diff options
| author | Samuel Thibault <samuel.thibault@ens-lyon.org> | 2012-04-08 23:09:39 +0000 |
|---|---|---|
| committer | Samuel Thibault <samuel.thibault@ens-lyon.org> | 2012-04-08 23:09:39 +0000 |
| commit | 5499792c357777dcfc5ee461fa64baaba1d2dde5 (patch) | |
| tree | 15363c44fe541b3adcd720edf39696e9230d7feb /unionfs/README | |
| parent | e5fca9a10c2622b85cc77e776ea56c8dbb9847e6 (diff) | |
| parent | c6db537edac054a03847b1b99f78c00703f48d6b (diff) | |
Merge branch 'dde' of git.debian.org:/git/pkg-hurd/hurd into dde
Diffstat (limited to 'unionfs/README')
| -rw-r--r-- | unionfs/README | 92 |
1 files changed, 0 insertions, 92 deletions
diff --git a/unionfs/README b/unionfs/README deleted file mode 100644 index 1849f4d9..00000000 --- a/unionfs/README +++ /dev/null @@ -1,92 +0,0 @@ -This is the unionfs translator for the GNU Hurd. - -The unionfs translator was originally written by Moritz Schulte -<moritz@duesseldorf.ccc.de> and currently mantained by Gianluca Guida -<glguida@gmail.com>. - - - -Introduction. - -An unionfs is a filesystems that joins many filesystems into one, meaning -that you can see into an "unionfs" all files contained in the filesystems -that are part of the union. - -When two or more directories with the same path are found in different -filesystems, their content is joined. -When two or more files with the same path are found in different filesystems, -unionfs has to solve this conflict. See below (Internals section) for -information on what this implementation does in such case. - -Example: -To join "foo/" "bar/" and "baz/" in the directory "quux/", just do: - - settrans -capfg quux/ /hurd/unionfs foo/ bar/ baz/ - -If you want to join even quux/ contents in the union itself, add -u as a -translator argument. -You can add filesystems at run-time with the fsysopts command. - - - -Stowing feature. - -This unionfs implements stowing feature, i.e. the translator will watch a -directory, called 'stow', that contains the directories to unite. -When a directory is added or removed in the stow, it will be added to or -removed from the unionfs. - -Example: -To use "/stow" as the stow for the directory "foo/", do: - - settrans -capfg foo/ /hurd/unionfs --stow=/stow - -All directories contained in /stow/ will then be joined together in foo/; -you can delete or add directory at run-time and you will see unionfs adding -or removing files in foo/ automatically. - -Another interesting feature of stow support of unionfs is the pattern matching -option. -For example, by using: - - settrans -capfg foo/ /hurd/unionfs -m bar --stow=/stow - -You will get joined in foo/ all sub-sub-directories of /stow matching "bar", -i.e. /stow/*/bar/; pattern matching will be done too in run-time added stow -subdirectories. -Furthermore, you can specify more complex matching pattern to the option, -like -m bar\* (to get all stow's sub-sub-directories beginning with "bar") -or specify multiple -m options. - -Example: -This command - - settrans -capfg /myfaketree/bin -m bin -m sbin --stow=/stow - -will join in /myfaketree/bin all files that are in /stow/*/bin and -/stow/*/sbin. It is equivalent to: - - settrans -capfg /myfaketree/bin -m [s]bin --stow=/stow - - - -Internals. - -This `unionfs' translator is simple, but it is definitely not a joke. - -It works by keeping in memory a dynamically updated tree of nodes, each -node representing a directory in the unionfs. A node contains an array -of ports that give access to the corrisponding directory in the underlying -filesystems. - -On lookup, the first entry found is chosen. Thus, it is very important the -underlying filesystems ordering, since the first underlying filesystem will -be the first one to be searched during lookups, and it is the filesystem -where new files and directories are written into. - -At the moment, underlying filesystem ordering is set by option ordering. - -See CAVEAT for other unexpected behaviour that could happen. - - -Please send all bug reports to Gianluca Guida <glguida@gmail.com>. |
