1 ==================================
2 User-Visible Changes in Tahoe-LAFS
3 ==================================
5 Release 1.8.2 (2011-01-30)
6 --------------------------
8 Compatibility and Dependencies
9 ''''''''''''''''''''''''''''''
11 - Tahoe is now compatible with Twisted-10.2 (released last month), as
12 well as with earlier versions. The previous Tahoe-1.8.1 release
13 failed to run against Twisted-10.2, raising an AttributeError on
14 StreamServerEndpointService (`#1286`_)
15 - Tahoe now depends upon the "mock" testing library, and the foolscap
16 dependency was raised to 0.6.1 . It no longer requires pywin32
17 (which was used only on windows). Future developers should note that
18 reactor.spawnProcess and derivatives may no longer be used inside
24 - the default reserved_space value for new storage nodes is 1 GB
26 - documentation is now in reStructuredText (.rst) format
27 - "tahoe cp" should now handle non-ASCII filenames
28 - the unmaintained Mac/Windows GUI applications have been removed
30 - tahoe processes should appear in top and ps as "tahoe", not
31 "python", on some unix platforms. (`#174`_)
32 - "tahoe debug trial" can be used to run the test suite (`#1296`_)
33 - the SFTP frontend now reports unknown sizes as "0" instead of "?",
34 to improve compatibility with clients like FileZilla (`#1337`_)
35 - "tahoe --version" should now report correct values in situations
36 where 1.8.1 might have been wrong (`#1287`_)
38 .. _`#174`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/174
39 .. _`#1208`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1208
40 .. _`#1282`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1282
41 .. _`#1286`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1286
42 .. _`#1287`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1287
43 .. _`#1296`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1296
44 .. _`#1337`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1337
47 Release 1.8.1 (2010-10-28)
48 --------------------------
50 Bugfixes and Improvements
51 '''''''''''''''''''''''''
53 - Allow the repairer to improve the health of a file by uploading some
54 shares, even if it cannot achieve the configured happiness
55 threshold. This fixes a regression introduced between v1.7.1 and
57 - Fix a memory leak in the ResponseCache which is used during mutable
58 file/directory operations. (`#1045`_)
59 - Fix a regression and add a performance improvement in the
60 downloader. This issue caused repair to fail in some special
62 - Fix a bug that caused 'tahoe cp' to fail for a grid-to-grid copy
63 involving a non-ASCII filename. (`#1224`_)
64 - Fix a rarely-encountered bug involving printing large strings to the
65 console on Windows. (`#1232`_)
66 - Perform ~ expansion in the --exclude-from filename argument to
67 'tahoe backup'. (`#1241`_)
68 - The CLI's 'tahoe mv' and 'tahoe ln' commands previously would try to
69 use an HTTP proxy if the HTTP_PROXY environment variable was set.
70 These now always connect directly to the WAPI, thus avoiding giving
71 caps to the HTTP proxy (and also avoiding failures in the case that
72 the proxy is failing or requires authentication). (`#1253`_)
73 - The CLI now correctly reports failure in the case that 'tahoe mv'
74 fails to unlink the file from its old location. (`#1255`_)
75 - 'tahoe start' now gives a more positive indication that the node has
77 - The arguments seen by 'ps' or other tools for node processes are now
78 more useful (in particular, they include the path of the 'tahoe'
79 script, rather than an obscure tool named 'twistd'). (`#174`_)
84 - The tahoe start/stop/restart and node creation commands no longer
85 accept the -m or --multiple option, for consistency between
91 - We now host binary packages so that users on certain operating
92 systems can install without having a compiler.
93 <http://tahoe-lafs.org/source/tahoe-lafs/deps/tahoe-lafs-dep-eggs/README.html>
94 - Use a newer version of a dependency if needed, even if an older
95 version is installed. This would previously cause a VersionConflict
97 - Use a precompiled binary of a dependency if one with a sufficiently
98 high version number is available, instead of attempting to compile
99 the dependency from source, even if the source version has a higher
100 version number. (`#1233`_)
105 - All current documentation in .txt format has been converted to .rst
107 - Added docs/backdoors.rst declaring that we won't add backdoors to
108 Tahoe-LAFS, or add anything to facilitate government access to data.
111 .. _`#71`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/71
112 .. _`#174`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/174
113 .. _`#1212`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1212
114 .. _`#1045`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1045
115 .. _`#1190`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1190
116 .. _`#1216`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1216
117 .. _`#1223`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1223
118 .. _`#1224`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1224
119 .. _`#1225`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1225
120 .. _`#1232`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1232
121 .. _`#1233`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1233
122 .. _`#1241`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1241
123 .. _`#1253`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1253
124 .. _`#1255`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1255
125 .. _`#1262`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1262
128 Release 1.8.0 (2010-09-23)
129 --------------------------
134 - A completely new downloader which improves performance and
135 robustness of immutable-file downloads. It uses the fastest K
136 servers to download the data in K-way parallel. It automatically
137 fails over to alternate servers if servers fail in mid-download. It
138 allows seeking to arbitrary locations in the file (the previous
139 downloader which would only read the entire file sequentially from
140 beginning to end). It minimizes unnecessary round trips and
141 unnecessary bytes transferred to improve performance. It sends
142 requests to fewer servers to reduce the load on servers (the
143 previous one would send a small request to every server for every
144 download) (`#287`_, `#288`_, `#448`_, `#798`_, `#800`_, `#990`_,
146 - Non-ASCII command-line arguments and non-ASCII outputs now work on
147 Windows. In addition, the command-line tool now works on 64-bit
150 Bugfixes and Improvements
151 '''''''''''''''''''''''''
153 - Document and clean up the command-line options for specifying the
154 node's base directory. (`#188`_, `#706`_, `#715`_, `#772`_,
156 - The default node directory for Windows is ".tahoe" in the user's
157 home directory, the same as on other platforms. (`#890`_)
158 - Fix a case in which full cap URIs could be logged. (`#685`_,
160 - Fix bug in WUI in Python 2.5 when the system clock is set back to
161 1969. Now you can use Tahoe-LAFS with Python 2.5 and set your system
162 clock to 1969 and still use the WUI. (`#1055`_)
163 - Many improvements in code organization, tests, logging,
164 documentation, and packaging. (`#983`_, `#1074`_, `#1108`_,
165 `#1127`_, `#1129`_, `#1131`_, `#1166`_, `#1175`_)
170 - on x86 and x86-64 platforms, pycryptopp >= 0.5.20
171 - pycrypto 2.2 is excluded due to a bug
173 .. _`#188`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/188
174 .. _`#287`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/287
175 .. _`#288`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/288
176 .. _`#448`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/448
177 .. _`#685`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/685
178 .. _`#706`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/706
179 .. _`#715`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/715
180 .. _`#772`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/772
181 .. _`#798`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/798
182 .. _`#800`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/800
183 .. _`#890`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/890
184 .. _`#983`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/983
185 .. _`#990`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/990
186 .. _`#1055`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1055
187 .. _`#1074`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1074
188 .. _`#1108`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1108
189 .. _`#1155`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1155
190 .. _`#1170`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1170
191 .. _`#1191`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1191
192 .. _`#1127`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1127
193 .. _`#1129`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1129
194 .. _`#1131`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1131
195 .. _`#1166`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1166
196 .. _`#1175`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1175
198 Release 1.7.1 (2010-07-18)
199 --------------------------
201 Bugfixes and Improvements
202 '''''''''''''''''''''''''
204 - Fix bug in which uploader could fail with AssertionFailure or report
205 that it had achieved servers-of-happiness when it hadn't. (`#1118`_)
206 - Fix bug in which servers could get into a state where they would
207 refuse to accept shares of a certain file (`#1117`_)
208 - Add init scripts for managing the gateway server on Debian/Ubuntu
210 - Fix bug where server version number was always 0 on the welcome page
212 - Add new command-line command "tahoe unlink" as a synonym for "tahoe
214 - The FTP frontend now encrypts its temporary files, protecting their
215 contents from an attacker who is able to read the disk. (`#1083`_)
216 - Fix IP address detection on FreeBSD 7, 8, and 9 (`#1098`_)
217 - Fix minor layout issue in the Web User Interface with Internet
219 - Fix rarely-encountered incompatibility between Twisted logging
220 utility and the new unicode support added in v1.7.0 (`#1099`_)
221 - Forward-compatibility improvements for non-ASCII caps (`#1051`_)
226 - Simplify and tidy-up directories, unicode support, test code
227 (`#923`_, `#967`_, `#1072`_)
229 .. _`#776`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/776
230 .. _`#923`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/923
231 .. _`#961`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/961
232 .. _`#967`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/967
233 .. _`#1051`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1051
234 .. _`#1067`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1067
235 .. _`#1072`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1072
236 .. _`#1083`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1083
237 .. _`#1097`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1097
238 .. _`#1098`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1098
239 .. _`#1099`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1099
240 .. _`#1117`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1117
241 .. _`#1118`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1118
244 Release 1.7.0 (2010-06-18)
245 --------------------------
250 - SFTP support (`#1037`_)
251 Your Tahoe-LAFS gateway now acts like a full-fledged SFTP server. It
252 has been tested with sshfs to provide a virtual filesystem in Linux.
253 Many users have asked for this feature. We hope that it serves them
254 well! See the `FTP-and-SFTP.rst`_ document to get
256 - support for non-ASCII character encodings (`#534`_)
257 Tahoe-LAFS now correctly handles filenames containing non-ASCII
258 characters on all supported platforms:
260 - when reading files in from the local filesystem (such as when you
261 run "tahoe backup" to back up your local files to a Tahoe-LAFS
263 - when writing files out to the local filesystem (such as when you
264 run "tahoe cp -r" to recursively copy files out of a Tahoe-LAFS
266 - when displaying filenames to the terminal (such as when you run
267 "tahoe ls"), subject to limitations of the terminal and locale;
268 - when parsing command-line arguments, except on Windows.
270 - Servers of Happiness (`#778`_)
271 Tahoe-LAFS now measures during immutable file upload to see how well
272 distributed it is across multiple servers. It aborts the upload if
273 the pieces of the file are not sufficiently well-distributed.
274 This behavior is controlled by a configuration parameter called
275 "servers of happiness". With the default settings for its erasure
276 coding, Tahoe-LAFS generates 10 shares for each file, such that any
277 3 of those shares are sufficient to recover the file. The default
278 value of "servers of happiness" is 7, which means that Tahoe-LAFS
279 will guarantee that there are at least 7 servers holding some of the
280 shares, such that any 3 of those servers can completely recover your
281 file. The new upload code also distributes the shares better than the
282 previous version in some cases and takes better advantage of
283 pre-existing shares (when a file has already been previously
284 uploaded). See the `architecture.rst`_ document [3] for details.
286 Bugfixes and Improvements
287 '''''''''''''''''''''''''
289 - Premature abort of upload if some shares were already present and
290 some servers fail. (`#608`_)
291 - python ./setup.py install -- can't create or remove files in install
293 - Network failure => internal TypeError. (`#902`_)
294 - Install of Tahoe on CentOS 5.4. (`#933`_)
295 - CLI option --node-url now supports https url. (`#1028`_)
296 - HTML/CSS template files were not correctly installed under
298 - MetadataSetter does not enforce restriction on setting "tahoe"
300 - ImportError: No module named
301 setuptools_darcs.setuptools_darcs. (`#1054`_)
302 - Renamed Title in xhtml files. (`#1062`_)
303 - Increase Python version dependency to 2.4.4, to avoid a critical
304 CPython security bug. (`#1066`_)
305 - Typo correction for the munin plugin tahoe_storagespace. (`#968`_)
306 - Fix warnings found by pylint. (`#973`_)
307 - Changing format of some documentation files. (`#1027`_)
308 - the misc/ directory was tied up. (`#1068`_)
309 - The 'ctime' and 'mtime' metadata fields are no longer written except
310 by "tahoe backup". (`#924`_)
311 - Unicode filenames in Tahoe-LAFS directories are normalized so that
312 names that differ only in how accents are encoded are treated as the
314 - Various small improvements to documentation. (`#937`_, `#911`_,
320 - The 'tahoe debug consolidate' subcommand (for converting old
321 allmydata Windows client backups to a newer format) has been
327 - the Python version dependency is raised to 2.4.4 in some cases
328 (2.4.3 for Redhat-based Linux distributions, 2.4.2 for UCS-2 builds)
332 - mock (only required by unit tests)
334 .. _`#534`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/534
335 .. _`#608`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/608
336 .. _`#778`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/778
337 .. _`#803`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/803
338 .. _`#902`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/902
339 .. _`#911`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/911
340 .. _`#924`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/924
341 .. _`#937`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/937
342 .. _`#933`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/933
343 .. _`#968`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/968
344 .. _`#973`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/973
345 .. _`#1024`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1024
346 .. _`#1027`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1027
347 .. _`#1028`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1028
348 .. _`#1033`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1033
349 .. _`#1034`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1034
350 .. _`#1037`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1037
351 .. _`#1054`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1054
352 .. _`#1062`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1062
353 .. _`#1066`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1066
354 .. _`#1068`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1068
355 .. _`#1076`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1076
356 .. _`#1082`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/1082
357 .. _architecture.rst: docs/architecture.rst
358 .. _FTP-and-SFTP.rst: docs/frontends/FTP-and-SFTP.rst
360 Release 1.6.1 (2010-02-27)
361 --------------------------
366 - Correct handling of Small Immutable Directories
368 Immutable directories can now be deep-checked and listed in the web
369 UI in all cases. (In v1.6.0, some operations, such as deep-check, on
370 a directory graph that included very small immutable directories,
371 would result in an exception causing the whole operation to abort.)
374 Usability Improvements
375 ''''''''''''''''''''''
377 - Improved user interface messages and error reporting. (`#681`_,
379 - The timeouts for operation handles have been greatly increased, so
380 that you can view the results of an operation up to 4 days after it
381 has completed. After viewing them for the first time, the results
382 are retained for a further day. (`#577`_)
384 Release 1.6.0 (2010-02-01)
385 --------------------------
390 - Immutable Directories
392 Tahoe-LAFS can now create and handle immutable
393 directories. (`#607`_, `#833`_, `#931`_) These are read just like
394 normal directories, but are "deep-immutable", meaning that all their
395 children (and everything reachable from those children) must be
396 immutable objects (i.e. immutable or literal files, and other
397 immutable directories).
399 These directories must be created in a single webapi call that
400 provides all of the children at once. (Since they cannot be changed
401 after creation, the usual create/add/add sequence cannot be used.)
402 They have URIs that start with "URI:DIR2-CHK:" or "URI:DIR2-LIT:",
403 and are described on the human-facing web interface (aka the "WUI")
404 with a "DIR-IMM" abbreviation (as opposed to "DIR" for the usual
405 read-write directories and "DIR-RO" for read-only directories).
407 Tahoe-LAFS releases before 1.6.0 cannot read the contents of an
408 immutable directory. 1.5.0 will tolerate their presence in a
409 directory listing (and display it as "unknown"). 1.4.1 and earlier
410 cannot tolerate them: a DIR-IMM child in any directory will prevent
411 the listing of that directory.
413 Immutable directories are repairable, just like normal immutable
416 The webapi "POST t=mkdir-immutable" call is used to create immutable
417 directories. See `webapi.rst`_ for details.
419 - "tahoe backup" now creates immutable directories, backupdb has
422 The "tahoe backup" command has been enhanced to create immutable
423 directories (in previous releases, it created read-only mutable
424 directories) (`#828`_). This is significantly faster, since it does
425 not need to create an RSA keypair for each new directory. Also
426 "DIR-IMM" immutable directories are repairable, unlike "DIR-RO"
427 read-only mutable directories at present. (A future Tahoe-LAFS
428 release should also be able to repair DIR-RO.)
430 In addition, the backupdb (used by "tahoe backup" to remember what
431 it has already copied) has been enhanced to store information about
432 existing immutable directories. This allows it to re-use directories
433 that have moved but still contain identical contents, or that have
434 been deleted and later replaced. (The 1.5.0 "tahoe backup" command
435 could only re-use directories that were in the same place as they
436 were in the immediately previous backup.) With this change, the
437 backup process no longer needs to read the previous snapshot out of
438 the Tahoe-LAFS grid, reducing the network load
439 considerably. (`#606`_)
441 A "null backup" (in which nothing has changed since the previous
442 backup) will require only two Tahoe-side operations: one to add an
443 Archives/$TIMESTAMP entry, and a second to update the Latest/
444 link. On the local disk side, it will readdir() all your local
445 directories and stat() all your local files.
447 If you've been using "tahoe backup" for a while, you will notice
448 that your first use of it after upgrading to 1.6.0 may take a long
449 time: it must create proper immutable versions of all the old
450 read-only mutable directories. This process won't take as long as
451 the initial backup (where all the file contents had to be uploaded
452 too): it will require time proportional to the number and size of
453 your directories. After this initial pass, all subsequent passes
454 should take a tiny fraction of the time.
456 As noted above, Tahoe-LAFS versions earlier than 1.5.0 cannot list a
457 directory containing an immutable subdirectory. Tahoe-LAFS versions
458 earlier than 1.6.0 cannot read the contents of an immutable
461 The "tahoe backup" command has been improved to skip over unreadable
462 objects (like device files, named pipes, and files with permissions
463 that prevent the command from reading their contents), instead of
464 throwing an exception and terminating the backup process. It also
465 skips over symlinks, because these cannot be represented faithfully
466 in the Tahoe-side filesystem. A warning message will be emitted each
467 time something is skipped. (`#729`_, `#850`_, `#641`_)
469 - "create-node" command added, "create-client" now implies
472 The basic idea behind Tahoe-LAFS's client+server and client-only
473 processes is that you are creating a general-purpose Tahoe-LAFS
474 "node" process, which has several components that can be
475 activated. Storage service is one of these optional components, as
476 is the Helper, FTP server, and SFTP server. Web gateway
477 functionality is nominally on this list, but it is always active; a
478 future release will make it optional. There are three special
479 purpose servers that can't currently be run as a component in a
480 node: introducer, key-generator, and stats-gatherer.
482 So now "tahoe create-node" will create a Tahoe-LAFS node process,
483 and after creation you can edit its tahoe.cfg to enable or disable
484 the desired services. It is a more general-purpose replacement for
485 "tahoe create-client". The default configuration has storage
486 service enabled. For convenience, the "--no-storage" argument makes
487 a tahoe.cfg file that disables storage service. (`#760`_)
489 "tahoe create-client" has been changed to create a Tahoe-LAFS node
490 without a storage service. It is equivalent to "tahoe create-node
491 --no-storage". This helps to reduce the confusion surrounding the
492 use of a command with "client" in its name to create a storage
493 *server*. Use "tahoe create-client" to create a purely client-side
494 node. If you want to offer storage to the grid, use "tahoe
495 create-node" instead.
497 In the future, other services will be added to the node, and they
498 will be controlled through options in tahoe.cfg . The most important
499 of these services may get additional --enable-XYZ or --disable-XYZ
500 arguments to "tahoe create-node".
502 - Performance Improvements
504 Download of immutable files begins as soon as the downloader has
505 located the K necessary shares (`#928`_, `#287`_). In both the
506 previous and current releases, a downloader will first issue queries
507 to all storage servers on the grid to locate shares before it begins
508 downloading the shares. In previous releases of Tahoe-LAFS, download
509 would not begin until all storage servers on the grid had replied to
510 the query, at which point K shares would be chosen for download from
511 among the shares that were located. In this release, download begins
512 as soon as any K shares are located. This means that downloads start
513 sooner, which is particularly important if there is a server on the
514 grid that is extremely slow or even hung in such a way that it will
515 never respond. In previous releases such a server would have a
516 negative impact on all downloads from that grid. In this release,
517 such a server will have no impact on downloads, as long as K shares
518 can be found on other, quicker, servers. This also means that
519 downloads now use the "best-alacrity" servers that they talk to, as
520 measured by how quickly the servers reply to the initial query. This
521 might cause downloads to go faster, especially on grids with
522 heterogeneous servers or geographical dispersion.
527 - The webapi acquired a new "t=mkdir-with-children" command, to create
528 and populate a directory in a single call. This is significantly
529 faster than using separate "t=mkdir" and "t=set-children" operations
530 (it uses one gateway-to-grid roundtrip, instead of three or
533 - The t=set-children (note the hyphen) operation is now documented in
534 webapi.rst, and is the new preferred spelling of the
535 old t=set_children (with an underscore). The underscore version
536 remains for backwards compatibility. (`#381`_, `#927`_)
538 - The tracebacks produced by errors in CLI tools should now be in
539 plain text, instead of HTML (which is unreadable outside of a
542 - The [storage]reserved_space configuration knob (which causes the
543 storage server to refuse shares when available disk space drops
544 below a threshold) should work on Windows now, not just
547 - "tahoe cp" should now exit with status "1" if it cannot figure out a
548 suitable target filename, such as when you copy from a bare
551 - "tahoe get" no longer creates a zero-length file upon
554 - "tahoe ls" can now list single files. (`#457`_)
556 - "tahoe deep-check --repair" should tolerate repair failures now,
557 instead of halting traversal. (`#874`_, `#786`_)
559 - "tahoe create-alias" no longer corrupts the aliases file if it had
560 previously been edited to have no trailing newline. (`#741`_)
562 - Many small packaging improvements were made to facilitate the
563 "tahoe-lafs" package being included in Ubuntu. Several mac/win32
564 binary libraries were removed, some figleaf code-coverage files were
565 removed, a bundled copy of darcsver-1.2.1 was removed, and
566 additional licensing text was added.
568 - Several DeprecationWarnings for python2.6 were silenced. (`#859`_)
570 - The checker --add-lease option would sometimes fail for shares
571 stored on old (Tahoe v1.2.0) servers. (`#875`_)
573 - The documentation for installing on Windows (docs/quickstart.rst)
574 has been improved. (`#773`_)
576 For other changes not mentioned here, see
577 <http://tahoe-lafs.org/trac/tahoe/query?milestone=1.6.0&keywords=!~news-done>.
578 To include the tickets mentioned above, go to
579 <http://tahoe-lafs.org/trac/tahoe/query?milestone=1.6.0>.
581 .. _`#121`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/121
582 .. _`#287`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/287
583 .. _`#381`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/381
584 .. _`#457`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/457
585 .. _`#533`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/533
586 .. _`#577`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/577
587 .. _`#606`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/606
588 .. _`#607`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/607
589 .. _`#637`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/637
590 .. _`#641`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/641
591 .. _`#646`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/646
592 .. _`#681`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/681
593 .. _`#729`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/729
594 .. _`#741`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/741
595 .. _`#760`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/760
596 .. _`#761`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/761
597 .. _`#768`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/768
598 .. _`#773`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/773
599 .. _`#786`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/786
600 .. _`#828`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/828
601 .. _`#833`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/833
602 .. _`#859`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/859
603 .. _`#874`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/874
604 .. _`#875`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/875
605 .. _`#931`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/931
606 .. _`#837`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/837
607 .. _`#850`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/850
608 .. _`#927`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/927
609 .. _`#928`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/928
610 .. _`#939`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/939
611 .. _`#948`: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/948
612 .. _webapi.rst: docs/frontends/webapi.rst
614 Release 1.5.0 (2009-08-01)
615 --------------------------
620 - Uploads of immutable files now use pipelined writes, improving
621 upload speed slightly (10%) over high-latency connections. (`#392`_)
623 - Processing large directories has been sped up, by removing a O(N^2)
624 algorithm from the dirnode decoding path and retaining unmodified
625 encrypted entries. (`#750`_, `#752`_)
627 - The human-facing web interface (aka the "WUI") received a
628 significant CSS makeover by Kevin Reid, making it much prettier and
629 easier to read. The WUI "check" and "deep-check" forms now include a
630 "Renew Lease" checkbox, mirroring the CLI --add-lease option, so
631 leases can be added or renewed from the web interface.
633 - The CLI "tahoe mv" command now refuses to overwrite
634 directories. (`#705`_)
636 - The CLI "tahoe webopen" command, when run without arguments, will
637 now bring up the "Welcome Page" (node status and mkdir/upload
640 - The 3.5MB limit on mutable files was removed, so it should be
641 possible to upload arbitrarily-sized mutable files. Note, however,
642 that the data format and algorithm remains the same, so using
643 mutable files still requires bandwidth, computation, and RAM in
644 proportion to the size of the mutable file. (`#694`_)
646 - This version of Tahoe-LAFS will tolerate directory entries that
647 contain filecap formats which it does not recognize: files and
648 directories from the future. This should improve the user
649 experience (for 1.5.0 users) when we add new cap formats in the
650 future. Previous versions would fail badly, preventing the user from
651 seeing or editing anything else in those directories. These
652 unrecognized objects can be renamed and deleted, but obviously not
653 read or written. Also they cannot generally be copied. (`#683`_)
658 - deep-check-and-repair now tolerates read-only directories, such as
659 the ones produced by the "tahoe backup" CLI command. Read-only
660 directories and mutable files are checked, but not
661 repaired. Previous versions threw an exception when attempting the
662 repair and failed to process the remaining contents. We cannot yet
663 repair these read-only objects, but at least this version allows the
664 rest of the check+repair to proceed. (`#625`_)
666 - A bug in 1.4.1 which caused a server to be listed multiple times
667 (and frequently broke all connections to that server) was
670 - The plaintext-hashing code was removed from the Helper interface,
671 removing the Helper's ability to mount a
672 partial-information-guessing attack. (`#722`_)
674 Platform/packaging changes
675 ''''''''''''''''''''''''''
677 - Tahoe-LAFS now runs on NetBSD, OpenBSD, ArchLinux, and NixOS, and on
678 an embedded system based on an ARM CPU running at 266 MHz.
680 - Unit test timeouts have been raised to allow the tests to complete
681 on extremely slow platforms like embedded ARM-based NAS boxes, which
682 may take several hours to run the test suite. An ARM-specific
683 data-corrupting bug in an older version of Crypto++ (5.5.2) was
684 identified: ARM-users are encouraged to use recent
685 Crypto++/pycryptopp which avoids this problem.
687 - Tahoe-LAFS now requires a SQLite library, either the sqlite3 that
688 comes built-in with python2.5/2.6, or the add-on pysqlite2 if you're
689 using python2.4. In the previous release, this was only needed for
690 the "tahoe backup" command: now it is mandatory.
692 - Several minor documentation updates were made.
694 - To help get Tahoe-LAFS into Linux distributions like Fedora and
695 Debian, packaging improvements are being made in both Tahoe-LAFS and
696 related libraries like pycryptopp and zfec.
698 - The Crypto++ library included in the pycryptopp package has been
699 upgraded to version 5.6.0 of Crypto++, which includes a more
700 efficient implementation of SHA-256 in assembly for x86 or amd64
707 - no python-2.4.0 or 2.4.1 (2.4.2 is good) (they contained a bug in base64.b32decode)
708 - avoid python-2.6 on windows with mingw: compiler issues
709 - python2.4 requires pysqlite2 (2.5,2.6 does not)
713 .. _#392: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/392
714 .. _#625: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/625
715 .. _#653: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/653
716 .. _#683: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/683
717 .. _#694: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/694
718 .. _#705: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/705
719 .. _#722: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/722
720 .. _#750: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/750
721 .. _#752: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/752
723 Release 1.4.1 (2009-04-13)
724 --------------------------
729 - The big feature for this release is the implementation of garbage
730 collection, allowing Tahoe storage servers to delete shares for old
731 deleted files. When enabled, this uses a "mark and sweep" process:
732 clients are responsible for updating the leases on their shares
733 (generally by running "tahoe deep-check --add-lease"), and servers
734 are allowed to delete any share which does not have an up-to-date
735 lease. The process is described in detail in
736 `garbage-collection.rst`_.
738 The server must be configured to enable garbage-collection, by
739 adding directives to the [storage] section that define an age limit
740 for shares. The default configuration will not delete any shares.
742 Both servers and clients should be upgraded to this release to make
743 the garbage-collection as pleasant as possible. 1.2.0 servers have
744 code to perform the update-lease operation but it suffers from a
745 fatal bug, while 1.3.0 servers have update-lease but will return an
746 exception for unknown storage indices, causing clients to emit an
747 Incident for each exception, slowing the add-lease process down to a
748 crawl. 1.1.0 servers did not have the add-lease operation at all.
750 Security/Usability Problems Fixed
751 '''''''''''''''''''''''''''''''''
753 - A super-linear algorithm in the Merkle Tree code was fixed, which
754 previously caused e.g. download of a 10GB file to take several hours
755 before the first byte of plaintext could be produced. The new
756 "alacrity" is about 2 minutes. A future release should reduce this
757 to a few seconds by fixing ticket `#442`_.
759 - The previous version permitted a small timing attack (due to our use
760 of strcmp) against the write-enabler and lease-renewal/cancel
761 secrets. An attacker who could measure response-time variations of
762 approximatly 3ns against a very noisy background time of about 15ms
763 might be able to guess these secrets. We do not believe this attack
764 was actually feasible. This release closes the attack by first
765 hashing the two strings to be compared with a random secret.
770 - In most cases, HTML tracebacks will only be sent if an "Accept:
771 text/html" header was provided with the HTTP request. This will
772 generally cause browsers to get an HTMLized traceback but send
773 regular text/plain tracebacks to non-browsers (like the CLI
774 clients). More errors have been mapped to useful HTTP error codes.
776 - The streaming webapi operations (deep-check and manifest) now have a
777 way to indicate errors (an output line that starts with "ERROR"
778 instead of being legal JSON). See `webapi.rst`_ for
781 - The storage server now has its own status page (at /storage), linked
782 from the Welcome page. This page shows progress and results of the
783 two new share-crawlers: one which merely counts shares (to give an
784 estimate of how many files/directories are being stored in the
785 grid), the other examines leases and reports how much space would be
786 freed if GC were enabled. The page also shows how much disk space is
787 present, used, reserved, and available for the Tahoe server, and
788 whether the server is currently running in "read-write" mode or
791 - When a directory node cannot be read (perhaps because of insufficent
792 shares), a minimal webapi page is created so that the "more-info"
793 links (including a Check/Repair operation) will still be accessible.
795 - A new "reliability" page was added, with the beginnings of work on a
796 statistical loss model. You can tell this page how many servers you
797 are using and their independent failure probabilities, and it will
798 tell you the likelihood that an arbitrary file will survive each
799 repair period. The "numpy" package must be installed to access this
800 page. A partial paper, written by Shawn Willden, has been added to
801 docs/proposed/lossmodel.lyx .
806 - "tahoe check" and "tahoe deep-check" now accept an "--add-lease"
807 argument, to update a lease on all shares. This is the "mark" side
808 of garbage collection.
810 - In many cases, CLI error messages have been improved: the ugly
811 HTMLized traceback has been replaced by a normal python traceback.
813 - "tahoe deep-check" and "tahoe manifest" now have better error
814 reporting. "tahoe cp" is now non-verbose by default.
816 - "tahoe backup" now accepts several "--exclude" arguments, to ignore
817 certain files (like editor temporary files and version-control
818 metadata) during backup.
820 - On windows, the CLI now accepts local paths like "c:\dir\file.txt",
821 which previously was interpreted as a Tahoe path using a "c:" alias.
823 - The "tahoe restart" command now uses "--force" by default (meaning
824 it will start a node even if it didn't look like there was one
827 - The "tahoe debug consolidate" command was added. This takes a series
828 of independent timestamped snapshot directories (such as those
829 created by the allmydata.com windows backup program, or a series of
830 "tahoe cp -r" commands) and creates new snapshots that used shared
831 read-only directories whenever possible (like the output of "tahoe
832 backup"). In the most common case (when the snapshots are fairly
833 similar), the result will use significantly fewer directories than
834 the original, allowing "deep-check" and similar tools to run much
835 faster. In some cases, the speedup can be an order of magnitude or
836 more. This tool is still somewhat experimental, and only needs to
837 be run on large backups produced by something other than "tahoe
838 backup", so it was placed under the "debug" category.
840 - "tahoe cp -r --caps-only tahoe:dir localdir" is a diagnostic tool
841 which, instead of copying the full contents of files into the local
842 directory, merely copies their filecaps. This can be used to verify
843 the results of a "consolidation" operation.
848 - The codebase no longer rauses RuntimeError as a kind of
849 assert(). Specific exception classes were created for each previous
850 instance of RuntimeError.
852 -Many unit tests were changed to use a non-network test harness,
853 speeding them up considerably.
855 - Deep-traversal operations (manifest and deep-check) now walk
856 individual directories in alphabetical order. Occasional turn breaks
857 are inserted to prevent a stack overflow when traversing directories
858 with hundreds of entries.
860 - The experimental SFTP server had its path-handling logic changed
861 slightly, to accomodate more SFTP clients, although there are still
864 .. _#442: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/442
865 .. _#645: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/645
866 .. _garbage-collection.rst: docs/garbage-collection.rst
868 Release 1.3.0 (2009-02-13)
869 --------------------------
871 Checker/Verifier/Repairer
872 '''''''''''''''''''''''''
874 - The primary focus of this release has been writing a checker /
875 verifier / repairer for files and directories. "Checking" is the
876 act of asking storage servers whether they have a share for the
877 given file or directory: if there are not enough shares available,
878 the file or directory will be unrecoverable. "Verifying" is the act
879 of downloading and cryptographically asserting that the server's
880 share is undamaged: it requires more work (bandwidth and CPU) than
881 checking, but can catch problems that simple checking
882 cannot. "Repair" is the act of replacing missing or damaged shares
885 - This release includes a full checker, a partial verifier, and a
886 partial repairer. The repairer is able to handle missing shares: new
887 shares are generated and uploaded to make up for the missing
888 ones. This is currently the best application of the repairer: to
889 replace shares that were lost because of server departure or
890 permanent drive failure.
892 - The repairer in this release is somewhat able to handle corrupted
893 shares. The limitations are:
895 - Immutable verifier is incomplete: not all shares are used, and not
896 all fields of those shares are verified. Therefore the immutable
897 verifier has only a moderate chance of detecting corrupted shares.
898 - The mutable verifier is mostly complete: all shares are examined,
899 and most fields of the shares are validated.
900 - The storage server protocol offers no way for the repairer to
901 replace or delete immutable shares. If corruption is detected, the
902 repairer will upload replacement shares to other servers, but the
903 corrupted shares will be left in place.
904 - read-only directories and read-only mutable files must be repaired
905 by someone who holds the write-cap: the read-cap is
906 insufficient. Moreover, the deep-check-and-repair operation will
907 halt with an error if it attempts to repair one of these read-only
909 - Some forms of corruption can cause both download and repair
910 operations to fail. A future release will fix this, since download
911 should be tolerant of any corruption as long as there are at least
912 'k' valid shares, and repair should be able to fix any file that is
915 - If the downloader, verifier, or repairer detects share corruption,
916 the servers which provided the bad shares will be notified (via a
917 file placed in the BASEDIR/storage/corruption-advisories directory)
918 so their operators can manually delete the corrupted shares and
919 investigate the problem. In addition, the "incident gatherer"
920 mechanism will automatically report share corruption to an incident
921 gatherer service, if one is configured. Note that corrupted shares
922 indicate hardware failures, serious software bugs, or malice on the
923 part of the storage server operator, so a corrupted share should be
924 considered highly unusual.
926 - By periodically checking/repairing all files and directories,
927 objects in the Tahoe filesystem remain resistant to recoverability
928 failures due to missing and/or broken servers.
930 - This release includes a wapi mechanism to initiate checks on
931 individual files and directories (with or without verification, and
932 with or without automatic repair). A related mechanism is used to
933 initiate a "deep-check" on a directory: recursively traversing the
934 directory and its children, checking (and/or verifying/repairing)
935 everything underneath. Both mechanisms can be run with an
936 "output=JSON" argument, to obtain machine-readable check/repair
937 status results. These results include a copy of the filesystem
938 statistics from the "deep-stats" operation (including total number
939 of files, size histogram, etc). If repair is possible, a "Repair"
940 button will appear on the results page.
942 - The client web interface now features some extra buttons to initiate
943 check and deep-check operations. When these operations finish, they
944 display a results page that summarizes any problems that were
945 encountered. All long-running deep-traversal operations, including
946 deep-check, use a start-and-poll mechanism, to avoid depending upon
947 a single long-lived HTTP connection. `webapi.rst`_ has
953 - The "tahoe backup" command is new in this release, which creates
954 efficient versioned backups of a local directory. Given a local
955 pathname and a target Tahoe directory, this will create a read-only
956 snapshot of the local directory in $target/Archives/$timestamp. It
957 will also create $target/Latest, which is a reference to the latest
958 such snapshot. Each time you run "tahoe backup" with the same source
959 and target, a new $timestamp snapshot will be added. These snapshots
960 will share directories that have not changed since the last backup,
961 to speed up the process and minimize storage requirements. In
962 addition, a small database is used to keep track of which local
963 files have been uploaded already, to avoid uploading them a second
964 time. This drastically reduces the work needed to do a "null backup"
965 (when nothing has changed locally), making "tahoe backup' suitable
966 to run from a daily cronjob.
968 Note that the "tahoe backup" CLI command must be used in conjunction
969 with a 1.3.0-or-newer Tahoe client node; there was a bug in the
970 1.2.0 webapi implementation that would prevent the last step (create
971 $target/Latest) from working.
976 - The 12GiB (approximate) immutable-file-size limitation is
977 lifted. This release knows how to handle so-called "v2 immutable
978 shares", which permit immutable files of up to about 18 EiB (about
979 3*10^14). These v2 shares are created if the file to be uploaded is
980 too large to fit into v1 shares. v1 shares are created if the file
981 is small enough to fit into them, so that files created with
982 tahoe-1.3.0 can still be read by earlier versions if they are not
983 too large. Note that storage servers also had to be changed to
984 support larger files, and this release is the first release in which
985 they are able to do that. Clients will detect which servers are
986 capable of supporting large files on upload and will not attempt to
987 upload shares of a large file to a server which doesn't support it.
992 - Tahoe now includes experimental FTP and SFTP servers. When
993 configured with a suitable method to translate username+password
994 into a root directory cap, it provides simple access to the virtual
995 filesystem. Remember that FTP is completely unencrypted: passwords,
996 filenames, and file contents are all sent over the wire in
997 cleartext, so FTP should only be used on a local (127.0.0.1)
998 connection. This feature is still in development: there are no unit
999 tests yet, and behavior with respect to Unicode filenames is
1000 uncertain. Please see `FTP-and-SFTP.rst`_ for
1001 configuration details. (`#512`_, `#531`_)
1006 - This release adds the 'tahoe create-alias' command, which is a
1007 combination of 'tahoe mkdir' and 'tahoe add-alias'. This also allows
1008 you to start using a new tahoe directory without exposing its URI in
1009 the argv list, which is publicly visible (through the process table)
1010 on most unix systems. Thanks to Kevin Reid for bringing this issue
1013 - The single-argument form of "tahoe put" was changed to create an
1014 unlinked file. I.e. "tahoe put bar.txt" will take the contents of a
1015 local "bar.txt" file, upload them to the grid, and print the
1016 resulting read-cap; the file will not be attached to any
1017 directories. This seemed a bit more useful than the previous
1018 behavior (copy stdin, upload to the grid, attach the resulting file
1019 into your default tahoe: alias in a child named 'bar.txt').
1021 - "tahoe put" was also fixed to handle mutable files correctly: "tahoe
1022 put bar.txt URI:SSK:..." will read the contents of the local bar.txt
1023 and use them to replace the contents of the given mutable file.
1025 - The "tahoe webopen" command was modified to accept aliases. This
1026 means "tahoe webopen tahoe:" will cause your web browser to open to
1027 a "wui" page that gives access to the directory associated with the
1028 default "tahoe:" alias. It should also accept leading slashes, like
1029 "tahoe webopen tahoe:/stuff".
1031 - Many esoteric debugging commands were moved down into a "debug"
1034 - tahoe debug dump-cap
1035 - tahoe debug dump-share
1036 - tahoe debug find-shares
1037 - tahoe debug catalog-shares
1038 - tahoe debug corrupt-share
1040 The last command ("tahoe debug corrupt-share") flips a random bit
1041 of the given local sharefile. This is used to test the file
1042 verifying/repairing code, and obviously should not be used on user
1045 The cli might not correctly handle arguments which contain non-ascii
1046 characters in Tahoe v1.3 (although depending on your platform it
1047 might, especially if your platform can be configured to pass such
1048 characters on the command-line in utf-8 encoding). See
1049 http://tahoe-lafs.org/trac/tahoe/ticket/565 for details.
1054 - The "default webapi port", used when creating a new client node (and
1055 in the getting-started documentation), was changed from 8123 to
1056 3456, to reduce confusion when Tahoe accessed through a Firefox
1057 browser on which the "Torbutton" extension has been installed. Port
1058 8123 is occasionally used as a Tor control port, so Torbutton adds
1059 8123 to Firefox's list of "banned ports" to avoid CSRF attacks
1060 against Tor. Once 8123 is banned, it is difficult to diagnose why
1061 you can no longer reach a Tahoe node, so the Tahoe default was
1062 changed. Note that 3456 is reserved by IANA for the "vat" protocol,
1063 but there are argueably more Torbutton+Tahoe users than vat users
1064 these days. Note that this will only affect newly-created client
1065 nodes. Pre-existing client nodes, created by earlier versions of
1066 tahoe, may still be listening on 8123.
1068 - All deep-traversal operations (start-manifest, start-deep-size,
1069 start-deep-stats, start-deep-check) now use a start-and-poll
1070 approach, instead of using a single (fragile) long-running
1071 synchronous HTTP connection. All these "start-" operations use POST
1072 instead of GET. The old "GET manifest", "GET deep-size", and "POST
1073 deep-check" operations have been removed.
1075 - The new "POST start-manifest" operation, when it finally completes,
1076 results in a table of (path,cap), instead of the list of verifycaps
1077 produced by the old "GET manifest". The table is available in
1078 several formats: use output=html, output=text, or output=json to
1079 choose one. The JSON output also includes stats, and a list of
1080 verifycaps and storage-index strings. The "return_to=" and
1081 "when_done=" arguments have been removed from the t=check and
1082 deep-check operations.
1084 - The top-level status page (/status) now has a machine-readable form,
1085 via "/status/?t=json". This includes information about the
1086 currently-active uploads and downloads, which may be useful for
1087 frontends that wish to display progress information. There is no
1088 easy way to correlate the activities displayed here with recent wapi
1091 - Any files in BASEDIR/public_html/ (configurable) will be served in
1092 response to requests in the /static/ portion of the URL space. This
1093 will simplify the deployment of javascript-based frontends that can
1094 still access wapi calls by conforming to the (regrettable)
1095 "same-origin policy".
1097 - The welcome page now has a "Report Incident" button, which is tied
1098 into the "Incident Gatherer" machinery. If the node is attached to
1099 an incident gatherer (via log_gatherer.furl), then pushing this
1100 button will cause an Incident to be signalled: this means recent log
1101 events are aggregated and sent in a bundle to the gatherer. The user
1102 can push this button after something strange takes place (and they
1103 can provide a short message to go along with it), and the relevant
1104 data will be delivered to a centralized incident-gatherer for later
1105 processing by operations staff.
1107 - The "HEAD" method should now work correctly, in addition to the
1108 usual "GET", "PUT", and "POST" methods. "HEAD" is supposed to return
1109 exactly the same headers as "GET" would, but without any of the
1110 actual response body data. For mutable files, this now does a brief
1111 mapupdate (to figure out the size of the file that would be
1112 returned), without actually retrieving the file's contents.
1114 - The "GET" operation on files can now support the HTTP "Range:"
1115 header, allowing requests for partial content. This allows certain
1116 media players to correctly stream audio and movies out of a Tahoe
1117 grid. The current implementation uses a disk-based cache in
1118 BASEDIR/private/cache/download , which holds the plaintext of the
1119 files being downloaded. Future implementations might not use this
1120 cache. GET for immutable files now returns an ETag header.
1122 - Each file and directory now has a "Show More Info" web page, which
1123 contains much of the information that was crammed into the directory
1124 page before. This includes readonly URIs, storage index strings,
1125 object type, buttons to control checking/verifying/repairing, and
1126 deep-check/deep-stats buttons (for directories). For mutable files,
1127 the "replace contents" upload form has been moved here too. As a
1128 result, the directory page is now much simpler and cleaner, and
1129 several potentially-misleading links (like t=uri) are now gone.
1131 - Slashes are discouraged in Tahoe file/directory names, since they
1132 cause problems when accessing the filesystem through the
1133 wapi. However, there are a couple of accidental ways to generate
1134 such names. This release tries to make it easier to correct such
1135 mistakes by escaping slashes in several places, allowing slashes in
1136 the t=info and t=delete commands, and in the source (but not the
1137 target) of a t=rename command.
1142 - Tahoe's dependencies have been extended to require the
1143 "[secure_connections]" feature from Foolscap, which will cause
1144 pyOpenSSL to be required and/or installed. If OpenSSL and its
1145 development headers are already installed on your system, this can
1146 occur automatically. Tahoe now uses pollreactor (instead of the
1147 default selectreactor) to work around a bug between pyOpenSSL and
1148 the most recent release of Twisted (8.1.0). This bug only affects
1149 unit tests (hang during shutdown), and should not impact regular
1152 - The Tahoe source code tarballs now come in two different forms:
1153 regular and "sumo". The regular tarball contains just Tahoe, nothing
1154 else. When building from the regular tarball, the build process will
1155 download any unmet dependencies from the internet (starting with the
1156 index at PyPI) so it can build and install them. The "sumo" tarball
1157 contains copies of all the libraries that Tahoe requires (foolscap,
1158 twisted, zfec, etc), so using the "sumo" tarball should not require
1159 any internet access during the build process. This can be useful if
1160 you want to build Tahoe while on an airplane, a desert island, or
1161 other bandwidth-limited environments.
1163 - Similarly, tahoe-lafs.org now hosts a "tahoe-deps" tarball which
1164 contains the latest versions of all these dependencies. This
1166 http://tahoe-lafs.org/source/tahoe/deps/tahoe-deps.tar.gz, can be
1167 unpacked in the tahoe source tree (or in its parent directory), and
1168 the build process should satisfy its downloading needs from it
1169 instead of reaching out to PyPI. This can be useful if you want to
1170 build Tahoe from a darcs checkout while on that airplane or desert
1173 - Because of the previous two changes ("sumo" tarballs and the
1174 "tahoe-deps" bundle), most of the files have been removed from
1175 misc/dependencies/ . This brings the regular Tahoe tarball down to
1176 2MB (compressed), and the darcs checkout (without history) to about
1177 7.6MB. A full darcs checkout will still be fairly large (because of
1178 the historical patches which included the dependent libraries), but
1179 a 'lazy' one should now be small.
1181 - The default "make" target is now an alias for "setup.py build",
1182 which itself is an alias for "setup.py develop --prefix support",
1183 with some extra work before and after (see setup.cfg). Most of the
1184 complicated platform-dependent code in the Makefile was rewritten in
1185 Python and moved into setup.py, simplifying things considerably.
1187 - Likewise, the "make test" target now delegates most of its work to
1188 "setup.py test", which takes care of getting PYTHONPATH configured
1189 to access the tahoe code (and dependencies) that gets put in
1190 support/lib/ by the build_tahoe step. This should allow unit tests
1191 to be run even when trial (which is part of Twisted) wasn't already
1192 installed (in this case, trial gets installed to support/bin because
1193 Twisted is a dependency of Tahoe).
1195 - Tahoe is now compatible with the recently-released Python 2.6 ,
1196 although it is recommended to use Tahoe on Python 2.5, on which it
1197 has received more thorough testing and deployment.
1199 - Tahoe is now compatible with simplejson-2.0.x . The previous release
1200 assumed that simplejson.loads always returned unicode strings, which
1201 is no longer the case in 2.0.x .
1203 Grid Management Tools
1204 '''''''''''''''''''''
1206 - Several tools have been added or updated in the misc/ directory,
1207 mostly munin plugins that can be used to monitor a storage grid.
1209 - The misc/spacetime/ directory contains a "disk watcher" daemon
1210 (startable with 'tahoe start'), which can be configured with a set
1211 of HTTP URLs (pointing at the wapi '/statistics' page of a bunch of
1212 storage servers), and will periodically fetch
1213 disk-used/disk-available information from all the servers. It keeps
1214 this information in an Axiom database (a sqlite-based library
1215 available from divmod.org). The daemon computes time-averaged rates
1216 of disk usage, as well as a prediction of how much time is left
1217 before the grid is completely full.
1219 - The misc/munin/ directory contains a new set of munin plugins
1220 (tahoe_diskleft, tahoe_diskusage, tahoe_doomsday) which talk to the
1221 disk-watcher and provide graphs of its calculations.
1223 - To support the disk-watcher, the Tahoe statistics component
1224 (visible through the wapi at the /statistics/ URL) now includes
1225 disk-used and disk-available information. Both are derived through
1226 an equivalent of the unix 'df' command (i.e. they ask the kernel
1227 for the number of free blocks on the partition that encloses the
1228 BASEDIR/storage directory). In the future, the disk-available
1229 number will be further influenced by the local storage policy: if
1230 that policy says that the server should refuse new shares when less
1231 than 5GB is left on the partition, then "disk-available" will
1232 report zero even though the kernel sees 5GB remaining.
1234 - The 'tahoe_overhead' munin plugin interacts with an
1235 allmydata.com-specific server which reports the total of the
1236 'deep-size' reports for all active user accounts, compares this
1237 with the disk-watcher data, to report on overhead percentages. This
1238 provides information on how much space could be recovered once
1239 Tahoe implements some form of garbage collection.
1241 Configuration Changes: single INI-format tahoe.cfg file
1242 '''''''''''''''''''''''''''''''''''''''''''''''''''''''
1244 - The Tahoe node is now configured with a single INI-format file,
1245 named "tahoe.cfg", in the node's base directory. Most of the
1246 previous multiple-separate-files are still read for backwards
1247 compatibility (the embedded SSH debug server and the
1248 advertised_ip_addresses files are the exceptions), but new
1249 directives will only be added to tahoe.cfg . The "tahoe
1250 create-client" command will create a tahoe.cfg for you, with sample
1251 values commented out. (ticket `#518`_)
1253 - tahoe.cfg now has controls for the foolscap "keepalive" and
1254 "disconnect" timeouts (`#521`_).
1256 - tahoe.cfg now has controls for the encoding parameters:
1257 "shares.needed" and "shares.total" in the "[client]" section. The
1258 default parameters are still 3-of-10.
1260 - The inefficient storage 'sizelimit' control (which established an
1261 upper bound on the amount of space that a storage server is allowed
1262 to consume) has been replaced by a lightweight 'reserved_space'
1263 control (which establishes a lower bound on the amount of remaining
1264 space). The storage server will reject all writes that would cause
1265 the remaining disk space (as measured by a '/bin/df' equivalent) to
1266 drop below this value. The "[storage]reserved_space=" tahoe.cfg
1267 parameter controls this setting. (note that this only affects
1268 immutable shares: it is an outstanding bug that reserved_space does
1269 not prevent the allocation of new mutable shares, nor does it
1270 prevent the growth of existing mutable shares).
1275 - Clients now declare which versions of the protocols they
1276 support. This is part of a new backwards-compatibility system:
1277 http://tahoe-lafs.org/trac/tahoe/wiki/Versioning .
1279 - The version strings for human inspection (as displayed on the
1280 Welcome web page, and included in logs) now includes a platform
1281 identifer (frequently including a linux distribution name, processor
1284 - Several bugs have been fixed, including one that would cause an
1285 exception (in the logs) if a wapi download operation was cancelled
1286 (by closing the TCP connection, or pushing the "stop" button in a
1289 - Tahoe now uses Foolscap "Incidents", writing an "incident report"
1290 file to logs/incidents/ each time something weird occurs. These
1291 reports are available to an "incident gatherer" through the flogtool
1292 command. For more details, please see the Foolscap logging
1293 documentation. An incident-classifying plugin function is provided
1294 in misc/incident-gatherer/classify_tahoe.py .
1296 - If clients detect corruption in shares, they now automatically
1297 report it to the server holding that share, if it is new enough to
1298 accept the report. These reports are written to files in
1299 BASEDIR/storage/corruption-advisories .
1301 - The 'nickname' setting is now defined to be a UTF-8 -encoded string,
1302 allowing non-ascii nicknames.
1304 - The 'tahoe start' command will now accept a --syslog argument and
1305 pass it through to twistd, making it easier to launch non-Tahoe
1306 nodes (like the cpu-watcher) and have them log to syslogd instead of
1307 a local file. This is useful when running a Tahoe node out of a USB
1310 - The Mac GUI in src/allmydata/gui/ has been improved.
1312 .. _#512: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/512
1313 .. _#518: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/518
1314 .. _#521: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/521
1315 .. _#531: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/531
1317 Release 1.2.0 (2008-07-21)
1318 --------------------------
1323 - This release makes the immutable-file "ciphertext hash tree"
1324 mandatory. Previous releases allowed the uploader to decide whether
1325 their file would have an integrity check on the ciphertext or not. A
1326 malicious uploader could use this to create a readcap that would
1327 download as one file or a different one, depending upon which shares
1328 the client fetched first, with no errors raised. There are other
1329 integrity checks on the shares themselves, preventing a storage
1330 server or other party from violating the integrity properties of the
1331 read-cap: this failure was only exploitable by the uploader who
1332 gives you a carefully constructed read-cap. If you download the file
1333 with Tahoe 1.2.0 or later, you will not be vulnerable to this
1336 This change does not introduce a compatibility issue, because all
1337 existing versions of Tahoe will emit the ciphertext hash tree in
1343 - Tahoe now requires Foolscap-0.2.9 . It also requires pycryptopp 0.5
1344 or newer, since earlier versions had a bug that interacted with
1345 specific compiler versions that could sometimes result in incorrect
1346 encryption behavior. Both packages are included in the Tahoe source
1347 tarball in misc/dependencies/ , and should be built automatically
1353 - Web API directory pages should now contain properly-slash-terminated
1354 links to other directories. They have also stopped using absolute
1355 links in forms and pages (which interfered with the use of a
1356 front-end load-balancing proxy).
1358 - The behavior of the "Check This File" button changed, in conjunction
1359 with larger internal changes to file checking/verification. The
1360 button triggers an immediate check as before, but the outcome is
1361 shown on its own page, and does not get stored anywhere. As a
1362 result, the web directory page no longer shows historical checker
1365 - A new "Deep-Check" button has been added, which allows a user to
1366 initiate a recursive check of the given directory and all files and
1367 directories reachable from it. This can cause quite a bit of work,
1368 and has no intermediate progress information or feedback about the
1369 process. In addition, the results of the deep-check are extremely
1370 limited. A later release will improve this behavior.
1372 - The web server's behavior with respect to non-ASCII (unicode)
1373 filenames in the "GET save=true" operation has been improved. To
1374 achieve maximum compatibility with variously buggy web browsers, the
1375 server does not try to figure out the character set of the inbound
1376 filename. It just echoes the same bytes back to the browser in the
1377 Content-Disposition header. This seems to make both IE7 and Firefox
1380 Checker/Verifier/Repairer
1381 '''''''''''''''''''''''''
1383 - Tahoe is slowly acquiring convenient tools to check up on file
1384 health, examine existing shares for errors, and repair files that
1385 are not fully healthy. This release adds a mutable
1386 checker/verifier/repairer, although testing is very limited, and
1387 there are no web interfaces to trigger repair yet. The "Check"
1388 button next to each file or directory on the wapi page will perform
1389 a file check, and the "deep check" button on each directory will
1390 recursively check all files and directories reachable from there
1391 (which may take a very long time).
1393 Future releases will improve access to this functionality.
1395 Operations/Packaging
1396 ''''''''''''''''''''
1398 - A "check-grid" script has been added, along with a Makefile
1399 target. This is intended (with the help of a pre-configured node
1400 directory) to check upon the health of a Tahoe grid, uploading and
1401 downloading a few files. This can be used as a monitoring tool for a
1402 deployed grid, to be run periodically and to signal an error if it
1403 ever fails. It also helps with compatibility testing, to verify that
1404 the latest Tahoe code is still able to handle files created by an
1407 - The munin plugins from misc/munin/ are now copied into any generated
1408 debian packages, and are made executable (and uncompressed) so they
1409 can be symlinked directly from /etc/munin/plugins/ .
1411 - Ubuntu "Hardy" was added as a supported debian platform, with a
1412 Makefile target to produce hardy .deb packages. Some notes have been
1413 added to `debian.rst`_ about building Tahoe on a debian/ubuntu
1416 - Storage servers now measure operation rates and
1417 latency-per-operation, and provides results through the /statistics
1418 web page as well as the stats gatherer. Munin plugins have been
1424 - Tahoe nodes now use Foolscap "incident logging" to record unusual
1425 events to their NODEDIR/logs/incidents/ directory. These incident
1426 files can be examined by Foolscap logging tools, or delivered to an
1427 external log-gatherer for further analysis. Note that Tahoe now
1428 requires Foolscap-0.2.9, since 0.2.8 had a bug that complained about
1429 "OSError: File exists" when trying to create the incidents/
1430 directory for a second time.
1432 - If no servers are available when retrieving a mutable file (like a
1433 directory), the node now reports an error instead of hanging
1434 forever. Earlier releases would not only hang (causing the wapi
1435 directory listing to get stuck half-way through), but the internal
1436 dirnode serialization would cause all subsequent attempts to
1437 retrieve or modify the same directory to hang as well. `#463`_
1439 - A minor internal exception (reported in logs/twistd.log, in the
1440 "stopProducing" method) was fixed, which complained about
1441 "self._paused_at not defined" whenever a file download was stopped
1442 from the web browser end.
1444 .. _#463: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/463
1445 .. _#491: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/491
1446 .. _debian.rst: docs/debian.rst
1448 Release 1.1.0 (2008-06-11)
1449 --------------------------
1451 CLI: new "alias" model
1452 ''''''''''''''''''''''
1454 - The new CLI code uses an scp/rsync -like interface, in which
1455 directories in the Tahoe storage grid are referenced by a
1456 colon-suffixed alias. The new commands look like:
1458 - tahoe cp local.txt tahoe:virtual.txt
1459 - tahoe ls work:subdir
1461 - More functionality is available through the CLI: creating unlinked
1462 files and directories, recursive copy in or out of the storage grid,
1463 hardlinks, and retrieving the raw read- or write- caps through the
1464 'ls' command. Please read `CLI.rst`_ for complete details.
1466 wapi: new pages, new commands
1467 '''''''''''''''''''''''''''''
1469 - Several new pages were added to the web API:
1471 - /helper_status : to describe what a Helper is doing
1472 - /statistics : reports node uptime, CPU usage, other stats
1473 - /file : for easy file-download URLs, see `#221`_
1474 - /cap == /uri : future compatibility
1476 - The localdir=/localfile= and t=download operations were
1477 removed. These required special configuration to enable anyways, but
1478 this feature was a security problem, and was mostly obviated by the
1479 new "cp -r" command.
1481 - Several new options to the GET command were added:
1483 - t=deep-size : add up the size of all immutable files reachable from the directory
1484 - t=deep-stats : return a JSON-encoded description of number of files, size distribution, total size, etc
1486 - POST is now preferred over PUT for most operations which cause
1489 - Most wapi calls now accept overwrite=, and default to overwrite=true
1491 - "POST /uri/DIRCAP/parent/child?t=mkdir" is now the preferred API to
1492 create multiple directories at once, rather than ...?t=mkdir-p .
1494 - PUT to a mutable file ("PUT /uri/MUTABLEFILECAP", "PUT
1495 /uri/DIRCAP/child") will modify the file in-place.
1497 - more munin graphs in misc/munin/
1500 - tahoe-rootdir-space
1501 - tahoe_estimate_files
1502 - mutable files published/retrieved
1511 - setuptools (now required at runtime)
1513 New Mutable-File Code
1514 '''''''''''''''''''''
1516 - The mutable-file handling code (mostly used for directories) has
1517 been completely rewritten. The new scheme has a better API (with a
1518 modify() method) and is less likely to lose data when several
1519 uncoordinated writers change a file at the same time.
1521 - In addition, a single Tahoe process will coordinate its own
1522 writes. If you make two concurrent directory-modifying wapi calls to
1523 a single tahoe node, it will internally make one of them wait for
1524 the other to complete. This prevents auto-collision (`#391`_).
1526 - The new mutable-file code also detects errors during publish
1527 better. Earlier releases might believe that a mutable file was
1528 published when in fact it failed.
1533 - The node now monitors its own CPU usage, as a percentage, measured
1534 every 60 seconds. 1/5/15 minute moving averages are available on the
1535 /statistics web page and via the stats-gathering interface.
1537 - Clients now accelerate reconnection to all servers after being
1538 offline (`#374`_). When a client is offline for a long time, it
1539 scales back reconnection attempts to approximately once per hour, so
1540 it may take a while to make the first attempt, but once any attempt
1541 succeeds, the other server connections will be retried immediately.
1543 - A new "offloaded KeyGenerator" facility can be configured, to move
1544 RSA key generation out from, say, a wapi node, into a separate
1545 process. RSA keys can take several seconds to create, and so a wapi
1546 node which is being used for directory creation will be unavailable
1547 for anything else during this time. The Key Generator process will
1548 pre-compute a small pool of keys, to speed things up further. This
1549 also takes better advantage of multi-core CPUs, or SMP hosts.
1551 - The node will only use a potentially-slow "du -s" command at startup
1552 (to measure how much space has been used) if the "sizelimit"
1553 parameter has been configured (to limit how much space is
1554 used). Large storage servers should turn off sizelimit until a later
1555 release improves the space-management code, since "du -s" on a
1556 terabyte filesystem can take hours.
1558 - The Introducer now allows new announcements to replace old ones, to
1559 avoid buildups of obsolete announcements.
1561 - Immutable files are limited to about 12GiB (when using the default
1562 3-of-10 encoding), because larger files would be corrupted by the
1563 four-byte share-size field on the storage servers (`#439`_). A later
1564 release will remove this limit. Earlier releases would allow >12GiB
1565 uploads, but the resulting file would be unretrievable.
1567 - The docs/ directory has been rearranged, with old docs put in
1568 docs/historical/ and not-yet-implemented ones in docs/proposed/ .
1570 - The Mac OS-X FUSE plugin has a significant bug fix: earlier versions
1571 would corrupt writes that used seek() instead of writing the file in
1572 linear order. The rsync tool is known to perform writes in this
1573 order. This has been fixed.
1575 .. _#221: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/221
1576 .. _#374: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/374
1577 .. _#391: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/391
1578 .. _#439: http://tahoe-lafs.org/trac/tahoe-lafs/ticket/439
1579 .. _CLI.rst: docs/CLI.rst