tahoe-lafs/README
2008-01-01 01:49:21 -07:00

350 lines
15 KiB
Plaintext

Welcome to the Tahoe project, a secure, decentralized, fault-tolerant
filesystem. All of the source code is available under a Free
Software, Open Source licence.
This filesystem is encrypted and spread over multiple peers in such a
way that it remains available even when some of the peers are
unavailable, malfunctioning, or malicious.
See the web site for information, news, and discussion:
http://allmydata.org
DEPENDENCIES:
If you aren't getting a pre-compiled binary, then you'll have to ensure that
the following packages are installed before you install Tahoe.
There are two kinds of dependencies, "manual dependencies" and
"easy_install-able dependencies". The latter kind are normally automatically
satisfied for you when you install Tahoe, but if something goes wrong, please
see the EASY_INSTALLABLE DEPENDENCIES section below.
All of the manual dependencies can probably be installed through your
standard package management tool if you are running on a modern Unix
operating system. For example, on an debian-like system, you can do "sudo
apt-get install build-essential python-dev python-twisted python-pyopenssl".
The Manual Dependencies:
+ a C compiler (language)
+ GNU make (build tool)
+ Python 2.4 or newer (tested against 2.4.4, and 2.5.1 -- note that 2.4.1 is
known not to work due to a bug in its base-32 encoder), including
development headers i.e. "Python.h" (language)
http://python.org/
+ Twisted Python (tested against 2.2.0, 2.4.0, and 2.5.0) (network and
operating system integration library)
http://twistedmatrix.com/
Installing from the Twisted source tarball works on all known supported
platforms, including cygwin.
You need the following subpackages, which are included in the default
Twisted distribution:
* core (the standard Twisted package)
* web, trial, conch
Twisted requires that you manually install zope.interface, a copy of which
is included in the Twisted distribution. Note that Twisted does *not*
require the entire Zope distribution, merely the much smaller
zope.interface component.
+ OpenSSL, including development headers (cryptography library)
http://openssl.org
+ Crypto++, including development headers (cryptography library)
http://cryptopp.com
+ Python PyOpenSSL (0.6 or later) (secure transport layer)
http://pyopenssl.sourceforge.net
To install PyOpenSSL on cygwin, install the OpenSSL development libraries
with the cygwin package management tool, then get the pyOpenSSL source
code, cd into it, and run "python ./setup.py install".
GETTING THE SOURCE CODE:
You need the source code if you are going to install The Debian Way, The
Setuptools Way, or The Running-In-Place Way (see below). You do not need the
source code if you are getting precompiled binaries for Debian or Ubuntu (see
above), or if you are going to install The easy_install Way (see below).
The code is available via darcs by running the following command:
darcs get http://allmydata.org/source/tahoe/trunk tahoe
This will create a directory named "tahoe" in the current working directory
and put a copy of the latest source code into it. Later, if you want to get
any new changes, then cd into that directory and run the command "darcs
pull".
Tarballs of sources are available at:
http://allmydata.org/source/tahoe/
INSTALLING:
There are four ways to do it: The easy_install Way, The Setuptools Way, The
Running-In-Place Way, and The Debian Way. Choose one. If you're not sure,
choose the easy_install way.
The easy_install Way:
You don't need to download the source code first. You do need to have the
"easy_install" tool installed first:
http://peak.telecommunity.com/DevCenter/EasyInstall#installing-easy-install
The 'easy_install' tool can download and install tahoe for you. Just type
'easy_install allmydata-tahoe' from any shell. That will download the most
recent Tahoe source tarball, unpack it in a temporary directory, install it
to the standard location, then download and install any easy_install-able
dependencies that you need (setuptools, zfec, foolscap, simplejson, nevow,
and pycryptopp). (This will work only if you have already installed the
dependencies listed in the MANUAL DEPENDENCIES section, above.)
The end result will be that the Tahoe code is installed to the standard
location for libraries on your operating system (on unix, that is somewhere
inside /usr/lib/), and the "tahoe" executable will be installed to the
standard location for executables on operating system.
The Setuptools Way:
Get the source code (see above).
Run 'python setup.py install'. This will compile and install the Tahoe code
to the standard location for your operating system (on unix, that is
somewhere inside /usr/lib/). It will also acquire and install the
easy_install-able dependencies (setuptools, zfec, foolscap, simplejson,
nevow, and pycryptopp) to the same place. (This will work only if you have
already installed the dependencies listed in the MANUAL DEPENDENCIES
section, above.)
(To install it to a non-standard location, see
http://allmydata.org/trac/tahoe/wiki/SetuptoolsAndGNUStow .)
The end result will be that the Tahoe code is installed to the standard
location for libraries on your operating system (on unix, that is somewhere
inside /usr/lib/), and the "tahoe" executable will be installed to the
standard location for executables on operating system.
The Running-In-Place Way:
You can use Tahoe without installing it. The steps are these:
1. Get the source code (see above).
2. Run "make build-auto-deps" to install the easy_install-able dependencies
(setuptools, zfec, foolscap, simplejson, nevow, and pycryptopp) into a
local subdirectory of the Tahoe source distribution. (Note that when the
dependent libraries are updated, you should use "make clean" before
"make build-auto-deps" to make sure you will get the newest versions).
3. Build Tahoe by running "make".
4. Once you've built it then you can execute "./bin/tahoe". (When the tahoe
script is in a Tahoe source distribution, it adds the necessary
directory to the Python "sys.path". It also looks for any dependencies
that you installed by "make build-auto-deps" and includes them in the
sys.path.) See the RUNNING section, below.
The Debian Way:
The Debian Way is to build .deb files which you can then install with
"dpkg".
This requires certain debian packages (build-essential, fakeroot,
devscripts, debhelper, cdbs) to be installed first, since they are used to
construct the Tahoe .deb files. A full list of these required packages can
be found in the "Build-Depends" line in the misc/DIST/debian/control in the
top-level tahoe directory (replacing the word DIST with etch, dapper, edgy,
or feisty as appropriate).
Get the source code (see above).
If you're running on a debian system, run 'make deb-etch', 'make deb-sid',
'make deb-edgy', or 'make deb-feisty' from within the tahoe top-level
directory to construct a debian package named 'allmydata-tahoe' which you
can then install with dpkg.
TESTING THAT IT IS PROPERLY INSTALLED
If you have gotten the source code, then you can run 'make check-deps'
checks that all of the required Python package dependencies are installed.
You can run 'make test' runs the unit test suites. (This can take a long
time on slow computers. There are a lot of tests and some of them do a lot
of public-key cryptography.)
Executing the tahoe script from the "bin" subdirectory will work only if
Tahoe itself is installed, either because it is installed into the local
subdirectory (as per "The Running-In-Place Way") or because it is installed
into your system (as per the other three ways of installing).
RUNNING:
Run the "tahoe" executable.
If you installed "The Running-In-Place Way", then it is in your source tree,
in the "bin" subdirectory thereof. If you installed in one of the other
three ways, then it has been installed into your operating system's
filesystem, perhaps in "/usr/bin" on Unix, or in "C:\Python25\Scripts" on
Window.
The "tahoe" utility is used to create, start, and stop nodes. Each node
lives in a separate base directory, inside of which you can add files to
configure and control the node. Nodes also read and write files within that
directory.
A grid consists of a single central 'introducer and vdrive' node and one or
more 'client' nodes. If you are joining an existing grid, the
introducer-and-vdrive node will already be running, and you'll just need to
create a client node. If you're creating a brand new grid, you'll need to
create both an introducer-and-vdrive and a client (and then invite other
people to create their own client nodes and join your grid).
The introducer (-and-vdrive) node is constructed by running 'tahoe
create-introducer --basedir $HERE'. Once constructed, you can start the
introducer by running 'tahoe start --basedir $HERE' (or, if you are already
in the introducer's base directory, just type 'tahoe start'). Inside that
base directory, there will be a pair of files 'introducer.furl' and
'vdrive.furl'. Make a copy of these, as they'll be needed on the client
nodes. (If you want to use a publically available test grid, get the
introducer.furl and vdrive.furl files from
http://allmydata.org/trac/tahoe/wiki/TestGrid instead of running your own
introducer.)
To construct a client node, pick a new working directory for it, then run
'tahoe create-client --basedir $HERE'. Copy the two .furl files from the
introducer into this new directory, then run 'tahoe start --basedir $HERE'.
After that, the client node should be off and running. The first thing it
will do is connect to the introducer and introduce itself to all other nodes
on the grid. You can follow its progress by looking at the
$HERE/logs/twistd.log file.
create-client will put port specification into a file named $HERE/webport,
unless overridden by the --webport option to create-client. The presence of
a port specification in the webport file prompts the client node to run a
webserver on the desired port, through which you can view, upload, download,
and delete files. The contents of the webport file is actually a "strports
specification", defined in
http://twistedmatrix.com/documents/current/api/twisted.application.strports.html
, so you can have it only listen on a local interface by writing
"tcp:8123:interface=127.0.0.1" to this file (that's what create-client does
by default), or make it use SSL by writing
"ssl:8123:privateKey=mykey.pem:certKey=cert.pem" instead.
A client node directory can also be created without installing the code
first. Just use 'make create-client', and a new directory named 'CLIENTDIR'
will be created inside the top of the source tree. Copy the relevant .furl
files in, set the webport, then start the node by using 'make start-client'.
To stop it again, use 'make stop-client'. Similar makefile targets exist
for making and running an introducer node.
If you are behind a firewall and you can configure your firewall to forward
TCP connections on a port to the computer running your Tahoe node, then you
can configure the Tahoe node to announce itself as being available on that
IP address and port. The way to do this is to create a file named
$HERE/advertised_ip_addresses, in which you can put IP addresses and port
numbers in "dotted-quad:port" form, e.g. "209.97.232.113:1345". You can put
multiple IP-address-and-port-number entries into this file, on separate
lines.
There is a public grid available for testing. The necessary .furl files are
in docs/testgrid/*.furl . More information is available on
http://allmydata.org/trac/tahoe/wiki/TestGrid .
LICENCE:
This program is free software; you can redistribute it and/or modify it
under the terms of the GNU General Public License as published by the Free
Software Foundation; either version 2 of the License, or (at your option)
any later version, with the added permission that, if you become obligated
to release a derived work under this licence (as per section 2.b), you may
delay the fulfillment of this obligation for up to 12 months. If you are
obligated to release code under section 2.b of this licence, you are
obligated to release it under these same terms, including the 12-month grace
period clause. See the COPYING file for details.
EASY_INSTALLABLE DEPENDENCIES
The following Python packages are required, but normally they are
automatically installed as a side-effect of installing Tahoe.
+ Python setuptools (build and distribution tool)
http://peak.telecommunity.com/DevCenter/EasyInstall#installation-instructions
The Tahoe install process will automatically download and install
setuptools if it is not present. However, if an old, incompatible version
of setuptools is present (< v0.6c6 on Cygwin, or < v0.6a9 on other
platforms), then the install will fail.
If the install fails due to your current version of setuptools being
incompatible, please either upgrade or uninstall your version of
setuptools and re-run the install.
+ zfec (erasure coding library)
http://cheeseshop.python.org/pypi/zfec
zfec is packaged in a setuptools-compatible way and included in the Python
Package Index (PyPI), so it will be automatically installed when you
install Tahoe (see INSTALLING). It can be manually installed by running
"easy_install zfec".
+ Python foolscap (secure remote object library)
http://cheeseshop.python.org/pypi/foolscap
foolscape is packaged in a setuptools-compatible way and included in the
Python Package Index (PyPI), so it will be automatically installed when
you install Tahoe (see INSTALLING). It can be manually installed by
running "easy_install foolscap".
+ Python simplejson (JSON parser)
http://cheeseshop.python.org/pypi/simplejson
simplejson is packaged in a setuptools-compatible way and included in the
Python Package Index (PyPI), so it will be automatically installed when
you install Tahoe (see INSTALLING). It can be manually installed by
running "easy_install simplejson".
+ Python Nevow (0.6.0 or later) (web presentation language)
http://divmod.org/trac/wiki/DivmodNevow
Note that the current version of Nevow (0.9.18) requires Twisted 2.4.0 or
later.
Nevow is packaged in a setuptools-compatible way and included in the
Python Package Index (PyPI), so it will be automatically installed when
you install Tahoe (see INSTALLING). It can be manually installed by
running "easy_install nevow".
+ Python pycryptopp (0.2.1 or later) (Python crypto library)
http://pypi.python.org/pypi/pycryptopp
pycryptopp is packaged in a setuptools-compatible way and included in the
Python Package Index (PyPI), so it will be automatically installed when
you install Tahoe (see INSTALLING). It can be manually installed by
running "easy_install nevow".