Differences between revisions 4 and 10 (spanning 6 versions)
⇤ ← Revision 4 as of 2006-01-27 18:39:26 →
Size: 331
Editor: RadekPodgorny
Comment:
← Revision 10 as of 2006-02-01 22:13:59 → ⇥
Size: 1151
Editor: RadekPodgorny
Comment:
Deletions are marked like this. Additions are marked like this.
Line 1: Line 1:
= unionfs-fuse = = Foreword =
Line 3: Line 3:
This is my effort to create a unionfs implementation which is probably slower but way more flexible than the current in-kernel unionfs solution. More info will come soon... This is my effort to create a unionfs implementation which is probably slower but way more flexible than the current in-kernel unionfs solution. Currently it's almost fully usable, the only problem is new file/directory creation which I will fix soon...
Line 5: Line 5:
This is still INCOMPLETE!!! == Why choose this stuff ==

 * The filesystem has to be mounted after the roots are mounted when using the standard module. With unionfs-fuse, you can mount the roots later and their contents will appear seamlesly
 * You get caching which speeds things up a lot for free
 * You get nice stats (optional)

== Why NOT choose it ==

 * Compared to kernel-space solution we need lots of useless context switches which makes kernel-only solution clear speed-winner
Line 12: Line 20:
attachment:unionfs.c == TODO ==

 * Add support for creating directories
 * Add support for choice which roots are read-only and which are not
 * Add cache controls
 * Handle writing to /stats intelligently

== Download ==

 * Latest version: 0.10 (released 2006-02-01)
 * Archive location: http://podgorny.cz/unionfs-fuse/releases/

Foreword

This is my effort to create a unionfs implementation which is probably slower but way more flexible than the current in-kernel unionfs solution. Currently it's almost fully usable, the only problem is new file/directory creation which I will fix soon...

Why choose this stuff

  • The filesystem has to be mounted after the roots are mounted when using the standard module. With unionfs-fuse, you can mount the roots later and their contents will appear seamlesly
  • You get caching which speeds things up a lot for free
  • You get nice stats (optional)

Why NOT choose it

  • Compared to kernel-space solution we need lots of useless context switches which makes kernel-only solution clear speed-winner

Available parameters

  • --roots=/root1,/root2

  • --stats

TODO

  • Add support for creating directories
  • Add support for choice which roots are read-only and which are not
  • Add cache controls
  • Handle writing to /stats intelligently

Download