Please note that these Trac pages are no longer being updated. Wiki contents/documentation have moved to GitHub.

Version 5 (modified by justinc, 7 years ago)


Installing Lind-FUSE

This document describes the steps you'll need to follow to mount the Lind POSIX file system on a normal machine. This will require the installation of FUSE and its Python bindings as well as the code for Lind. In the end, the Lind code (written in Repy V2) will be called to serve as a file system.

Install FUSE

You need to have python-fuse installed to use lind fuse. Right now there is not a lind-fuse installer. The mechanism to install FUSE depends on your OS.


First try to install it using your package manager. For example, on Ubuntu use the following command:

# apt-get install python-fuse

(On RedHat / Fedora, you'll need to use yum instead. On !OpenSuSE, you'll need to use yast.)

If none of these work, try to download the tarball at the following link instead:

If you don't use the package manager, to install the package:

$ tar zxvf fuse-python-0.2.tar.gz
$ cd fuse-python-0.2
$ sudo python install

Mac instructions (untested)

If you are using a Mac you can try installing  MacFuse. Then follow the Linux instructions above where it mentions to download the tarball and make fuse-python directly.

Load the Appropriate Seattle code

You will need to get the nacl_repy branch of Seattle from SVN.

$ svn co

Once that happens, run the script to copy the files into a directory of your choice (existing files there will be removed). For example, to use ~/test-fuse do the following:

$ cd nacl_repy
$ mkdir ~/test-fuse
$ python -t ~/test-fuse

Finding the Lind file system. Lind has many subsystems, and is install in a central location. For those subsystems to interact, they use an environment variable named REPY_PATH, which is set to the loca\ tion of the Lind and its Repy subfolder.

To check if you have fuse:

ls -l /dev/fuse


To use lind-fuse, you can simply mount a file system, and perform operations on it. Then unmount it. Some operations are not currently supported (such as df)


$ cd somepath
$ mkdir mountpoint   # make an empty directory
$ python <path_to_repy_install>/ mountpoint  # given the mountpoint you just made, mount a Lind-fuse filesystem into it.
$ cd mountpoint   # now try it out
$ ls
$ cp -r /var/logs .
$ ls
$ cd ..  # when you are done
$ fusermount -u mountpoint  # un-mount the file system.

This should work as any user (not just root).

The lind file system backing file used (or created) by the system will be whatever is in the local directory.


The core Lind-fuse files are:

  • the driver which all the filesystem calls are sent to
  • a small tester for the file system
  • the implementation of the file system calls
  • some useful file system related constants

lind fuse uses the python-fuse interface. The code is very simple. Two classes, one which handles filesystem related system calls, and the other which does file related system calls. To debug what is going on, you can run lind-fuse in the foreground, so you can see its output:

$ python mountpoint -f

Output will come out in this shell. Now go to another and do things in mountpoint.

To see which calls are being issued into your file system, set the logging global to true in You must reinstall before the changes will be reflected.



Many system calls are not implented in lind-fuse. Most of the basic ones related to file directory creation and read/write work though. More could be added easily.