1 =============================
2 Configuring a Tahoe-LAFS node
3 =============================
5 1. `Overall Node Configuration`_
6 2. `Client Configuration`_
7 3. `Storage Server Configuration`_
8 4. `Frontend Configuration`_
10 6. `Running An Introducer`_
11 7. `Other Files in BASEDIR`_
13 9. `Backwards Compatibility Files`_
16 A Tahoe-LAFS node is configured by writing to files in its base
17 directory. These files are read by the node when it starts, so each time you
18 change them, you need to restart the node.
20 The node also writes state to its base directory, so it will create files on
23 This document contains a complete list of the config files that are examined
24 by the client node, as well as the state files that you'll observe in its
27 The main file is named "``tahoe.cfg``", and is an ".INI"-style configuration
28 file (parsed by the Python stdlib 'ConfigParser' module: "``[name]``" section
29 markers, lines with "``key.subkey: value``", rfc822-style
30 continuations). There are also other files containing information that does
31 not easily fit into this format. The "``tahoe create-node``" or "``tahoe
32 create-client``" command will create an initial ``tahoe.cfg`` file for
33 you. After creation, the node will never modify the ``tahoe.cfg`` file: all
34 persistent state is put in other files.
36 The item descriptions below use the following types:
40 one of (True, yes, on, 1, False, off, no, 0), case-insensitive
44 a Twisted listening-port specification string, like "``tcp:80``" or
45 "``tcp:3456:interface=127.0.0.1``". For a full description of the format,
46 see `the Twisted strports documentation
47 <http://twistedmatrix.com/documents/current/api/twisted.application.strports.html>`_.
51 a Foolscap endpoint identifier, like
52 ``pb://soklj4y7eok5c3xkmjeqpw@192.168.69.247:44801/eqpwqtzm``
55 Overall Node Configuration
56 ==========================
58 This section controls the network behavior of the node overall: which ports
59 and IP addresses are used, when connections are timed out, etc. This
60 configuration is independent of the services that the node is offering: the
61 same controls are used for client and introducer nodes.
63 If your node is behind a firewall or NAT device and you want other clients to
64 connect to it, you'll need to open a port in the firewall or NAT, and specify
65 that port number in the tub.port option. If behind a NAT, you *may* need to
66 set the ``tub.location`` option described below.
70 ``nickname = (UTF-8 string, optional)``
72 This value will be displayed in management tools as this node's
73 "nickname". If not provided, the nickname will be set to "<unspecified>".
74 This string shall be a UTF-8 encoded Unicode string.
76 ``web.port = (strports string, optional)``
78 This controls where the node's webserver should listen, providing
79 filesystem access and node status as defined in
80 `<frontends/webapi.rst>`_. This file contains a Twisted "strports"
81 specification such as "``3456``" or "``tcp:3456:interface=127.0.0.1``".
82 The "``tahoe create-node``" or "``tahoe create-client``" commands set the
83 ``web.port`` to "``tcp:3456:interface=127.0.0.1``" by default; this is
84 overridable by the ``--webport`` option. You can make it use SSL by
85 writing "``ssl:3456:privateKey=mykey.pem:certKey=cert.pem``" instead.
87 If this is not provided, the node will not run a web server.
89 ``web.static = (string, optional)``
91 This controls where the ``/static`` portion of the URL space is
92 served. The value is a directory name (``~username`` is allowed, and
93 non-absolute names are interpreted relative to the node's basedir), which
94 can contain HTML and other files. This can be used to serve a
95 Javascript-based frontend to the Tahoe-LAFS node, or other services.
97 The default value is "``public_html``", which will serve
98 ``BASEDIR/public_html`` . With the default settings,
99 ``http://127.0.0.1:3456/static/foo.html`` will serve the contents of
100 ``BASEDIR/public_html/foo.html`` .
102 ``tub.port = (integer, optional)``
104 This controls which port the node uses to accept Foolscap connections
105 from other nodes. If not provided, the node will ask the kernel for any
106 available port. The port will be written to a separate file (named
107 ``client.port`` or ``introducer.port``), so that subsequent runs will
108 re-use the same port.
110 ``tub.location = (string, optional)``
112 In addition to running as a client, each Tahoe-LAFS node also runs as a
113 server, listening for connections from other Tahoe-LAFS clients. The node
114 announces its location by publishing a "FURL" (a string with some
115 connection hints) to the Introducer. The string it publishes can be found
116 in ``BASEDIR/private/storage.furl`` . The ``tub.location`` configuration
117 controls what location is published in this announcement.
119 If you don't provide ``tub.location``, the node will try to figure out a
120 useful one by itself, by using tools like "``ifconfig``" to determine the
121 set of IP addresses on which it can be reached from nodes both near and
122 far. It will also include the TCP port number on which it is listening
123 (either the one specified by ``tub.port``, or whichever port was assigned
124 by the kernel when ``tub.port`` is left unspecified).
126 You might want to override this value if your node lives behind a
127 firewall that is doing inbound port forwarding, or if you are using other
128 proxies such that the local IP address or port number is not the same one
129 that remote clients should use to connect. You might also want to control
130 this when using a Tor proxy to avoid revealing your actual IP address
131 through the Introducer announcement.
133 The value is a comma-separated string of host:port location hints, like
136 123.45.67.89:8098,tahoe.example.com:8098,127.0.0.1:8098
140 * Emulate default behavior, assuming your host has IP address
141 123.45.67.89 and the kernel-allocated port number was 8098::
144 tub.location = 123.45.67.89:8098,127.0.0.1:8098
146 * Use a DNS name so you can change the IP address more easily::
149 tub.location = tahoe.example.com:8098
151 * Run a node behind a firewall (which has an external IP address) that
152 has been configured to forward port 7912 to our internal node's port
156 tub.location = external-firewall.example.com:7912
158 * Run a node behind a Tor proxy (perhaps via ``torsocks``), in
159 client-only mode (i.e. we can make outbound connections, but other
160 nodes will not be able to connect to us). The literal
161 '``unreachable.example.org``' will not resolve, but will serve as a
162 reminder to human observers that this node cannot be reached. "Don't
163 call us.. we'll call you"::
166 tub.location = unreachable.example.org:0
168 * Run a node behind a Tor proxy, and make the server available as a Tor
169 "hidden service". (This assumes that other clients are running their
170 node with ``torsocks``, such that they are prepared to connect to a
171 ``.onion`` address.) The hidden service must first be configured in
172 Tor, by giving it a local port number and then obtaining a ``.onion``
173 name, using something in the ``torrc`` file like::
175 HiddenServiceDir /var/lib/tor/hidden_services/tahoe
176 HiddenServicePort 29212 127.0.0.1:8098
178 once Tor is restarted, the ``.onion`` hostname will be in
179 ``/var/lib/tor/hidden_services/tahoe/hostname``. Then set up your
183 tub.location = ualhejtq2p7ohfbb.onion:29212
185 Most users will not need to set ``tub.location``.
187 ``log_gatherer.furl = (FURL, optional)``
189 If provided, this contains a single FURL string that is used to contact a
190 "log gatherer", which will be granted access to the logport. This can be
191 used to gather operational logs in a single place. Note that in previous
192 releases of Tahoe-LAFS, if an old-style ``BASEDIR/log_gatherer.furl``
193 file existed it would also be used in addition to this value, allowing
194 multiple log gatherers to be used at once. As of Tahoe-LAFS v1.9.0, an
195 old-style file is ignored and a warning will be emitted if one is
196 detected. This means that as of Tahoe-LAFS v1.9.0 you can have at most
197 one log gatherer per node. See ticket `#1423`_ about lifting this
198 restriction and letting you have multiple log gatherers.
200 .. _`#1423`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1423
202 ``timeout.keepalive = (integer in seconds, optional)``
204 ``timeout.disconnect = (integer in seconds, optional)``
206 If ``timeout.keepalive`` is provided, it is treated as an integral number
207 of seconds, and sets the Foolscap "keepalive timer" to that value. For
208 each connection to another node, if nothing has been heard for a while,
209 we will attempt to provoke the other end into saying something. The
210 duration of silence that passes before sending the PING will be between
211 KT and 2*KT. This is mainly intended to keep NAT boxes from expiring idle
212 TCP sessions, but also gives TCP's long-duration keepalive/disconnect
213 timers some traffic to work with. The default value is 240 (i.e. 4
216 If timeout.disconnect is provided, this is treated as an integral number
217 of seconds, and sets the Foolscap "disconnect timer" to that value. For
218 each connection to another node, if nothing has been heard for a while,
219 we will drop the connection. The duration of silence that passes before
220 dropping the connection will be between DT-2*KT and 2*DT+2*KT (please see
221 ticket `#521`_ for more details). If we are sending a large amount of
222 data to the other end (which takes more than DT-2*KT to deliver), we
223 might incorrectly drop the connection. The default behavior (when this
224 value is not provided) is to disable the disconnect timer.
226 See ticket `#521`_ for a discussion of how to pick these timeout values.
227 Using 30 minutes means we'll disconnect after 22 to 68 minutes of
228 inactivity. Receiving data will reset this timeout, however if we have
229 more than 22min of data in the outbound queue (such as 800kB in two
230 pipelined segments of 10 shares each) and the far end has no need to
231 contact us, our ping might be delayed, so we may disconnect them by
234 .. _`#521`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/521
236 ``ssh.port = (strports string, optional)``
238 ``ssh.authorized_keys_file = (filename, optional)``
240 This enables an SSH-based interactive Python shell, which can be used to
241 inspect the internal state of the node, for debugging. To cause the node
242 to accept SSH connections on port 8022 from the same keys as the rest of
247 ssh.authorized_keys_file = ~/.ssh/authorized_keys
249 ``tempdir = (string, optional)``
251 This specifies a temporary directory for the web-API server to use, for
252 holding large files while they are being uploaded. If a web-API client
253 attempts to upload a 10GB file, this tempdir will need to have at least
254 10GB available for the upload to complete.
256 The default value is the ``tmp`` directory in the node's base directory
257 (i.e. ``BASEDIR/tmp``), but it can be placed elsewhere. This directory is
258 used for files that usually (on a Unix system) go into ``/tmp``. The
259 string will be interpreted relative to the node's base directory.
267 ``introducer.furl = (FURL string, mandatory)``
269 This FURL tells the client how to connect to the introducer. Each
270 Tahoe-LAFS grid is defined by an introducer. The introducer's FURL is
271 created by the introducer node and written into its base directory when
272 it starts, whereupon it should be published to everyone who wishes to
273 attach a client to that grid
275 ``helper.furl = (FURL string, optional)``
277 If provided, the node will attempt to connect to and use the given helper
278 for uploads. See `<helper.rst>`_ for details.
280 ``key_generator.furl = (FURL string, optional)``
282 If provided, the node will attempt to connect to and use the given
283 key-generator service, using RSA keys from the external process rather
284 than generating its own.
286 ``stats_gatherer.furl = (FURL string, optional)``
288 If provided, the node will connect to the given stats gatherer and
289 provide it with operational statistics.
291 ``shares.needed = (int, optional) aka "k", default 3``
293 ``shares.total = (int, optional) aka "N", N >= k, default 10``
295 ``shares.happy = (int, optional) 1 <= happy <= N, default 7``
297 These three values set the default encoding parameters. Each time a new
298 file is uploaded, erasure-coding is used to break the ciphertext into
299 separate shares. There will be ``N`` (i.e. ``shares.total``) shares
300 created, and the file will be recoverable if any ``k``
301 (i.e. ``shares.needed``) shares are retrieved. The default values are
302 3-of-10 (i.e. ``shares.needed = 3``, ``shares.total = 10``). Setting
303 ``k`` to 1 is equivalent to simple replication (uploading ``N`` copies of
306 These values control the tradeoff between storage overhead and
307 reliability. To a first approximation, a 1MB file will use (1MB *
308 ``N``/``k``) of backend storage space (the actual value will be a bit
309 more, because of other forms of overhead). Up to ``N``-``k`` shares can
310 be lost before the file becomes unrecoverable. So large ``N``/``k``
311 ratios are more reliable, and small ``N``/``k`` ratios use less disk
312 space. ``N`` cannot be larger than 256, because of the 8-bit
313 erasure-coding algorithm that Tahoe-LAFS uses. ``k`` can not be greater
314 than ``N``. See `<performance.rst>`_ for more details.
316 ``shares.happy`` allows you control over how well to "spread out" the
317 shares of an immutable file. For a successful upload, shares are
318 guaranteed to be initially placed on at least ``shares.happy`` distinct
319 servers, the correct functioning of any ``k`` of which is sufficient to
320 guarantee the availability of the uploaded file. This value should not be
321 larger than the number of servers on your grid.
323 A value of ``shares.happy`` <= ``k`` is allowed, but does not provide any
324 redundancy if some servers fail or lose shares.
326 (Mutable files use a different share placement algorithm that does not
327 currently consider this parameter.)
329 ``mutable.format = sdmf or mdmf``
331 This value tells Tahoe-LAFS what the default mutable file format should
332 be. If ``mutable.format=sdmf``, then newly created mutable files will be
333 in the old SDMF format. This is desirable for clients that operate on
334 grids where some peers run older versions of Tahoe-LAFS, as these older
335 versions cannot read the new MDMF mutable file format. If
336 ``mutable.format`` is ``mdmf``, then newly created mutable files will use
337 the new MDMF format, which supports efficient in-place modification and
338 streaming downloads. You can overwrite this value using a special
339 mutable-type parameter in the webapi. If you do not specify a value here,
340 Tahoe-LAFS will use SDMF for all newly-created mutable files.
342 Note that this parameter only applies to mutable files. Mutable
343 directories, which are stored as mutable files, are not controlled by
344 this parameter and will always use SDMF. We may revisit this decision
345 in future versions of Tahoe-LAFS.
347 Frontend Configuration
348 ======================
350 The Tahoe client process can run a variety of frontend file-access protocols.
351 You will use these to create and retrieve files from the virtual filesystem.
352 Configuration details for each are documented in the following
353 protocol-specific guides:
357 Tahoe runs a webserver by default on port 3456. This interface provides a
358 human-oriented "WUI", with pages to create, modify, and browse
359 directories and files, as well as a number of pages to check on the
360 status of your Tahoe node. It also provides a machine-oriented "WAPI",
361 with a REST-ful HTTP interface that can be used by other programs
362 (including the CLI tools). Please see `<frontends/webapi.rst>`_ for full
363 details, and the ``web.port`` and ``web.static`` config variables above.
364 The `<frontends/download-status.rst>`_ document also describes a few WUI
369 The main "bin/tahoe" executable includes subcommands for manipulating the
370 filesystem, uploading/downloading files, and creating/running Tahoe
371 nodes. See `<frontends/CLI.rst>`_ for details.
375 Tahoe can also run both FTP and SFTP servers, and map a username/password
376 pair to a top-level Tahoe directory. See `<frontends/FTP-and-SFTP.rst>`_
377 for instructions on configuring these services, and the ``[ftpd]`` and
378 ``[sftpd]`` sections of ``tahoe.cfg``.
382 As of Tahoe-LAFS v1.9.0, a node running on Linux can be configured to
383 automatically upload files that are created or changed in a specified
384 local directory. See `<frontends/drop_upload.rst>`_ for details.
388 Storage Server Configuration
389 ============================
393 ``enabled = (boolean, optional)``
395 If this is ``True``, the node will run a storage server, offering space
396 to other clients. If it is ``False``, the node will not run a storage
397 server, meaning that no shares will be stored on this node. Use ``False``
398 for clients who do not wish to provide storage service. The default value
401 ``readonly = (boolean, optional)``
403 If ``True``, the node will run a storage server but will not accept any
404 shares, making it effectively read-only. Use this for storage servers
405 that are being decommissioned: the ``storage/`` directory could be
406 mounted read-only, while shares are moved to other servers. Note that
407 this currently only affects immutable shares. Mutable shares (used for
408 directories) will be written and modified anyway. See ticket `#390
409 <http://tahoe-lafs.org/trac/tahoe-lafs/ticket/390>`_ for the current
410 status of this bug. The default value is ``False``.
412 ``reserved_space = (str, optional)``
414 If provided, this value defines how much disk space is reserved: the
415 storage server will not accept any share that causes the amount of free
416 disk space to drop below this value. (The free space is measured by a
417 call to statvfs(2) on Unix, or GetDiskFreeSpaceEx on Windows, and is the
418 space available to the user account under which the storage server runs.)
420 This string contains a number, with an optional case-insensitive scale
421 suffix like "K" or "M" or "G", and an optional "B" or "iB" suffix. So
422 "100MB", "100M", "100000000B", "100000000", and "100000kb" all mean the
423 same thing. Likewise, "1MiB", "1024KiB", and "1048576B" all mean the same
426 "``tahoe create-node``" generates a tahoe.cfg with
427 "``reserved_space=1G``", but you may wish to raise, lower, or remove the
428 reservation to suit your needs.
434 ``expire.override_lease_duration =``
436 ``expire.cutoff_date =``
438 ``expire.immutable =``
442 These settings control garbage collection, in which the server will
443 delete shares that no longer have an up-to-date lease on them. Please see
444 `<garbage-collection.rst>`_ for full details.
450 A "helper" is a regular client node that also offers the "upload helper"
455 ``enabled = (boolean, optional)``
457 If ``True``, the node will run a helper (see `<helper.rst>`_ for
458 details). The helper's contact FURL will be placed in
459 ``private/helper.furl``, from which it can be copied to any clients that
460 wish to use it. Clearly nodes should not both run a helper and attempt to
461 use one: do not create ``helper.furl`` and also define
462 ``[helper]enabled`` in the same node. The default is ``False``.
465 Running An Introducer
466 =====================
468 The introducer node uses a different ``.tac`` file (named
469 "``introducer.tac``"), and pays attention to the ``[node]`` section, but not
472 The Introducer node maintains some different state than regular client nodes.
474 ``BASEDIR/introducer.furl``
476 This is generated the first time the introducer node is started, and used
477 again on subsequent runs, to give the introduction service a persistent
478 long-term identity. This file should be published and copied into new
479 client nodes before they are started for the first time.
482 Other Files in BASEDIR
483 ======================
485 Some configuration is not kept in ``tahoe.cfg``, for the following reasons:
487 * it is generated by the node at startup, e.g. encryption keys. The node
488 never writes to ``tahoe.cfg``.
489 * it is generated by user action, e.g. the "``tahoe create-alias``" command.
491 In addition, non-configuration persistent state is kept in the node's base
492 directory, next to the configuration knobs.
494 This section describes these other files.
498 This contains an SSL private-key certificate. The node generates this the
499 first time it is started, and re-uses it on subsequent runs. This
500 certificate allows the node to have a cryptographically-strong identifier
501 (the Foolscap "TubID"), and to establish secure connections to other nodes.
505 Nodes that host StorageServers will create this directory to hold shares of
506 files on behalf of other clients. There will be a directory underneath it
507 for each StorageIndex for which this node is holding shares. There is also
508 an "incoming" directory where partially-completed shares are held while
509 they are being received.
513 This file defines the client, by constructing the actual Client instance
514 each time the node is started. It is used by the "``twistd``" daemonization
515 program (in the ``-y`` mode), which is run internally by the "``tahoe
516 start``" command. This file is created by the "``tahoe create-node``" or
517 "``tahoe create-client``" commands.
519 ``tahoe-introducer.tac``
521 This file is used to construct an introducer, and is created by the
522 "``tahoe create-introducer``" command.
524 ``tahoe-key-generator.tac``
526 This file is used to construct a key generator, and is created by the
527 "``tahoe create-key-gernerator``" command.
529 ``tahoe-stats-gatherer.tac``
531 This file is used to construct a statistics gatherer, and is created by the
532 "``tahoe create-stats-gatherer``" command.
534 ``private/control.furl``
536 This file contains a FURL that provides access to a control port on the
537 client node, from which files can be uploaded and downloaded. This file is
538 created with permissions that prevent anyone else from reading it (on
539 operating systems that support such a concept), to insure that only the
540 owner of the client node can use this feature. This port is intended for
541 debugging and testing use.
543 ``private/logport.furl``
545 This file contains a FURL that provides access to a 'log port' on the
546 client node, from which operational logs can be retrieved. Do not grant
547 logport access to strangers, because occasionally secret information may be
550 ``private/helper.furl``
552 If the node is running a helper (for use by other clients), its contact
553 FURL will be placed here. See `<helper.rst>`_ for more details.
555 ``private/root_dir.cap`` (optional)
557 The command-line tools will read a directory cap out of this file and use
558 it, if you don't specify a '--dir-cap' option or if you specify
561 ``private/convergence`` (automatically generated)
563 An added secret for encrypting immutable files. Everyone who has this same
564 string in their ``private/convergence`` file encrypts their immutable files
565 in the same way when uploading them. This causes identical files to
566 "converge" -- to share the same storage space since they have identical
567 ciphertext -- which conserves space and optimizes upload time, but it also
568 exposes file contents to the possibility of a brute-force attack by people
569 who know that string. In this attack, if the attacker can guess most of the
570 contents of a file, then they can use brute-force to learn the remaining
573 So the set of people who know your ``private/convergence`` string is the
574 set of people who converge their storage space with you when you and they
575 upload identical immutable files, and it is also the set of people who
576 could mount such an attack.
578 The content of the ``private/convergence`` file is a base-32 encoded
579 string. If the file doesn't exist, then when the Tahoe-LAFS client starts
580 up it will generate a random 256-bit string and write the base-32 encoding
581 of this string into the file. If you want to converge your immutable files
582 with as many people as possible, put the empty string (so that
583 ``private/convergence`` is a zero-length file).
591 Each Tahoe-LAFS node creates a directory to hold the log messages produced
592 as the node runs. These logfiles are created and rotated by the
593 "``twistd``" daemonization program, so ``logs/twistd.log`` will contain the
594 most recent messages, ``logs/twistd.log.1`` will contain the previous ones,
595 ``logs/twistd.log.2`` will be older still, and so on. ``twistd`` rotates
596 logfiles after they grow beyond 1MB in size. If the space consumed by
597 logfiles becomes troublesome, they should be pruned: a cron job to delete
598 all files that were created more than a month ago in this ``logs/``
599 directory should be sufficient.
603 this is written by all nodes after startup, and contains a base32-encoded
604 (i.e. human-readable) NodeID that identifies this specific node. This
605 NodeID is the same string that gets displayed on the web page (in the
606 "which peers am I connected to" list), and the shortened form (the first
607 few characters) is recorded in various log messages.
611 Gateway nodes may find it necessary to prohibit access to certain files. The
612 web-API has a facility to block access to filecaps by their storage index,
613 returning a 403 "Forbidden" error instead of the original file. For more
614 details, see the "Access Blacklist" section of `<frontends/webapi.rst>`_.
620 The following is a sample ``tahoe.cfg`` file, containing values for some of
621 the keys described in the previous section. Note that this is not a
622 recommended configuration (most of these are not the default values), merely
628 nickname = Bob's Tahoe-LAFS Node
630 tub.location = 123.45.67.89:8098,44.55.66.77:8098
632 log_gatherer.furl = pb://soklj4y7eok5c3xkmjeqpw@192.168.69.247:44801/eqpwqtzm
633 timeout.keepalive = 240
634 timeout.disconnect = 1800
636 ssh.authorized_keys_file = ~/.ssh/authorized_keys
640 introducer.furl = pb://ok45ssoklj4y7eok5c3xkmj@tahoe.example:44801/ii3uumo
641 helper.furl = pb://ggti5ssoklj4y7eok5c3xkmj@helper.tahoe.example:7054/kk8lhr
647 sizelimit = 10000000000
654 Old Configuration Files
655 =======================
657 Tahoe-LAFS releases before v1.3.0 had no ``tahoe.cfg`` file, and used
658 distinct files for each item. This is no longer supported and if you have
659 configuration in the old format you must manually convert it to the new
660 format for Tahoe-LAFS to detect it. See `<historical/configuration.rst>`_.