a fileserver converts the contents of a junction into a referral for a file-access client.


A duple consisting of a fileserver host (DNS hostname or IP address) and the export path of a file system on that host. File-access clients mount a location.


A filename/directory tree that a sufficiently authorized client can access.

Namespace Database

A Namespace Database, also known as an NSDB, is the central repository of FedFS domain namespace information that is shared among fileservers in a FedFS domain. It is accessed via the LDAP protocol.


A referral is a fileserver response that tells a file-access client to look elsewhere for the shared file system it wants. A referral event can occur, for example, when a fileserver reports to a client that the object the client is attempting to access has moved. The client responds by requesting a list of locations where it can find the object.


A replica is a copy of a fileset. Replicas are used to increase availability or performance. Updates to replicas appear to occur in the same order, but do not necessarily occur simultaneously.

Personal tools