Union mount: Difference between revisions
→top: Sounds better |
accuracy; moved to new category |
||
Line 1: | Line 1: | ||
A '''union mount''' is a [[Mount (computing)|mount]] that allows several [[filesystem]]s to be mounted |
A '''union mount''' is a [[Mount (computing)|mount]] that allows several [[filesystem]]s to be mounted one , appearing to be one filesystem.<ref name="usenix"/> |
||
Rather than mounting each filesystem at a different place in the directory hierarchy, a union mount overlays the filesystems, creating a unified hierarchy. Thus, any given directory (or "folder") in the resulting filesystem may contain files and subdirectories from any or all of the underlying filesystems.<ref name="fsl"/><ref name="lwn1"/><ref name="lwn2"/><ref name="lwn3"/> |
Rather than mounting each filesystem at a different place in the directory hierarchy, a union mount overlays the filesystems, creating a unified hierarchy. Thus, any given directory (or "folder") in the resulting filesystem may contain files and subdirectories from any or all of the underlying filesystems.<ref name="fsl"/><ref name="lwn1"/><ref name="lwn2"/><ref name="lwn3"/> |
||
Generally one of the filesystems will be |
Generally one of the filesystems will be read-write, while other filesystems are read-only. are by . |
||
Union mounts were invented |
Union mounts were invented ca. 1990, appearing in [[Berkeley Software Distribution#4.4BSD and descendants|4.4BSD-Lite]]; the authors of the BSD implementation cite previous work on the [[3-d filesystem]] and the [[Translucent File System]] (TFS) done at [[Bell Labs]] and [[Sun Microsystems|Sun]], respectively.<ref name="usenix"/> They are also a central concept in [[Plan 9 from Bell Labs|Plan 9]], which replaces several Unix staples with union mounts (e.g., several directories unioned together at a single <code>/bin</code> directory replace the <code>PATH</code> variable).<ref name= "use-of-namespaces">{{cite web | url= http://doc.cat-v.org/plan_9/4th_edition/papers/names | title = The Use of Name Spaces in Plan 9 | first1= R. |last1= Pike |authorlink= Ken Thompson |first2= D. | last2 = Presotto | first3 = K. |last3=Thompson |first4=H. |last4= Trickey |first5= P. |last5= Winterbottom |work= Bell Labs | publisher = [http://cat-v.org/ cat-v.org] |accessdate= 2011-12-05}}</ref> |
||
Similarly, [[GlusterFS]] offers a possibility to mount different filesystems distributed across a network, rather than being located on the same machine.<ref name="glusterfs" /> |
Similarly, [[GlusterFS]] offers a possibility to mount different filesystems distributed across a network, rather than being located on the same machine.<ref name="glusterfs" /> |
||
Line 20: | Line 20: | ||
{{use dmy dates|date=March 2012}} |
{{use dmy dates|date=March 2012}} |
||
⚫ | |||
{{Compu-storage-stub}} |
{{Compu-storage-stub}} |
||
⚫ |
Revision as of 22:33, 17 February 2015
A union mount is a mount that allows several filesystems or directories to be simultaneously mounted and visible through one mount point, appearing to be one filesystem.[1] Union mounts are implemented by union filesystems.
Rather than mounting each filesystem at a different place in the directory hierarchy, a union mount overlays the filesystems, creating a unified hierarchy. Thus, any given directory (or "folder") in the resulting filesystem may contain files and subdirectories from any or all of the underlying filesystems.[2][3][4][5]
Generally one of the constituent directories or filesystems will be treated as read-write, while other directories or filesystems are treated as read-only. If changes are written by applications to the union filesystem, they are recorded in the writeable overlay.
Union mounts were invented ca. 1990, appearing in 4.4BSD-Lite; the authors of the BSD implementation cite previous work on the 3-d filesystem and the Translucent File System (TFS) done at Bell Labs and Sun, respectively.[1] They are also a central concept in Plan 9, which replaces several Unix staples with union mounts (e.g., several directories unioned together at a single /bin
directory replace the PATH
variable).[6]
Similarly, GlusterFS offers a possibility to mount different filesystems distributed across a network, rather than being located on the same machine.[7]
References
- ^ a b Pendry, Jan-Simon; Marshall Kirk McKusick (December 1995). "Union Mounts in 4.4BSD-Lite". Proceedings of the USENIX Technical Conference on UNIX and Advanced Computing Systems: 25–33. Retrieved 25 November 2007.
- ^ Wright, Charles P.; Jay Dave; Puja Gupta; Harikesavan Krishnan; Erez Zadok; Mohammad Nayyer Zubair. "Versatility and Unix Semantics in a Fan-Out Unification File System". Stony Brook University Technical Report FSL-04-01b. Retrieved 25 November 2007.
- ^ Aurora, Valerie; Henson (March 2009). "Unioning file systems: Architecture, features, and design choices". lwn.net. Retrieved 21 December 2009.
- ^ Aurora, Valerie; Henson (March 2009). "Union file systems: Implementations, part I". lwn.net. Retrieved 21 December 2009.
- ^ Aurora, Valerie; Henson (April 2009). "Unioning file systems: Implementations, part 2". lwn.net. Retrieved 21 December 2009.
- ^ Pike, R.; Presotto, D.; Thompson, K.; Trickey, H.; Winterbottom, P. "The Use of Name Spaces in Plan 9". Bell Labs. cat-v.org. Retrieved 5 December 2011.
{{cite web}}
: External link in
(help)|publisher=
- ^ "About GlusterFS". November 2009. Retrieved 4 March 2013.
{{cite journal}}
: Cite journal requires|journal=
(help)