1 =============================
2 Configuring a Tahoe-LAFS node
3 =============================
5 1. `Overall Node Configuration`_
6 2. `Client Configuration`_
7 3. `Storage Server Configuration`_
9 5. `Running An Introducer`_
10 6. `Other Files in BASEDIR`_
12 8. `Backwards Compatibility Files`_
15 A Tahoe-LAFS node is configured by writing to files in its base directory. These
16 files are read by the node when it starts, so each time you change them, you
17 need to restart the node.
19 The node also writes state to its base directory, so it will create files on
22 This document contains a complete list of the config files that are examined
23 by the client node, as well as the state files that you'll observe in its
26 The main file is named "``tahoe.cfg``", and is an ".INI"-style configuration
27 file (parsed by the Python stdlib 'ConfigParser' module: "``[name]``" section
28 markers, lines with "``key.subkey: value``", rfc822-style continuations). There
29 are also other files containing information that does not easily fit into this
30 format. The "``tahoe create-node``" or "``tahoe create-client``" command will
31 create an initial ``tahoe.cfg`` file for you. After creation, the node will
32 never modify the ``tahoe.cfg`` file: all persistent state is put in other files.
34 The item descriptions below use the following types:
37 one of (True, yes, on, 1, False, off, no, 0), case-insensitive
40 a Twisted listening-port specification string, like "``tcp:80``"
41 or "``tcp:3456:interface=127.0.0.1``". For a full description of
42 the format, see `the Twisted strports documentation
43 <http://twistedmatrix.com/documents/current/api/twisted.application.strports.html>`_.
46 a Foolscap endpoint identifier, like
47 ``pb://soklj4y7eok5c3xkmjeqpw@192.168.69.247:44801/eqpwqtzm``
50 Overall Node Configuration
51 ==========================
53 This section controls the network behavior of the node overall: which ports
54 and IP addresses are used, when connections are timed out, etc. This
55 configuration is independent of the services that the node is offering: the
56 same controls are used for client and introducer nodes.
58 If your node is behind a firewall or NAT device and you want other clients to
59 connect to it, you'll need to open a port in the firewall or NAT, and specify
60 that port number in the tub.port option. If behind a NAT, you *may* need to
61 set the ``tub.location`` option described below.
65 ``nickname = (UTF-8 string, optional)``
67 This value will be displayed in management tools as this node's
68 "nickname". If not provided, the nickname will be set to "<unspecified>".
69 This string shall be a UTF-8 encoded Unicode string.
71 ``web.port = (strports string, optional)``
73 This controls where the node's webserver should listen, providing
74 filesystem access and node status as defined in `webapi.rst
75 <frontends/webapi.rst>`_. This file contains a Twisted "strports"
76 specification such as "``3456``" or "``tcp:3456:interface=127.0.0.1``".
77 The "``tahoe create-node``" or "``tahoe create-client``" commands set
78 the ``web.port`` to "``tcp:3456:interface=127.0.0.1``" by default; this
79 is overridable by the ``--webport`` option. You can make it use SSL by
80 writing "``ssl:3456:privateKey=mykey.pem:certKey=cert.pem``" instead.
82 If this is not provided, the node will not run a web server.
84 ``web.static = (string, optional)``
86 This controls where the ``/static`` portion of the URL space is served. The
87 value is a directory name (``~username`` is allowed, and non-absolute names
88 are interpreted relative to the node's basedir), which can contain HTML
89 and other files. This can be used to serve a Javascript-based frontend to
90 the Tahoe-LAFS node, or other services.
92 The default value is "``public_html``", which will serve ``BASEDIR/public_html`` .
93 With the default settings, ``http://127.0.0.1:3456/static/foo.html`` will
94 serve the contents of ``BASEDIR/public_html/foo.html`` .
96 ``tub.port = (integer, optional)``
98 This controls which port the node uses to accept Foolscap connections
99 from other nodes. If not provided, the node will ask the kernel for any
100 available port. The port will be written to a separate file (named
101 ``client.port`` or ``introducer.port``), so that subsequent runs will
102 re-use the same port.
104 ``tub.location = (string, optional)``
106 In addition to running as a client, each Tahoe-LAFS node also runs as a
107 server, listening for connections from other Tahoe-LAFS clients. The node
108 announces its location by publishing a "FURL" (a string with some
109 connection hints) to the Introducer. The string it publishes can be found
110 in ``BASEDIR/private/storage.furl`` . The ``tub.location`` configuration
111 controls what location is published in this announcement.
113 If you don't provide ``tub.location``, the node will try to figure out a
114 useful one by itself, by using tools like "``ifconfig``" to determine the
115 set of IP addresses on which it can be reached from nodes both near and far.
116 It will also include the TCP port number on which it is listening (either
117 the one specified by ``tub.port``, or whichever port was assigned by the
118 kernel when ``tub.port`` is left unspecified).
120 You might want to override this value if your node lives behind a
121 firewall that is doing inbound port forwarding, or if you are using other
122 proxies such that the local IP address or port number is not the same one
123 that remote clients should use to connect. You might also want to control
124 this when using a Tor proxy to avoid revealing your actual IP address
125 through the Introducer announcement.
127 The value is a comma-separated string of host:port location hints, like
130 123.45.67.89:8098,tahoe.example.com:8098,127.0.0.1:8098
134 * Emulate default behavior, assuming your host has IP address
135 123.45.67.89 and the kernel-allocated port number was 8098::
138 tub.location = 123.45.67.89:8098,127.0.0.1:8098
140 * Use a DNS name so you can change the IP address more easily::
143 tub.location = tahoe.example.com:8098
145 * Run a node behind a firewall (which has an external IP address) that
146 has been configured to forward port 7912 to our internal node's port
150 tub.location = external-firewall.example.com:7912
152 * Run a node behind a Tor proxy (perhaps via ``torsocks``), in client-only
153 mode (i.e. we can make outbound connections, but other nodes will not
154 be able to connect to us). The literal '``unreachable.example.org``' will
155 not resolve, but will serve as a reminder to human observers that this
156 node cannot be reached. "Don't call us.. we'll call you"::
159 tub.location = unreachable.example.org:0
161 * Run a node behind a Tor proxy, and make the server available as a Tor
162 "hidden service". (This assumes that other clients are running their
163 node with ``torsocks``, such that they are prepared to connect to a
164 ``.onion`` address.) The hidden service must first be configured in
165 Tor, by giving it a local port number and then obtaining a ``.onion``
166 name, using something in the ``torrc`` file like::
168 HiddenServiceDir /var/lib/tor/hidden_services/tahoe
169 HiddenServicePort 29212 127.0.0.1:8098
171 once Tor is restarted, the ``.onion`` hostname will be in
172 ``/var/lib/tor/hidden_services/tahoe/hostname``. Then set up your
176 tub.location = ualhejtq2p7ohfbb.onion:29212
178 Most users will not need to set ``tub.location``.
180 Note that the old ``advertised_ip_addresses`` file from earlier releases is
181 no longer supported. Tahoe-LAFS v1.3.0 and later will ignore this file.
183 ``log_gatherer.furl = (FURL, optional)``
185 If provided, this contains a single FURL string that is used to contact
186 a "log gatherer", which will be granted access to the logport. This can
187 be used by centralized storage grids to gather operational logs in a
188 single place. Note that when an old-style ``BASEDIR/log_gatherer.furl`` file
189 exists (see `Backwards Compatibility Files`_, below), both are used. (For
190 most other items, the separate config file overrides the entry in
193 ``timeout.keepalive = (integer in seconds, optional)``
195 ``timeout.disconnect = (integer in seconds, optional)``
197 If ``timeout.keepalive`` is provided, it is treated as an integral number of
198 seconds, and sets the Foolscap "keepalive timer" to that value. For each
199 connection to another node, if nothing has been heard for a while, we
200 will attempt to provoke the other end into saying something. The duration
201 of silence that passes before sending the PING will be between KT and
202 2*KT. This is mainly intended to keep NAT boxes from expiring idle TCP
203 sessions, but also gives TCP's long-duration keepalive/disconnect timers
204 some traffic to work with. The default value is 240 (i.e. 4 minutes).
206 If timeout.disconnect is provided, this is treated as an integral number
207 of seconds, and sets the Foolscap "disconnect timer" to that value. For
208 each connection to another node, if nothing has been heard for a while,
209 we will drop the connection. The duration of silence that passes before
210 dropping the connection will be between DT-2*KT and 2*DT+2*KT (please see
211 ticket `#521`_ for more details). If we are sending a large amount of data
212 to the other end (which takes more than DT-2*KT to deliver), we might
213 incorrectly drop the connection. The default behavior (when this value is
214 not provided) is to disable the disconnect timer.
216 See ticket `#521`_ for a discussion of how to pick these timeout values.
217 Using 30 minutes means we'll disconnect after 22 to 68 minutes of
218 inactivity. Receiving data will reset this timeout, however if we have
219 more than 22min of data in the outbound queue (such as 800kB in two
220 pipelined segments of 10 shares each) and the far end has no need to
221 contact us, our ping might be delayed, so we may disconnect them by
224 .. _`#521`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/521
226 ``ssh.port = (strports string, optional)``
228 ``ssh.authorized_keys_file = (filename, optional)``
230 This enables an SSH-based interactive Python shell, which can be used to
231 inspect the internal state of the node, for debugging. To cause the node
232 to accept SSH connections on port 8022 from the same keys as the rest of
237 ssh.authorized_keys_file = ~/.ssh/authorized_keys
239 ``tempdir = (string, optional)``
241 This specifies a temporary directory for the web-API server to use, for
242 holding large files while they are being uploaded. If a web-API client
243 attempts to upload a 10GB file, this tempdir will need to have at least
244 10GB available for the upload to complete.
246 The default value is the ``tmp`` directory in the node's base directory
247 (i.e. ``BASEDIR/tmp``), but it can be placed elsewhere. This directory is
248 used for files that usually (on a Unix system) go into ``/tmp``. The string
249 will be interpreted relative to the node's base directory.
256 ``introducer.furl = (FURL string, mandatory)``
258 This FURL tells the client how to connect to the introducer. Each Tahoe-LAFS
259 grid is defined by an introducer. The introducer's FURL is created by the
260 introducer node and written into its base directory when it starts,
261 whereupon it should be published to everyone who wishes to attach a
264 ``helper.furl = (FURL string, optional)``
266 If provided, the node will attempt to connect to and use the given helper
267 for uploads. See `<helper.rst>`_ for details.
269 ``key_generator.furl = (FURL string, optional)``
271 If provided, the node will attempt to connect to and use the given
272 key-generator service, using RSA keys from the external process rather
273 than generating its own.
275 ``stats_gatherer.furl = (FURL string, optional)``
277 If provided, the node will connect to the given stats gatherer and
278 provide it with operational statistics.
280 ``shares.needed = (int, optional) aka "k", default 3``
282 ``shares.total = (int, optional) aka "N", N >= k, default 10``
284 ``shares.happy = (int, optional) 1 <= happy <= N, default 7``
286 These three values set the default encoding parameters. Each time a new
287 file is uploaded, erasure-coding is used to break the ciphertext into
288 separate pieces. There will be ``N`` (i.e. ``shares.total``) pieces created,
289 and the file will be recoverable if any ``k`` (i.e. ``shares.needed``)
290 pieces are retrieved. The default values are 3-of-10 (i.e.
291 ``shares.needed = 3``, ``shares.total = 10``). Setting ``k`` to 1 is
292 equivalent to simple replication (uploading ``N`` copies of the file).
294 These values control the tradeoff between storage overhead, performance,
295 and reliability. To a first approximation, a 1MB file will use (1MB * ``N``/``k``)
296 of backend storage space (the actual value will be a bit more, because of
297 other forms of overhead). Up to ``N``-``k`` shares can be lost before the file
298 becomes unrecoverable, so assuming there are at least ``N`` servers, up to
299 ``N``-``k`` servers can be offline without losing the file. So large ``N``/``k``
300 ratios are more reliable, and small ``N``/``k`` ratios use less disk space.
301 Clearly, ``k`` must never be smaller than ``N``.
303 Large values of ``N`` will slow down upload operations slightly, since more
304 servers must be involved, and will slightly increase storage overhead due
305 to the hash trees that are created. Large values of ``k`` will cause
306 downloads to be marginally slower, because more servers must be involved.
307 ``N`` cannot be larger than 256, because of the 8-bit erasure-coding
308 algorithm that Tahoe-LAFS uses.
310 ``shares.happy`` allows you control over the distribution of your immutable
311 file. For a successful upload, shares are guaranteed to be initially
312 placed on at least ``shares.happy`` distinct servers, the correct
313 functioning of any ``k`` of which is sufficient to guarantee the availability
314 of the uploaded file. This value should not be larger than the number of
315 servers on your grid.
317 A value of ``shares.happy`` <= ``k`` is allowed, but does not provide any
318 redundancy if some servers fail or lose shares.
320 (Mutable files use a different share placement algorithm that does not
321 currently consider this parameter.)
324 Storage Server Configuration
325 ============================
329 ``enabled = (boolean, optional)``
331 If this is ``True``, the node will run a storage server, offering space to
332 other clients. If it is ``False``, the node will not run a storage server,
333 meaning that no shares will be stored on this node. Use ``False`` for
334 clients who do not wish to provide storage service. The default value is
337 ``readonly = (boolean, optional)``
339 If ``True``, the node will run a storage server but will not accept any
340 shares, making it effectively read-only. Use this for storage servers
341 that are being decommissioned: the ``storage/`` directory could be mounted
342 read-only, while shares are moved to other servers. Note that this
343 currently only affects immutable shares. Mutable shares (used for
344 directories) will be written and modified anyway. See ticket `#390
345 <http://tahoe-lafs.org/trac/tahoe-lafs/ticket/390>`_ for the current
346 status of this bug. The default value is ``False``.
348 ``reserved_space = (str, optional)``
350 If provided, this value defines how much disk space is reserved: the
351 storage server will not accept any share that causes the amount of free
352 disk space to drop below this value. (The free space is measured by a
353 call to statvfs(2) on Unix, or GetDiskFreeSpaceEx on Windows, and is the
354 space available to the user account under which the storage server runs.)
356 This string contains a number, with an optional case-insensitive scale
357 suffix like "K" or "M" or "G", and an optional "B" or "iB" suffix. So
358 "100MB", "100M", "100000000B", "100000000", and "100000kb" all mean the
359 same thing. Likewise, "1MiB", "1024KiB", and "1048576B" all mean the same
362 "``tahoe create-node``" generates a tahoe.cfg with
363 "``reserved_space=1G``", but you may wish to raise, lower, or remove the
364 reservation to suit your needs.
370 ``expire.override_lease_duration =``
372 ``expire.cutoff_date =``
374 ``expire.immutable =``
378 These settings control garbage collection, in which the server will
379 delete shares that no longer have an up-to-date lease on them. Please see
380 `<garbage-collection.rst>`_ for full details.
386 A "helper" is a regular client node that also offers the "upload helper"
391 ``enabled = (boolean, optional)``
393 If ``True``, the node will run a helper (see `<helper.rst>`_ for details).
394 The helper's contact FURL will be placed in ``private/helper.furl``, from
395 which it can be copied to any clients that wish to use it. Clearly nodes
396 should not both run a helper and attempt to use one: do not create
397 ``helper.furl`` and also define ``[helper]enabled`` in the same node.
398 The default is ``False``.
401 Running An Introducer
402 =====================
404 The introducer node uses a different ``.tac`` file (named "``introducer.tac``"),
405 and pays attention to the ``[node]`` section, but not the others.
407 The Introducer node maintains some different state than regular client nodes.
409 ``BASEDIR/introducer.furl``
410 This is generated the first time the introducer node is started, and used
411 again on subsequent runs, to give the introduction service a persistent
412 long-term identity. This file should be published and copied into new client
413 nodes before they are started for the first time.
416 Other Files in BASEDIR
417 ======================
419 Some configuration is not kept in ``tahoe.cfg``, for the following reasons:
421 * it is generated by the node at startup, e.g. encryption keys. The node
422 never writes to ``tahoe.cfg``.
423 * it is generated by user action, e.g. the "``tahoe create-alias``" command.
425 In addition, non-configuration persistent state is kept in the node's base
426 directory, next to the configuration knobs.
428 This section describes these other files.
431 This contains an SSL private-key certificate. The node
432 generates this the first time it is started, and re-uses it on subsequent
433 runs. This certificate allows the node to have a cryptographically-strong
434 identifier (the Foolscap "TubID"), and to establish secure connections to
438 Nodes that host StorageServers will create this directory to hold shares
439 of files on behalf of other clients. There will be a directory underneath
440 it for each StorageIndex for which this node is holding shares. There is
441 also an "incoming" directory where partially-completed shares are held
442 while they are being received.
445 This file defines the client, by constructing the actual Client instance
446 each time the node is started. It is used by the "``twistd``" daemonization
447 program (in the ``-y`` mode), which is run internally by the "``tahoe start``"
448 command. This file is created by the "``tahoe create-node``" or
449 "``tahoe create-client``" commands.
451 ``tahoe-introducer.tac``
452 This file is used to construct an introducer, and is created by the
453 "``tahoe create-introducer``" command.
455 ``tahoe-key-generator.tac``
456 This file is used to construct a key generator, and is created by the
457 "``tahoe create-key-gernerator``" command.
459 ``tahoe-stats-gatherer.tac``
460 This file is used to construct a statistics gatherer, and is created by
461 the "``tahoe create-stats-gatherer``" command.
463 ``private/control.furl``
464 This file contains a FURL that provides access to a control port on the
465 client node, from which files can be uploaded and downloaded. This file is
466 created with permissions that prevent anyone else from reading it (on
467 operating systems that support such a concept), to insure that only the
468 owner of the client node can use this feature. This port is intended for
469 debugging and testing use.
471 ``private/logport.furl``
472 This file contains a FURL that provides access to a 'log port' on the
473 client node, from which operational logs can be retrieved. Do not grant
474 logport access to strangers, because occasionally secret information may be
477 ``private/helper.furl``
478 If the node is running a helper (for use by other clients), its contact
479 FURL will be placed here. See `<helper.rst>`_ for more details.
481 ``private/root_dir.cap`` (optional)
482 The command-line tools will read a directory cap out of this file and use
483 it, if you don't specify a '--dir-cap' option or if you specify
486 ``private/convergence`` (automatically generated)
487 An added secret for encrypting immutable files. Everyone who has this same
488 string in their ``private/convergence`` file encrypts their immutable files
489 in the same way when uploading them. This causes identical files to
490 "converge" -- to share the same storage space since they have identical
491 ciphertext -- which conserves space and optimizes upload time, but it also
492 exposes file contents to the possibility of a brute-force attack by people
493 who know that string. In this attack, if the attacker can guess most of the
494 contents of a file, then they can use brute-force to learn the remaining
497 So the set of people who know your ``private/convergence`` string is the set
498 of people who converge their storage space with you when you and they upload
499 identical immutable files, and it is also the set of people who could mount
502 The content of the ``private/convergence`` file is a base-32 encoded string.
503 If the file doesn't exist, then when the Tahoe-LAFS client starts up it will
504 generate a random 256-bit string and write the base-32 encoding of this
505 string into the file. If you want to converge your immutable files with as
506 many people as possible, put the empty string (so that ``private/convergence``
507 is a zero-length file).
513 Each Tahoe-LAFS node creates a directory to hold the log messages produced as
514 the node runs. These logfiles are created and rotated by the "``twistd``"
515 daemonization program, so ``logs/twistd.log`` will contain the most recent
516 messages, ``logs/twistd.log.1`` will contain the previous ones,
517 ``logs/twistd.log.2`` will be older still, and so on. ``twistd`` rotates
518 logfiles after they grow beyond 1MB in size. If the space consumed by logfiles
519 becomes troublesome, they should be pruned: a cron job to delete all files
520 that were created more than a month ago in this ``logs/`` directory should be
524 this is written by all nodes after startup, and contains a base32-encoded
525 (i.e. human-readable) NodeID that identifies this specific node. This
526 NodeID is the same string that gets displayed on the web page (in the
527 "which peers am I connected to" list), and the shortened form (the first
528 few characters) is recorded in various log messages.
530 Backwards Compatibility Files
531 =============================
533 Tahoe-LAFS releases before v1.3.0 had no ``tahoe.cfg`` file, and used distinct
534 files for each item listed below. For each configuration knob, if the distinct
535 file exists, it will take precedence over the corresponding item in ``tahoe.cfg``.
537 =============================== =================================== =================
538 Config setting File Comment
539 =============================== =================================== =================
540 ``[node]nickname`` ``BASEDIR/nickname``
541 ``[node]web.port`` ``BASEDIR/webport``
542 ``[node]tub.port`` ``BASEDIR/client.port`` (for Clients, not Introducers)
543 ``[node]tub.port`` ``BASEDIR/introducer.port`` (for Introducers, not Clients) (note that, unlike other keys, ``tahoe.cfg`` overrides this file)
544 ``[node]tub.location`` ``BASEDIR/advertised_ip_addresses``
545 ``[node]log_gatherer.furl`` ``BASEDIR/log_gatherer.furl`` (one per line)
546 ``[node]timeout.keepalive`` ``BASEDIR/keepalive_timeout``
547 ``[node]timeout.disconnect`` ``BASEDIR/disconnect_timeout``
548 ``[client]introducer.furl`` ``BASEDIR/introducer.furl``
549 ``[client]helper.furl`` ``BASEDIR/helper.furl``
550 ``[client]key_generator.furl`` ``BASEDIR/key_generator.furl``
551 ``[client]stats_gatherer.furl`` ``BASEDIR/stats_gatherer.furl``
552 ``[storage]enabled`` ``BASEDIR/no_storage`` (``False`` if ``no_storage`` exists)
553 ``[storage]readonly`` ``BASEDIR/readonly_storage`` (``True`` if ``readonly_storage`` exists)
554 ``[storage]sizelimit`` ``BASEDIR/sizelimit``
555 ``[storage]debug_discard`` ``BASEDIR/debug_discard_storage``
556 ``[helper]enabled`` ``BASEDIR/run_helper`` (``True`` if ``run_helper`` exists)
557 =============================== =================================== =================
559 Note: the functionality of ``[node]ssh.port`` and ``[node]ssh.authorized_keys_file``
560 were previously combined, controlled by the presence of a
561 ``BASEDIR/authorized_keys.SSHPORT`` file, in which the suffix of the filename
562 indicated which port the ssh server should listen on, and the contents of the
563 file provided the ssh public keys to accept. Support for these files has been
564 removed completely. To ``ssh`` into your Tahoe-LAFS node, add ``[node]ssh.port``
565 and ``[node].ssh_authorized_keys_file`` statements to your ``tahoe.cfg``.
567 Likewise, the functionality of ``[node]tub.location`` is a variant of the
568 now-unsupported ``BASEDIR/advertised_ip_addresses`` . The old file was additive
569 (the addresses specified in ``advertised_ip_addresses`` were used in addition to
570 any that were automatically discovered), whereas the new ``tahoe.cfg`` directive
571 is not (``tub.location`` is used verbatim).
577 The following is a sample ``tahoe.cfg`` file, containing values for all keys
578 described above. Note that this is not a recommended configuration (most of
579 these are not the default values), merely a legal one.
584 nickname = Bob's Tahoe-LAFS Node
586 tub.location = 123.45.67.89:8098,44.55.66.77:8098
588 log_gatherer.furl = pb://soklj4y7eok5c3xkmjeqpw@192.168.69.247:44801/eqpwqtzm
589 timeout.keepalive = 240
590 timeout.disconnect = 1800
592 ssh.authorized_keys_file = ~/.ssh/authorized_keys
596 introducer.furl = pb://ok45ssoklj4y7eok5c3xkmj@tahoe.example:44801/ii3uumo
597 helper.furl = pb://ggti5ssoklj4y7eok5c3xkmj@helper.tahoe.example:7054/kk8lhr
602 readonly_storage = True
603 sizelimit = 10000000000