1 ==========================
2 The Tahoe REST-ful Web API
3 ==========================
5 1. `Enabling the web-API port`_
6 2. `Basic Concepts: GET, PUT, DELETE, POST`_
11 4. `Slow Operations, Progress, and Cancelling`_
12 5. `Programmatic Operations`_
15 2. `Writing/Uploading a File`_
16 3. `Creating a New Directory`_
17 4. `Getting Information About A File Or Directory (as JSON)`_
18 5. `Attaching an Existing File or Directory by its read- or write-cap`_
19 6. `Adding Multiple Files or Directories to a Parent Directory at Once`_
20 7. `Unlinking a File or Directory`_
22 6. `Browser Operations: Human-Oriented Interfaces`_
24 1. `Viewing A Directory (as HTML)`_
25 2. `Viewing/Downloading a File`_
26 3. `Getting Information About A File Or Directory (as HTML)`_
27 4. `Creating a Directory`_
28 5. `Uploading a File`_
29 6. `Attaching An Existing File Or Directory (by URI)`_
30 7. `Unlinking A Child`_
31 8. `Renaming A Child`_
33 10. `Debugging and Testing Features`_
35 7. `Other Useful Pages`_
36 8. `Static Files in /public_html`_
37 9. `Safety and Security Issues -- Names vs. URIs`_
38 10. `Concurrency Issues`_
41 Enabling the web-API port
42 =========================
44 Every Tahoe node is capable of running a built-in HTTP server. To enable
45 this, just write a port number into the "[node]web.port" line of your node's
46 tahoe.cfg file. For example, writing "web.port = 3456" into the "[node]"
47 section of $NODEDIR/tahoe.cfg will cause the node to run a webserver on port
50 This string is actually a Twisted "strports" specification, meaning you can
51 get more control over the interface to which the server binds by supplying
52 additional arguments. For more details, see the documentation on
53 `twisted.application.strports
54 <http://twistedmatrix.com/documents/current/api/twisted.application.strports.html>`_.
56 Writing "tcp:3456:interface=127.0.0.1" into the web.port line does the same
57 but binds to the loopback interface, ensuring that only the programs on the
58 local host can connect. Using "ssl:3456:privateKey=mykey.pem:certKey=cert.pem"
61 This webport can be set when the node is created by passing a --webport
62 option to the 'tahoe create-node' command. By default, the node listens on
63 port 3456, on the loopback (127.0.0.1) interface.
66 Basic Concepts: GET, PUT, DELETE, POST
67 ======================================
69 As described in `docs/architecture.rst <../architecture.rst>`_, each file
70 and directory in a Tahoe virtual filesystem is referenced by an identifier
71 that combines the designation of the object with the authority to do something
72 with it (such as read or modify the contents). This identifier is called a
73 "read-cap" or "write-cap", depending upon whether it enables read-only or
74 read-write access. These "caps" are also referred to as URIs (which may be
75 confusing because they are not currently `RFC3986
76 <http://tools.ietf.org/html/rfc3986>`_-compliant URIs).
78 The Tahoe web-based API is "REST-ful", meaning it implements the concepts of
79 "REpresentational State Transfer": the original scheme by which the World
80 Wide Web was intended to work. Each object (file or directory) is referenced
81 by a URL that includes the read- or write- cap. HTTP methods (GET, PUT, and
82 DELETE) are used to manipulate these objects. You can think of the URL as a
83 noun, and the method as a verb.
85 In REST, the GET method is used to retrieve information about an object, or
86 to retrieve some representation of the object itself. When the object is a
87 file, the basic GET method will simply return the contents of that file.
88 Other variations (generally implemented by adding query parameters to the
89 URL) will return information about the object, such as metadata. GET
90 operations are required to have no side-effects.
92 PUT is used to upload new objects into the filesystem, or to replace an
93 existing link or the contents of a mutable file. DELETE is used to unlink
94 objects from directories. Both PUT and DELETE are required to be idempotent:
95 performing the same operation multiple times must have the same side-effects
96 as only performing it once.
98 POST is used for more complicated actions that cannot be expressed as a GET,
99 PUT, or DELETE. POST operations can be thought of as a method call: sending
100 some message to the object referenced by the URL. In Tahoe, POST is also used
101 for operations that must be triggered by an HTML form (including upload and
102 unlinking), because otherwise a regular web browser has no way to accomplish
103 these tasks. In general, everything that can be done with a PUT or DELETE can
104 also be done with a POST.
106 Tahoe's web API is designed for two different kinds of consumer. The first is
107 a program that needs to manipulate the virtual file system. Such programs are
108 expected to use the RESTful interface described above. The second is a human
109 using a standard web browser to work with the filesystem. This user is given
110 a series of HTML pages with links to download files, and forms that use POST
111 actions to upload, rename, and unlink files.
113 When an error occurs, the HTTP response code will be set to an appropriate
114 400-series code (like 404 Not Found for an unknown childname, or 400 Bad Request
115 when the parameters to a web-API operation are invalid), and the HTTP response
116 body will usually contain a few lines of explanation as to the cause of the
117 error and possible responses. Unusual exceptions may result in a 500 Internal
118 Server Error as a catch-all, with a default response body containing
119 a Nevow-generated HTML-ized representation of the Python exception stack trace
120 that caused the problem. CLI programs which want to copy the response body to
121 stderr should provide an "Accept: text/plain" header to their requests to get
122 a plain text stack trace instead. If the Accept header contains ``*/*``, or
123 ``text/*``, or text/html (or if there is no Accept header), HTML tracebacks will
130 Tahoe uses a variety of read- and write- caps to identify files and
131 directories. The most common of these is the "immutable file read-cap", which
132 is used for most uploaded files. These read-caps look like the following::
134 URI:CHK:ime6pvkaxuetdfah2p2f35pe54:4btz54xk3tew6nd4y2ojpxj4m6wxjqqlwnztgre6gnjgtucd5r4a:3:10:202
136 The next most common is a "directory write-cap", which provides both read and
137 write access to a directory, and look like this::
139 URI:DIR2:djrdkfawoqihigoett4g6auz6a:jx5mplfpwexnoqff7y5e4zjus4lidm76dcuarpct7cckorh2dpgq
141 There are also "directory read-caps", which start with "URI:DIR2-RO:", and
142 give read-only access to a directory. Finally there are also mutable file
143 read- and write- caps, which start with "URI:SSK", and give access to mutable
146 (Later versions of Tahoe will make these strings shorter, and will remove the
147 unfortunate colons, which must be escaped when these caps are embedded in
150 To refer to any Tahoe object through the web API, you simply need to combine
151 a prefix (which indicates the HTTP server to use) with the cap (which
152 indicates which object inside that server to access). Since the default Tahoe
153 webport is 3456, the most common prefix is one that will use a local node
154 listening on this port::
156 http://127.0.0.1:3456/uri/ + $CAP
158 So, to access the directory named above (which happens to be the
159 publically-writeable sample directory on the Tahoe test grid, described at
160 http://allmydata.org/trac/tahoe/wiki/TestGrid), the URL would be::
162 http://127.0.0.1:3456/uri/URI%3ADIR2%3Adjrdkfawoqihigoett4g6auz6a%3Ajx5mplfpwexnoqff7y5e4zjus4lidm76dcuarpct7cckorh2dpgq/
164 (note that the colons in the directory-cap are url-encoded into "%3A"
167 Likewise, to access the file named above, use::
169 http://127.0.0.1:3456/uri/URI%3ACHK%3Aime6pvkaxuetdfah2p2f35pe54%3A4btz54xk3tew6nd4y2ojpxj4m6wxjqqlwnztgre6gnjgtucd5r4a%3A3%3A10%3A202
171 In the rest of this document, we'll use "$DIRCAP" as shorthand for a read-cap
172 or write-cap that refers to a directory, and "$FILECAP" to abbreviate a cap
173 that refers to a file (whether mutable or immutable). So those URLs above can
176 http://127.0.0.1:3456/uri/$DIRCAP/
177 http://127.0.0.1:3456/uri/$FILECAP
179 The operation summaries below will abbreviate these further, by eliding the
180 server prefix. They will be displayed like this::
189 Tahoe directories contain named child entries, just like directories in a regular
190 local filesystem. These child entries, called "dirnodes", consist of a name,
191 metadata, a write slot, and a read slot. The write and read slots normally contain
192 a write-cap and read-cap referring to the same object, which can be either a file
193 or a subdirectory. The write slot may be empty (actually, both may be empty,
194 but that is unusual).
196 If you have a Tahoe URL that refers to a directory, and want to reference a
197 named child inside it, just append the child name to the URL. For example, if
198 our sample directory contains a file named "welcome.txt", we can refer to
201 http://127.0.0.1:3456/uri/$DIRCAP/welcome.txt
203 (or http://127.0.0.1:3456/uri/URI%3ADIR2%3Adjrdkfawoqihigoett4g6auz6a%3Ajx5mplfpwexnoqff7y5e4zjus4lidm76dcuarpct7cckorh2dpgq/welcome.txt)
205 Multiple levels of subdirectories can be handled this way::
207 http://127.0.0.1:3456/uri/$DIRCAP/tahoe-source/docs/architecture.rst
209 In this document, when we need to refer to a URL that references a file using
210 this child-of-some-directory format, we'll use the following string::
212 /uri/$DIRCAP/[SUBDIRS../]FILENAME
214 The "[SUBDIRS../]" part means that there are zero or more (optional)
215 subdirectory names in the middle of the URL. The "FILENAME" at the end means
216 that this whole URL refers to a file of some sort, rather than to a
219 When we need to refer specifically to a directory in this way, we'll write::
221 /uri/$DIRCAP/[SUBDIRS../]SUBDIR
224 Note that all components of pathnames in URLs are required to be UTF-8
225 encoded, so "resume.doc" (with an acute accent on both E's) would be accessed
228 http://127.0.0.1:3456/uri/$DIRCAP/r%C3%A9sum%C3%A9.doc
230 Also note that the filenames inside upload POST forms are interpreted using
231 whatever character set was provided in the conventional '_charset' field, and
232 defaults to UTF-8 if not otherwise specified. The JSON representation of each
233 directory contains native Unicode strings. Tahoe directories are specified to
234 contain Unicode filenames, and cannot contain binary strings that are not
235 representable as such.
237 All Tahoe operations that refer to existing files or directories must include
238 a suitable read- or write- cap in the URL: the web-API server won't add one
239 for you. If you don't know the cap, you can't access the file. This allows
240 the security properties of Tahoe caps to be extended across the web-API
244 Slow Operations, Progress, and Cancelling
245 =========================================
247 Certain operations can be expected to take a long time. The "t=deep-check",
248 described below, will recursively visit every file and directory reachable
249 from a given starting point, which can take minutes or even hours for
250 extremely large directory structures. A single long-running HTTP request is a
251 fragile thing: proxies, NAT boxes, browsers, and users may all grow impatient
252 with waiting and give up on the connection.
254 For this reason, long-running operations have an "operation handle", which
255 can be used to poll for status/progress messages while the operation
256 proceeds. This handle can also be used to cancel the operation. These handles
257 are created by the client, and passed in as a an "ophandle=" query argument
258 to the POST or PUT request which starts the operation. The following
259 operations can then be used to retrieve status:
261 ``GET /operations/$HANDLE?output=HTML (with or without t=status)``
263 ``GET /operations/$HANDLE?output=JSON (same)``
265 These two retrieve the current status of the given operation. Each operation
266 presents a different sort of information, but in general the page retrieved
269 * whether the operation is complete, or if it is still running
270 * how much of the operation is complete, and how much is left, if possible
272 Note that the final status output can be quite large: a deep-manifest of a
273 directory structure with 300k directories and 200k unique files is about
274 275MB of JSON, and might take two minutes to generate. For this reason, the
275 full status is not provided until the operation has completed.
277 The HTML form will include a meta-refresh tag, which will cause a regular
278 web browser to reload the status page about 60 seconds later. This tag will
279 be removed once the operation has completed.
281 There may be more status information available under
282 /operations/$HANDLE/$ETC : i.e., the handle forms the root of a URL space.
284 ``POST /operations/$HANDLE?t=cancel``
286 This terminates the operation, and returns an HTML page explaining what was
287 cancelled. If the operation handle has already expired (see below), this
288 POST will return a 404, which indicates that the operation is no longer
289 running (either it was completed or terminated). The response body will be
290 the same as a GET /operations/$HANDLE on this operation handle, and the
291 handle will be expired immediately afterwards.
293 The operation handle will eventually expire, to avoid consuming an unbounded
294 amount of memory. The handle's time-to-live can be reset at any time, by
295 passing a retain-for= argument (with a count of seconds) to either the
296 initial POST that starts the operation, or the subsequent GET request which
297 asks about the operation. For example, if a 'GET
298 /operations/$HANDLE?output=JSON&retain-for=600' query is performed, the
299 handle will remain active for 600 seconds (10 minutes) after the GET was
302 In addition, if the GET includes a release-after-complete=True argument, and
303 the operation has completed, the operation handle will be released
306 If a retain-for= argument is not used, the default handle lifetimes are:
308 * handles will remain valid at least until their operation finishes
309 * uncollected handles for finished operations (i.e. handles for
310 operations that have finished but for which the GET page has not been
311 accessed since completion) will remain valid for four days, or for
312 the total time consumed by the operation, whichever is greater.
313 * collected handles (i.e. the GET page has been retrieved at least once
314 since the operation completed) will remain valid for one day.
316 Many "slow" operations can begin to use unacceptable amounts of memory when
317 operating on large directory structures. The memory usage increases when the
318 ophandle is polled, as the results must be copied into a JSON string, sent
319 over the wire, then parsed by a client. So, as an alternative, many "slow"
320 operations have streaming equivalents. These equivalents do not use operation
321 handles. Instead, they emit line-oriented status results immediately. Client
322 code can cancel the operation by simply closing the HTTP connection.
325 Programmatic Operations
326 =======================
328 Now that we know how to build URLs that refer to files and directories in a
329 Tahoe virtual filesystem, what sorts of operations can we do with those URLs?
330 This section contains a catalog of GET, PUT, DELETE, and POST operations that
331 can be performed on these URLs. This set of operations are aimed at programs
332 that use HTTP to communicate with a Tahoe node. A later section describes
333 operations that are intended for web browsers.
339 ``GET /uri/$FILECAP``
341 ``GET /uri/$DIRCAP/[SUBDIRS../]FILENAME``
343 This will retrieve the contents of the given file. The HTTP response body
344 will contain the sequence of bytes that make up the file.
346 To view files in a web browser, you may want more control over the
347 Content-Type and Content-Disposition headers. Please see the next section
348 "Browser Operations", for details on how to modify these URLs for that
352 Writing/Uploading A File
353 ------------------------
355 ``PUT /uri/$FILECAP``
357 ``PUT /uri/$DIRCAP/[SUBDIRS../]FILENAME``
359 Upload a file, using the data from the HTTP request body, and add whatever
360 child links and subdirectories are necessary to make the file available at
361 the given location. Once this operation succeeds, a GET on the same URL will
362 retrieve the same contents that were just uploaded. This will create any
363 necessary intermediate subdirectories.
365 To use the /uri/$FILECAP form, $FILECAP must be a write-cap for a mutable file.
367 In the /uri/$DIRCAP/[SUBDIRS../]FILENAME form, if the target file is a
368 writeable mutable file, that file's contents will be overwritten
369 in-place. If it is a read-cap for a mutable file, an error will occur.
370 If it is an immutable file, the old file will be discarded, and a new
371 one will be put in its place. If the target file is a writable mutable
372 file, you may also specify an "offset" parameter -- a byte offset that
373 determines where in the mutable file the data from the HTTP request
374 body is placed. This operation is relatively efficient for MDMF mutable
375 files, and is relatively inefficient (but still supported) for SDMF
376 mutable files. If no offset parameter is specified, then the entire
377 file is replaced with the data from the HTTP request body. For an
378 immutable file, the "offset" parameter is not valid.
380 When creating a new file, if "mutable=true" is in the query arguments, the
381 operation will create a mutable file instead of an immutable one.
383 This returns the file-cap of the resulting file. If a new file was created
384 by this method, the HTTP response code (as dictated by rfc2616) will be set
385 to 201 CREATED. If an existing file was replaced or modified, the response
388 Note that the 'curl -T localfile http://127.0.0.1:3456/uri/$DIRCAP/foo.txt'
389 command can be used to invoke this operation.
393 This uploads a file, and produces a file-cap for the contents, but does not
394 attach the file into the filesystem. No directories will be modified by
395 this operation. The file-cap is returned as the body of the HTTP response.
397 If "mutable=true" is in the query arguments, the operation will create a
398 mutable file, and return its write-cap in the HTTP respose. The default is
399 to create an immutable file, returning the read-cap as a response. If
400 you create a mutable file, you can also use the "mutable-type" query
401 parameter. If "mutable-type=sdmf", then the mutable file will be created
402 in the old SDMF mutable file format. This is desirable for files that
403 need to be read by old clients. If "mutable-type=mdmf", then the file
404 will be created in the new MDMF mutable file format. MDMF mutable files
405 can be downloaded more efficiently, and modified in-place efficiently,
406 but are not compatible with older versions of Tahoe-LAFS. If no
407 "mutable-type" argument is given, the file is created in whatever
408 format was configured in tahoe.cfg.
411 Creating A New Directory
412 ------------------------
414 ``POST /uri?t=mkdir``
418 Create a new empty directory and return its write-cap as the HTTP response
419 body. This does not make the newly created directory visible from the
420 filesystem. The "PUT" operation is provided for backwards compatibility:
421 new code should use POST.
423 ``POST /uri?t=mkdir-with-children``
425 Create a new directory, populated with a set of child nodes, and return its
426 write-cap as the HTTP response body. The new directory is not attached to
427 any other directory: the returned write-cap is the only reference to it.
429 Initial children are provided as the body of the POST form (this is more
430 efficient than doing separate mkdir and set_children operations). If the
431 body is empty, the new directory will be empty. If not empty, the body will
432 be interpreted as a UTF-8 JSON-encoded dictionary of children with which the
433 new directory should be populated, using the same format as would be
434 returned in the 'children' value of the t=json GET request, described below.
435 Each dictionary key should be a child name, and each value should be a list
436 of [TYPE, PROPDICT], where PROPDICT contains "rw_uri", "ro_uri", and
437 "metadata" keys (all others are ignored). For example, the PUT request body
441 "Fran\u00e7ais": [ "filenode", {
442 "ro_uri": "URI:CHK:...",
445 "ctime": 1202777696.7564139,
446 "mtime": 1202777696.7564139,
448 "linkcrtime": 1202777696.7564139,
449 "linkmotime": 1202777696.7564139
451 "subdir": [ "dirnode", {
452 "rw_uri": "URI:DIR2:...",
453 "ro_uri": "URI:DIR2-RO:...",
455 "ctime": 1202778102.7589991,
456 "mtime": 1202778111.2160511,
458 "linkcrtime": 1202777696.7564139,
459 "linkmotime": 1202777696.7564139
463 For forward-compatibility, a mutable directory can also contain caps in
464 a format that is unknown to the web-API server. When such caps are retrieved
465 from a mutable directory in a "ro_uri" field, they will be prefixed with
466 the string "ro.", indicating that they must not be decoded without
467 checking that they are read-only. The "ro." prefix must not be stripped
468 off without performing this check. (Future versions of the web-API server
469 will perform it where necessary.)
471 If both the "rw_uri" and "ro_uri" fields are present in a given PROPDICT,
472 and the web-API server recognizes the rw_uri as a write cap, then it will
473 reset the ro_uri to the corresponding read cap and discard the original
474 contents of ro_uri (in order to ensure that the two caps correspond to the
475 same object and that the ro_uri is in fact read-only). However this may not
476 happen for caps in a format unknown to the web-API server. Therefore, when
477 writing a directory the web-API client should ensure that the contents
478 of "rw_uri" and "ro_uri" for a given PROPDICT are a consistent
479 (write cap, read cap) pair if possible. If the web-API client only has
480 one cap and does not know whether it is a write cap or read cap, then
481 it is acceptable to set "rw_uri" to that cap and omit "ro_uri". The
482 client must not put a write cap into a "ro_uri" field.
484 The metadata may have a "no-write" field. If this is set to true in the
485 metadata of a link, it will not be possible to open that link for writing
486 via the SFTP frontend; see `<FTP-and-SFTP.rst>`_ for details.
487 Also, if the "no-write" field is set to true in the metadata of a link to
488 a mutable child, it will cause the link to be diminished to read-only.
490 Note that the web-API-using client application must not provide the
491 "Content-Type: multipart/form-data" header that usually accompanies HTML
492 form submissions, since the body is not formatted this way. Doing so will
493 cause a server error as the lower-level code misparses the request body.
495 Child file names should each be expressed as a Unicode string, then used as
496 keys of the dictionary. The dictionary should then be converted into JSON,
497 and the resulting string encoded into UTF-8. This UTF-8 bytestring should
498 then be used as the POST body.
500 ``POST /uri?t=mkdir-immutable``
502 Like t=mkdir-with-children above, but the new directory will be
503 deep-immutable. This means that the directory itself is immutable, and that
504 it can only contain objects that are treated as being deep-immutable, like
505 immutable files, literal files, and deep-immutable directories.
507 For forward-compatibility, a deep-immutable directory can also contain caps
508 in a format that is unknown to the web-API server. When such caps are retrieved
509 from a deep-immutable directory in a "ro_uri" field, they will be prefixed
510 with the string "imm.", indicating that they must not be decoded without
511 checking that they are immutable. The "imm." prefix must not be stripped
512 off without performing this check. (Future versions of the web-API server
513 will perform it where necessary.)
515 The cap for each child may be given either in the "rw_uri" or "ro_uri"
516 field of the PROPDICT (not both). If a cap is given in the "rw_uri" field,
517 then the web-API server will check that it is an immutable read-cap of a
518 *known* format, and give an error if it is not. If a cap is given in the
519 "ro_uri" field, then the web-API server will still check whether known
520 caps are immutable, but for unknown caps it will simply assume that the
521 cap can be stored, as described above. Note that an attacker would be
522 able to store any cap in an immutable directory, so this check when
523 creating the directory is only to help non-malicious clients to avoid
524 accidentally giving away more authority than intended.
526 A non-empty request body is mandatory, since after the directory is created,
527 it will not be possible to add more children to it.
529 ``POST /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=mkdir``
531 ``PUT /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=mkdir``
533 Create new directories as necessary to make sure that the named target
534 ($DIRCAP/SUBDIRS../SUBDIR) is a directory. This will create additional
535 intermediate mutable directories as necessary. If the named target directory
536 already exists, this will make no changes to it.
538 If the final directory is created, it will be empty.
540 This operation will return an error if a blocking file is present at any of
541 the parent names, preventing the server from creating the necessary parent
542 directory; or if it would require changing an immutable directory.
544 The write-cap of the new directory will be returned as the HTTP response
547 ``POST /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=mkdir-with-children``
549 Like /uri?t=mkdir-with-children, but the final directory is created as a
550 child of an existing mutable directory. This will create additional
551 intermediate mutable directories as necessary. If the final directory is
552 created, it will be populated with initial children from the POST request
553 body, as described above.
555 This operation will return an error if a blocking file is present at any of
556 the parent names, preventing the server from creating the necessary parent
557 directory; or if it would require changing an immutable directory; or if
558 the immediate parent directory already has a a child named SUBDIR.
560 ``POST /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=mkdir-immutable``
562 Like /uri?t=mkdir-immutable, but the final directory is created as a child
563 of an existing mutable directory. The final directory will be deep-immutable,
564 and will be populated with the children specified as a JSON dictionary in
565 the POST request body.
567 In Tahoe 1.6 this operation creates intermediate mutable directories if
568 necessary, but that behaviour should not be relied on; see ticket #920.
570 This operation will return an error if the parent directory is immutable,
571 or already has a child named SUBDIR.
573 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=mkdir&name=NAME``
575 Create a new empty mutable directory and attach it to the given existing
576 directory. This will create additional intermediate directories as necessary.
578 This operation will return an error if a blocking file is present at any of
579 the parent names, preventing the server from creating the necessary parent
580 directory, or if it would require changing any immutable directory.
582 The URL of this operation points to the parent of the bottommost new directory,
583 whereas the /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=mkdir operation above has a URL
584 that points directly to the bottommost new directory.
586 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=mkdir-with-children&name=NAME``
588 Like /uri/$DIRCAP/[SUBDIRS../]?t=mkdir&name=NAME, but the new directory will
589 be populated with initial children via the POST request body. This command
590 will create additional intermediate mutable directories as necessary.
592 This operation will return an error if a blocking file is present at any of
593 the parent names, preventing the server from creating the necessary parent
594 directory; or if it would require changing an immutable directory; or if
595 the immediate parent directory already has a a child named NAME.
597 Note that the name= argument must be passed as a queryarg, because the POST
598 request body is used for the initial children JSON.
600 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=mkdir-immutable&name=NAME``
602 Like /uri/$DIRCAP/[SUBDIRS../]?t=mkdir-with-children&name=NAME, but the
603 final directory will be deep-immutable. The children are specified as a
604 JSON dictionary in the POST request body. Again, the name= argument must be
605 passed as a queryarg.
607 In Tahoe 1.6 this operation creates intermediate mutable directories if
608 necessary, but that behaviour should not be relied on; see ticket #920.
610 This operation will return an error if the parent directory is immutable,
611 or already has a child named NAME.
614 Getting Information About A File Or Directory (as JSON)
615 -------------------------------------------------------
617 ``GET /uri/$FILECAP?t=json``
619 ``GET /uri/$DIRCAP?t=json``
621 ``GET /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=json``
623 ``GET /uri/$DIRCAP/[SUBDIRS../]FILENAME?t=json``
625 This returns a machine-parseable JSON-encoded description of the given
626 object. The JSON always contains a list, and the first element of the list is
627 always a flag that indicates whether the referenced object is a file or a
628 directory. If it is a capability to a file, then the information includes
629 file size and URI, like this::
631 GET /uri/$FILECAP?t=json :
635 "verify_uri": verify_uri,
640 If it is a capability to a directory followed by a path from that directory
641 to a file, then the information also includes metadata from the link to the
642 file in the parent directory, like this::
644 GET /uri/$DIRCAP/[SUBDIRS../]FILENAME?t=json
648 "verify_uri": verify_uri,
652 "ctime": 1202777696.7564139,
653 "mtime": 1202777696.7564139,
655 "linkcrtime": 1202777696.7564139,
656 "linkmotime": 1202777696.7564139
659 If it is a directory, then it includes information about the children of
660 this directory, as a mapping from child name to a set of data about the
661 child (the same data that would appear in a corresponding GET?t=json of the
662 child itself). The child entries also include metadata about each child,
663 including link-creation- and link-change- timestamps. The output looks like
666 GET /uri/$DIRCAP?t=json :
667 GET /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=json :
670 "rw_uri": read_write_uri,
671 "ro_uri": read_only_uri,
672 "verify_uri": verify_uri,
675 "foo.txt": [ "filenode", {
679 "ctime": 1202777696.7564139,
680 "mtime": 1202777696.7564139,
682 "linkcrtime": 1202777696.7564139,
683 "linkmotime": 1202777696.7564139
685 "subdir": [ "dirnode", {
689 "ctime": 1202778102.7589991,
690 "mtime": 1202778111.2160511,
692 "linkcrtime": 1202777696.7564139,
693 "linkmotime": 1202777696.7564139
697 In the above example, note how 'children' is a dictionary in which the keys
698 are child names and the values depend upon whether the child is a file or a
699 directory. The value is mostly the same as the JSON representation of the
700 child object (except that directories do not recurse -- the "children"
701 entry of the child is omitted, and the directory view includes the metadata
702 that is stored on the directory edge).
704 The rw_uri field will be present in the information about a directory
705 if and only if you have read-write access to that directory. The verify_uri
706 field will be present if and only if the object has a verify-cap
707 (non-distributed LIT files do not have verify-caps).
709 If the cap is of an unknown format, then the file size and verify_uri will
712 GET /uri/$UNKNOWNCAP?t=json :
715 "ro_uri": unknown_read_uri
718 GET /uri/$DIRCAP/[SUBDIRS../]UNKNOWNCHILDNAME?t=json :
721 "rw_uri": unknown_write_uri,
722 "ro_uri": unknown_read_uri,
725 "ctime": 1202777696.7564139,
726 "mtime": 1202777696.7564139,
728 "linkcrtime": 1202777696.7564139,
729 "linkmotime": 1202777696.7564139
732 As in the case of file nodes, the metadata will only be present when the
733 capability is to a directory followed by a path. The "mutable" field is also
734 not always present; when it is absent, the mutability of the object is not
740 The value of the 'tahoe':'linkmotime' key is updated whenever a link to a
741 child is set. The value of the 'tahoe':'linkcrtime' key is updated whenever
742 a link to a child is created -- i.e. when there was not previously a link
745 Note however, that if the edge in the Tahoe filesystem points to a mutable
746 file and the contents of that mutable file is changed, then the
747 'tahoe':'linkmotime' value on that edge will *not* be updated, since the
748 edge itself wasn't updated -- only the mutable file was.
750 The timestamps are represented as a number of seconds since the UNIX epoch
751 (1970-01-01 00:00:00 UTC), with leap seconds not being counted in the long
754 In Tahoe earlier than v1.4.0, 'mtime' and 'ctime' keys were populated
755 instead of the 'tahoe':'linkmotime' and 'tahoe':'linkcrtime' keys. Starting
756 in Tahoe v1.4.0, the 'linkmotime'/'linkcrtime' keys in the 'tahoe' sub-dict
757 are populated. However, prior to Tahoe v1.7beta, a bug caused the 'tahoe'
758 sub-dict to be deleted by web-API requests in which new metadata is
759 specified, and not to be added to existing child links that lack it.
761 From Tahoe v1.7.0 onward, the 'mtime' and 'ctime' fields are no longer
762 populated or updated (see ticket #924), except by "tahoe backup" as
763 explained below. For backward compatibility, when an existing link is
764 updated and 'tahoe':'linkcrtime' is not present in the previous metadata
765 but 'ctime' is, the old value of 'ctime' is used as the new value of
766 'tahoe':'linkcrtime'.
768 The reason we added the new fields in Tahoe v1.4.0 is that there is a
769 "set_children" API (described below) which you can use to overwrite the
770 values of the 'mtime'/'ctime' pair, and this API is used by the
771 "tahoe backup" command (in Tahoe v1.3.0 and later) to set the 'mtime' and
772 'ctime' values when backing up files from a local filesystem into the
773 Tahoe filesystem. As of Tahoe v1.4.0, the set_children API cannot be used
774 to set anything under the 'tahoe' key of the metadata dict -- if you
775 include 'tahoe' keys in your 'metadata' arguments then it will silently
778 Therefore, if the 'tahoe' sub-dict is present, you can rely on the
779 'linkcrtime' and 'linkmotime' values therein to have the semantics described
780 above. (This is assuming that only official Tahoe clients have been used to
781 write those links, and that their system clocks were set to what you expected
782 -- there is nothing preventing someone from editing their Tahoe client or
783 writing their own Tahoe client which would overwrite those values however
784 they like, and there is nothing to constrain their system clock from taking
787 When an edge is created or updated by "tahoe backup", the 'mtime' and
788 'ctime' keys on that edge are set as follows:
790 * 'mtime' is set to the timestamp read from the local filesystem for the
791 "mtime" of the local file in question, which means the last time the
792 contents of that file were changed.
794 * On Windows, 'ctime' is set to the creation timestamp for the file
795 read from the local filesystem. On other platforms, 'ctime' is set to
796 the UNIX "ctime" of the local file, which means the last time that
797 either the contents or the metadata of the local file was changed.
799 There are several ways that the 'ctime' field could be confusing:
801 1. You might be confused about whether it reflects the time of the creation
802 of a link in the Tahoe filesystem (by a version of Tahoe < v1.7.0) or a
803 timestamp copied in by "tahoe backup" from a local filesystem.
805 2. You might be confused about whether it is a copy of the file creation
806 time (if "tahoe backup" was run on a Windows system) or of the last
807 contents-or-metadata change (if "tahoe backup" was run on a different
810 3. You might be confused by the fact that changing the contents of a
811 mutable file in Tahoe doesn't have any effect on any links pointing at
812 that file in any directories, although "tahoe backup" sets the link
813 'ctime'/'mtime' to reflect timestamps about the local file corresponding
814 to the Tahoe file to which the link points.
816 4. Also, quite apart from Tahoe, you might be confused about the meaning
817 of the "ctime" in UNIX local filesystems, which people sometimes think
818 means file creation time, but which actually means, in UNIX local
819 filesystems, the most recent time that the file contents or the file
820 metadata (such as owner, permission bits, extended attributes, etc.)
821 has changed. Note that although "ctime" does not mean file creation time
822 in UNIX, links created by a version of Tahoe prior to v1.7.0, and never
823 written by "tahoe backup", will have 'ctime' set to the link creation
827 Attaching an Existing File or Directory by its read- or write-cap
828 -----------------------------------------------------------------
830 ``PUT /uri/$DIRCAP/[SUBDIRS../]CHILDNAME?t=uri``
832 This attaches a child object (either a file or directory) to a specified
833 location in the virtual filesystem. The child object is referenced by its
834 read- or write- cap, as provided in the HTTP request body. This will create
835 intermediate directories as necessary.
837 This is similar to a UNIX hardlink: by referencing a previously-uploaded file
838 (or previously-created directory) instead of uploading/creating a new one,
839 you can create two references to the same object.
841 The read- or write- cap of the child is provided in the body of the HTTP
842 request, and this same cap is returned in the response body.
844 The default behavior is to overwrite any existing object at the same
845 location. To prevent this (and make the operation return an error instead
846 of overwriting), add a "replace=false" argument, as "?t=uri&replace=false".
847 With replace=false, this operation will return an HTTP 409 "Conflict" error
848 if there is already an object at the given location, rather than
849 overwriting the existing object. To allow the operation to overwrite a
850 file, but return an error when trying to overwrite a directory, use
851 "replace=only-files" (this behavior is closer to the traditional UNIX "mv"
852 command). Note that "true", "t", and "1" are all synonyms for "True", and
853 "false", "f", and "0" are synonyms for "False", and the parameter is
856 Note that this operation does not take its child cap in the form of
857 separate "rw_uri" and "ro_uri" fields. Therefore, it cannot accept a
858 child cap in a format unknown to the web-API server, unless its URI
859 starts with "ro." or "imm.". This restriction is necessary because the
860 server is not able to attenuate an unknown write cap to a read cap.
861 Unknown URIs starting with "ro." or "imm.", on the other hand, are
862 assumed to represent read caps. The client should not prefix a write
863 cap with "ro." or "imm." and pass it to this operation, since that
864 would result in granting the cap's write authority to holders of the
868 Adding Multiple Files or Directories to a Parent Directory at Once
869 ------------------------------------------------------------------
871 ``POST /uri/$DIRCAP/[SUBDIRS..]?t=set_children``
873 ``POST /uri/$DIRCAP/[SUBDIRS..]?t=set-children`` (Tahoe >= v1.6)
875 This command adds multiple children to a directory in a single operation.
876 It reads the request body and interprets it as a JSON-encoded description
877 of the child names and read/write-caps that should be added.
879 The body should be a JSON-encoded dictionary, in the same format as the
880 "children" value returned by the "GET /uri/$DIRCAP?t=json" operation
881 described above. In this format, each key is a child names, and the
882 corresponding value is a tuple of (type, childinfo). "type" is ignored, and
883 "childinfo" is a dictionary that contains "rw_uri", "ro_uri", and
884 "metadata" keys. You can take the output of "GET /uri/$DIRCAP1?t=json" and
885 use it as the input to "POST /uri/$DIRCAP2?t=set_children" to make DIR2
886 look very much like DIR1 (except for any existing children of DIR2 that
887 were not overwritten, and any existing "tahoe" metadata keys as described
890 When the set_children request contains a child name that already exists in
891 the target directory, this command defaults to overwriting that child with
892 the new value (both child cap and metadata, but if the JSON data does not
893 contain a "metadata" key, the old child's metadata is preserved). The
894 command takes a boolean "overwrite=" query argument to control this
895 behavior. If you use "?t=set_children&overwrite=false", then an attempt to
896 replace an existing child will instead cause an error.
898 Any "tahoe" key in the new child's "metadata" value is ignored. Any
899 existing "tahoe" metadata is preserved. The metadata["tahoe"] value is
900 reserved for metadata generated by the tahoe node itself. The only two keys
901 currently placed here are "linkcrtime" and "linkmotime". For details, see
902 the section above entitled "Get Information About A File Or Directory (as
903 JSON)", in the "About the metadata" subsection.
905 Note that this command was introduced with the name "set_children", which
906 uses an underscore rather than a hyphen as other multi-word command names
907 do. The variant with a hyphen is now accepted, but clients that desire
908 backward compatibility should continue to use "set_children".
911 Unlinking a File or Directory
912 -----------------------------
914 ``DELETE /uri/$DIRCAP/[SUBDIRS../]CHILDNAME``
916 This removes the given name from its parent directory. CHILDNAME is the
917 name to be removed, and $DIRCAP/SUBDIRS.. indicates the directory that will
920 Note that this does not actually delete the file or directory that the name
921 points to from the tahoe grid -- it only unlinks the named reference from
922 this directory. If there are other names in this directory or in other
923 directories that point to the resource, then it will remain accessible
924 through those paths. Even if all names pointing to this object are removed
925 from their parent directories, then someone with possession of its read-cap
926 can continue to access the object through that cap.
928 The object will only become completely unreachable once 1: there are no
929 reachable directories that reference it, and 2: nobody is holding a read-
930 or write- cap to the object. (This behavior is very similar to the way
931 hardlinks and anonymous files work in traditional UNIX filesystems).
933 This operation will not modify more than a single directory. Intermediate
934 directories which were implicitly created by PUT or POST methods will *not*
935 be automatically removed by DELETE.
937 This method returns the file- or directory- cap of the object that was just
941 Browser Operations: Human-oriented interfaces
942 =============================================
944 This section describes the HTTP operations that provide support for humans
945 running a web browser. Most of these operations use HTML forms that use POST
946 to drive the Tahoe node. This section is intended for HTML authors who want
947 to write web pages that contain forms and buttons which manipulate the Tahoe
950 Note that for all POST operations, the arguments listed can be provided
951 either as URL query arguments or as form body fields. URL query arguments are
952 separated from the main URL by "?", and from each other by "&". For example,
953 "POST /uri/$DIRCAP?t=upload&mutable=true". Form body fields are usually
954 specified by using <input type="hidden"> elements. For clarity, the
955 descriptions below display the most significant arguments as URL query args.
958 Viewing A Directory (as HTML)
959 -----------------------------
961 ``GET /uri/$DIRCAP/[SUBDIRS../]``
963 This returns an HTML page, intended to be displayed to a human by a web
964 browser, which contains HREF links to all files and directories reachable
965 from this directory. These HREF links do not have a t= argument, meaning
966 that a human who follows them will get pages also meant for a human. It also
967 contains forms to upload new files, and to unlink files and directories
968 from their parent directory. Those forms use POST methods to do their job.
971 Viewing/Downloading a File
972 --------------------------
974 ``GET /uri/$FILECAP``
976 ``GET /uri/$DIRCAP/[SUBDIRS../]FILENAME``
978 This will retrieve the contents of the given file. The HTTP response body
979 will contain the sequence of bytes that make up the file.
981 If you want the HTTP response to include a useful Content-Type header,
982 either use the second form (which starts with a $DIRCAP), or add a
983 "filename=foo" query argument, like "GET /uri/$FILECAP?filename=foo.jpg".
984 The bare "GET /uri/$FILECAP" does not give the Tahoe node enough information
985 to determine a Content-Type (since Tahoe immutable files are merely
986 sequences of bytes, not typed+named file objects).
988 If the URL has both filename= and "save=true" in the query arguments, then
989 the server to add a "Content-Disposition: attachment" header, along with a
990 filename= parameter. When a user clicks on such a link, most browsers will
991 offer to let the user save the file instead of displaying it inline (indeed,
992 most browsers will refuse to display it inline). "true", "t", "1", and other
993 case-insensitive equivalents are all treated the same.
995 Character-set handling in URLs and HTTP headers is a dubious art [1]_. For
996 maximum compatibility, Tahoe simply copies the bytes from the filename=
997 argument into the Content-Disposition header's filename= parameter, without
998 trying to interpret them in any particular way.
1001 ``GET /named/$FILECAP/FILENAME``
1003 This is an alternate download form which makes it easier to get the correct
1004 filename. The Tahoe server will provide the contents of the given file, with
1005 a Content-Type header derived from the given filename. This form is used to
1006 get browsers to use the "Save Link As" feature correctly, and also helps
1007 command-line tools like "wget" and "curl" use the right filename. Note that
1008 this form can *only* be used with file caps; it is an error to use a
1009 directory cap after the /named/ prefix.
1012 Getting Information About A File Or Directory (as HTML)
1013 -------------------------------------------------------
1015 ``GET /uri/$FILECAP?t=info``
1017 ``GET /uri/$DIRCAP/?t=info``
1019 ``GET /uri/$DIRCAP/[SUBDIRS../]SUBDIR/?t=info``
1021 ``GET /uri/$DIRCAP/[SUBDIRS../]FILENAME?t=info``
1023 This returns a human-oriented HTML page with more detail about the selected
1024 file or directory object. This page contains the following items:
1028 * JSON representation
1029 * raw contents (text/plain)
1030 * access caps (URIs): verify-cap, read-cap, write-cap (for mutable objects)
1031 * check/verify/repair form
1032 * deep-check/deep-size/deep-stats/manifest (for directories)
1033 * replace-conents form (for mutable files)
1036 Creating a Directory
1037 --------------------
1039 ``POST /uri?t=mkdir``
1041 This creates a new empty directory, but does not attach it to the virtual
1044 If a "redirect_to_result=true" argument is provided, then the HTTP response
1045 will cause the web browser to be redirected to a /uri/$DIRCAP page that
1046 gives access to the newly-created directory. If you bookmark this page,
1047 you'll be able to get back to the directory again in the future. This is the
1048 recommended way to start working with a Tahoe server: create a new unlinked
1049 directory (using redirect_to_result=true), then bookmark the resulting
1050 /uri/$DIRCAP page. There is a "create directory" button on the Welcome page
1051 to invoke this action.
1053 If "redirect_to_result=true" is not provided (or is given a value of
1054 "false"), then the HTTP response body will simply be the write-cap of the
1057 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=mkdir&name=CHILDNAME``
1059 This creates a new empty directory as a child of the designated SUBDIR. This
1060 will create additional intermediate directories as necessary.
1062 If a "when_done=URL" argument is provided, the HTTP response will cause the
1063 web browser to redirect to the given URL. This provides a convenient way to
1064 return the browser to the directory that was just modified. Without a
1065 when_done= argument, the HTTP response will simply contain the write-cap of
1066 the directory that was just created.
1072 ``POST /uri?t=upload``
1074 This uploads a file, and produces a file-cap for the contents, but does not
1075 attach the file into the filesystem. No directories will be modified by
1078 The file must be provided as the "file" field of an HTML encoded form body,
1079 produced in response to an HTML form like this::
1081 <form action="/uri" method="POST" enctype="multipart/form-data">
1082 <input type="hidden" name="t" value="upload" />
1083 <input type="file" name="file" />
1084 <input type="submit" value="Upload Unlinked" />
1087 If a "when_done=URL" argument is provided, the response body will cause the
1088 browser to redirect to the given URL. If the when_done= URL has the string
1089 "%(uri)s" in it, that string will be replaced by a URL-escaped form of the
1090 newly created file-cap. (Note that without this substitution, there is no
1091 way to access the file that was just uploaded).
1093 The default (in the absence of when_done=) is to return an HTML page that
1094 describes the results of the upload. This page will contain information
1095 about which storage servers were used for the upload, how long each
1096 operation took, etc.
1098 If a "mutable=true" argument is provided, the operation will create a
1099 mutable file, and the response body will contain the write-cap instead of
1100 the upload results page. The default is to create an immutable file,
1101 returning the upload results page as a response. If you create a
1102 mutable file, you may choose to specify the format of that mutable file
1103 with the "mutable-type" parameter. If "mutable-type=mdmf", then the
1104 file will be created as an MDMF mutable file. If "mutable-type=sdmf",
1105 then the file will be created as an SDMF mutable file. If no value is
1106 specified, the file will be created in whatever format is specified in
1110 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=upload``
1112 This uploads a file, and attaches it as a new child of the given directory,
1113 which must be mutable. The file must be provided as the "file" field of an
1114 HTML-encoded form body, produced in response to an HTML form like this::
1116 <form action="." method="POST" enctype="multipart/form-data">
1117 <input type="hidden" name="t" value="upload" />
1118 <input type="file" name="file" />
1119 <input type="submit" value="Upload" />
1122 A "name=" argument can be provided to specify the new child's name,
1123 otherwise it will be taken from the "filename" field of the upload form
1124 (most web browsers will copy the last component of the original file's
1125 pathname into this field). To avoid confusion, name= is not allowed to
1128 If there is already a child with that name, and it is a mutable file, then
1129 its contents are replaced with the data being uploaded. If it is not a
1130 mutable file, the default behavior is to remove the existing child before
1131 creating a new one. To prevent this (and make the operation return an error
1132 instead of overwriting the old child), add a "replace=false" argument, as
1133 "?t=upload&replace=false". With replace=false, this operation will return an
1134 HTTP 409 "Conflict" error if there is already an object at the given
1135 location, rather than overwriting the existing object. Note that "true",
1136 "t", and "1" are all synonyms for "True", and "false", "f", and "0" are
1137 synonyms for "False". the parameter is case-insensitive.
1139 This will create additional intermediate directories as necessary, although
1140 since it is expected to be triggered by a form that was retrieved by "GET
1141 /uri/$DIRCAP/[SUBDIRS../]", it is likely that the parent directory will
1144 If a "mutable=true" argument is provided, any new file that is created will
1145 be a mutable file instead of an immutable one. <input type="checkbox"
1146 name="mutable" /> will give the user a way to set this option.
1148 If a "when_done=URL" argument is provided, the HTTP response will cause the
1149 web browser to redirect to the given URL. This provides a convenient way to
1150 return the browser to the directory that was just modified. Without a
1151 when_done= argument, the HTTP response will simply contain the file-cap of
1152 the file that was just uploaded (a write-cap for mutable files, or a
1153 read-cap for immutable files).
1155 ``POST /uri/$DIRCAP/[SUBDIRS../]FILENAME?t=upload``
1157 This also uploads a file and attaches it as a new child of the given
1158 directory, which must be mutable. It is a slight variant of the previous
1159 operation, as the URL refers to the target file rather than the parent
1160 directory. It is otherwise identical: this accepts mutable= and when_done=
1163 ``POST /uri/$FILECAP?t=upload``
1165 This modifies the contents of an existing mutable file in-place. An error is
1166 signalled if $FILECAP does not refer to a mutable file. It behaves just like
1167 the "PUT /uri/$FILECAP" form, but uses a POST for the benefit of HTML forms
1171 Attaching An Existing File Or Directory (by URI)
1172 ------------------------------------------------
1174 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=uri&name=CHILDNAME&uri=CHILDCAP``
1176 This attaches a given read- or write- cap "CHILDCAP" to the designated
1177 directory, with a specified child name. This behaves much like the PUT t=uri
1178 operation, and is a lot like a UNIX hardlink. It is subject to the same
1179 restrictions as that operation on the use of cap formats unknown to the
1182 This will create additional intermediate directories as necessary, although
1183 since it is expected to be triggered by a form that was retrieved by "GET
1184 /uri/$DIRCAP/[SUBDIRS../]", it is likely that the parent directory will
1187 This accepts the same replace= argument as POST t=upload.
1193 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=delete&name=CHILDNAME``
1195 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=unlink&name=CHILDNAME``
1197 This instructs the node to remove a child object (file or subdirectory) from
1198 the given directory, which must be mutable. Note that the entire subtree is
1199 unlinked from the parent. Unlike deleting a subdirectory in a UNIX local
1200 filesystem, the subtree need not be empty; if it isn't, then other references
1201 into the subtree will see that the child subdirectories are not modified by
1202 this operation. Only the link from the given directory to its child is severed.
1204 In Tahoe-LAFS v1.9.0 and later, t=unlink can be used as a synonym for t=delete.
1205 If interoperability with older web-API servers is required, t=delete should
1212 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=rename&from_name=OLD&to_name=NEW``
1214 This instructs the node to rename a child of the given directory, which must
1215 be mutable. This has a similar effect to removing the child, then adding the
1216 same child-cap under the new name, except that it preserves metadata. This
1217 operation cannot move the child to a different directory.
1219 This operation will replace any existing child of the new name, making it
1220 behave like the UNIX "``mv -f``" command.
1225 ``GET /uri?uri=$CAP``
1227 This causes a redirect to /uri/$CAP, and retains any additional query
1228 arguments (like filename= or save=). This is for the convenience of web
1229 forms which allow the user to paste in a read- or write- cap (obtained
1230 through some out-of-band channel, like IM or email).
1232 Note that this form merely redirects to the specific file or directory
1233 indicated by the $CAP: unlike the GET /uri/$DIRCAP form, you cannot
1234 traverse to children by appending additional path segments to the URL.
1236 ``GET /uri/$DIRCAP/[SUBDIRS../]?t=rename-form&name=$CHILDNAME``
1238 This provides a useful facility to browser-based user interfaces. It
1239 returns a page containing a form targetting the "POST $DIRCAP t=rename"
1240 functionality described above, with the provided $CHILDNAME present in the
1241 'from_name' field of that form. I.e. this presents a form offering to
1242 rename $CHILDNAME, requesting the new name, and submitting POST rename.
1244 ``GET /uri/$DIRCAP/[SUBDIRS../]CHILDNAME?t=uri``
1246 This returns the file- or directory- cap for the specified object.
1248 ``GET /uri/$DIRCAP/[SUBDIRS../]CHILDNAME?t=readonly-uri``
1250 This returns a read-only file- or directory- cap for the specified object.
1251 If the object is an immutable file, this will return the same value as
1255 Debugging and Testing Features
1256 ------------------------------
1258 These URLs are less-likely to be helpful to the casual Tahoe user, and are
1259 mainly intended for developers.
1261 ``POST $URL?t=check``
1263 This triggers the FileChecker to determine the current "health" of the
1264 given file or directory, by counting how many shares are available. The
1265 page that is returned will display the results. This can be used as a "show
1266 me detailed information about this file" page.
1268 If a verify=true argument is provided, the node will perform a more
1269 intensive check, downloading and verifying every single bit of every share.
1271 If an add-lease=true argument is provided, the node will also add (or
1272 renew) a lease to every share it encounters. Each lease will keep the share
1273 alive for a certain period of time (one month by default). Once the last
1274 lease expires or is explicitly cancelled, the storage server is allowed to
1277 If an output=JSON argument is provided, the response will be
1278 machine-readable JSON instead of human-oriented HTML. The data is a
1279 dictionary with the following keys::
1281 storage-index: a base32-encoded string with the objects's storage index,
1282 or an empty string for LIT files
1283 summary: a string, with a one-line summary of the stats of the file
1284 results: a dictionary that describes the state of the file. For LIT files,
1285 this dictionary has only the 'healthy' key, which will always be
1286 True. For distributed files, this dictionary has the following
1288 count-shares-good: the number of good shares that were found
1289 count-shares-needed: 'k', the number of shares required for recovery
1290 count-shares-expected: 'N', the number of total shares generated
1291 count-good-share-hosts: this was intended to be the number of distinct
1292 storage servers with good shares. It is currently
1293 (as of Tahoe-LAFS v1.8.0) computed incorrectly;
1295 count-wrong-shares: for mutable files, the number of shares for
1296 versions other than the 'best' one (highest
1297 sequence number, highest roothash). These are
1299 count-recoverable-versions: for mutable files, the number of
1300 recoverable versions of the file. For
1301 a healthy file, this will equal 1.
1302 count-unrecoverable-versions: for mutable files, the number of
1303 unrecoverable versions of the file.
1304 For a healthy file, this will be 0.
1305 count-corrupt-shares: the number of shares with integrity failures
1306 list-corrupt-shares: a list of "share locators", one for each share
1307 that was found to be corrupt. Each share locator
1308 is a list of (serverid, storage_index, sharenum).
1309 needs-rebalancing: (bool) True if there are multiple shares on a single
1310 storage server, indicating a reduction in reliability
1311 that could be resolved by moving shares to new
1313 servers-responding: list of base32-encoded storage server identifiers,
1314 one for each server which responded to the share
1316 healthy: (bool) True if the file is completely healthy, False otherwise.
1317 Healthy files have at least N good shares. Overlapping shares
1318 do not currently cause a file to be marked unhealthy. If there
1319 are at least N good shares, then corrupt shares do not cause the
1320 file to be marked unhealthy, although the corrupt shares will be
1321 listed in the results (list-corrupt-shares) and should be manually
1322 removed to wasting time in subsequent downloads (as the
1323 downloader rediscovers the corruption and uses alternate shares).
1324 Future compatibility: the meaning of this field may change to
1325 reflect whether the servers-of-happiness criterion is met
1327 sharemap: dict mapping share identifier to list of serverids
1328 (base32-encoded strings). This indicates which servers are
1329 holding which shares. For immutable files, the shareid is
1330 an integer (the share number, from 0 to N-1). For
1331 immutable files, it is a string of the form
1332 'seq%d-%s-sh%d', containing the sequence number, the
1333 roothash, and the share number.
1335 ``POST $URL?t=start-deep-check`` (must add &ophandle=XYZ)
1337 This initiates a recursive walk of all files and directories reachable from
1338 the target, performing a check on each one just like t=check. The result
1339 page will contain a summary of the results, including details on any
1340 file/directory that was not fully healthy.
1342 t=start-deep-check can only be invoked on a directory. An error (400
1343 BAD_REQUEST) will be signalled if it is invoked on a file. The recursive
1344 walker will deal with loops safely.
1346 This accepts the same verify= and add-lease= arguments as t=check.
1348 Since this operation can take a long time (perhaps a second per object),
1349 the ophandle= argument is required (see "Slow Operations, Progress, and
1350 Cancelling" above). The response to this POST will be a redirect to the
1351 corresponding /operations/$HANDLE page (with output=HTML or output=JSON to
1352 match the output= argument given to the POST). The deep-check operation
1353 will continue to run in the background, and the /operations page should be
1354 used to find out when the operation is done.
1356 Detailed check results for non-healthy files and directories will be
1357 available under /operations/$HANDLE/$STORAGEINDEX, and the HTML status will
1358 contain links to these detailed results.
1360 The HTML /operations/$HANDLE page for incomplete operations will contain a
1361 meta-refresh tag, set to 60 seconds, so that a browser which uses
1362 deep-check will automatically poll until the operation has completed.
1364 The JSON page (/options/$HANDLE?output=JSON) will contain a
1365 machine-readable JSON dictionary with the following keys::
1367 finished: a boolean, True if the operation is complete, else False. Some
1368 of the remaining keys may not be present until the operation
1370 root-storage-index: a base32-encoded string with the storage index of the
1371 starting point of the deep-check operation
1372 count-objects-checked: count of how many objects were checked. Note that
1373 non-distributed objects (i.e. small immutable LIT
1374 files) are not checked, since for these objects,
1375 the data is contained entirely in the URI.
1376 count-objects-healthy: how many of those objects were completely healthy
1377 count-objects-unhealthy: how many were damaged in some way
1378 count-corrupt-shares: how many shares were found to have corruption,
1379 summed over all objects examined
1380 list-corrupt-shares: a list of "share identifiers", one for each share
1381 that was found to be corrupt. Each share identifier
1382 is a list of (serverid, storage_index, sharenum).
1383 list-unhealthy-files: a list of (pathname, check-results) tuples, for
1384 each file that was not fully healthy. 'pathname' is
1385 a list of strings (which can be joined by "/"
1386 characters to turn it into a single string),
1387 relative to the directory on which deep-check was
1388 invoked. The 'check-results' field is the same as
1389 that returned by t=check&output=JSON, described
1391 stats: a dictionary with the same keys as the t=start-deep-stats command
1394 ``POST $URL?t=stream-deep-check``
1396 This initiates a recursive walk of all files and directories reachable from
1397 the target, performing a check on each one just like t=check. For each
1398 unique object (duplicates are skipped), a single line of JSON is emitted to
1399 the HTTP response channel (or an error indication, see below). When the walk
1400 is complete, a final line of JSON is emitted which contains the accumulated
1401 file-size/count "deep-stats" data.
1403 This command takes the same arguments as t=start-deep-check.
1405 A CLI tool can split the response stream on newlines into "response units",
1406 and parse each response unit as JSON. Each such parsed unit will be a
1407 dictionary, and will contain at least the "type" key: a string, one of
1408 "file", "directory", or "stats".
1410 For all units that have a type of "file" or "directory", the dictionary will
1411 contain the following keys::
1413 "path": a list of strings, with the path that is traversed to reach the
1415 "cap": a write-cap URI for the file or directory, if available, else a
1417 "verifycap": a verify-cap URI for the file or directory
1418 "repaircap": an URI for the weakest cap that can still be used to repair
1420 "storage-index": a base32 storage index for the object
1421 "check-results": a copy of the dictionary which would be returned by
1422 t=check&output=json, with three top-level keys:
1423 "storage-index", "summary", and "results", and a variety
1424 of counts and sharemaps in the "results" value.
1426 Note that non-distributed files (i.e. LIT files) will have values of None
1427 for verifycap, repaircap, and storage-index, since these files can neither
1428 be verified nor repaired, and are not stored on the storage servers.
1429 Likewise the check-results dictionary will be limited: an empty string for
1430 storage-index, and a results dictionary with only the "healthy" key.
1432 The last unit in the stream will have a type of "stats", and will contain
1433 the keys described in the "start-deep-stats" operation, below.
1435 If any errors occur during the traversal (specifically if a directory is
1436 unrecoverable, such that further traversal is not possible), an error
1437 indication is written to the response body, instead of the usual line of
1438 JSON. This error indication line will begin with the string "ERROR:" (in all
1439 caps), and contain a summary of the error on the rest of the line. The
1440 remaining lines of the response body will be a python exception. The client
1441 application should look for the ERROR: and stop processing JSON as soon as
1442 it is seen. Note that neither a file being unrecoverable nor a directory
1443 merely being unhealthy will cause traversal to stop. The line just before
1444 the ERROR: will describe the directory that was untraversable, since the
1445 unit is emitted to the HTTP response body before the child is traversed.
1448 ``POST $URL?t=check&repair=true``
1450 This performs a health check of the given file or directory, and if the
1451 checker determines that the object is not healthy (some shares are missing
1452 or corrupted), it will perform a "repair". During repair, any missing
1453 shares will be regenerated and uploaded to new servers.
1455 This accepts the same verify=true and add-lease= arguments as t=check. When
1456 an output=JSON argument is provided, the machine-readable JSON response
1457 will contain the following keys::
1459 storage-index: a base32-encoded string with the objects's storage index,
1460 or an empty string for LIT files
1461 repair-attempted: (bool) True if repair was attempted
1462 repair-successful: (bool) True if repair was attempted and the file was
1463 fully healthy afterwards. False if no repair was
1464 attempted, or if a repair attempt failed.
1465 pre-repair-results: a dictionary that describes the state of the file
1466 before any repair was performed. This contains exactly
1467 the same keys as the 'results' value of the t=check
1468 response, described above.
1469 post-repair-results: a dictionary that describes the state of the file
1470 after any repair was performed. If no repair was
1471 performed, post-repair-results and pre-repair-results
1472 will be the same. This contains exactly the same keys
1473 as the 'results' value of the t=check response,
1476 ``POST $URL?t=start-deep-check&repair=true`` (must add &ophandle=XYZ)
1478 This triggers a recursive walk of all files and directories, performing a
1479 t=check&repair=true on each one.
1481 Like t=start-deep-check without the repair= argument, this can only be
1482 invoked on a directory. An error (400 BAD_REQUEST) will be signalled if it
1483 is invoked on a file. The recursive walker will deal with loops safely.
1485 This accepts the same verify= and add-lease= arguments as
1486 t=start-deep-check. It uses the same ophandle= mechanism as
1487 start-deep-check. When an output=JSON argument is provided, the response
1488 will contain the following keys::
1490 finished: (bool) True if the operation has completed, else False
1491 root-storage-index: a base32-encoded string with the storage index of the
1492 starting point of the deep-check operation
1493 count-objects-checked: count of how many objects were checked
1495 count-objects-healthy-pre-repair: how many of those objects were completely
1496 healthy, before any repair
1497 count-objects-unhealthy-pre-repair: how many were damaged in some way
1498 count-objects-healthy-post-repair: how many of those objects were completely
1499 healthy, after any repair
1500 count-objects-unhealthy-post-repair: how many were damaged in some way
1502 count-repairs-attempted: repairs were attempted on this many objects.
1503 count-repairs-successful: how many repairs resulted in healthy objects
1504 count-repairs-unsuccessful: how many repairs resulted did not results in
1505 completely healthy objects
1506 count-corrupt-shares-pre-repair: how many shares were found to have
1507 corruption, summed over all objects
1508 examined, before any repair
1509 count-corrupt-shares-post-repair: how many shares were found to have
1510 corruption, summed over all objects
1511 examined, after any repair
1512 list-corrupt-shares: a list of "share identifiers", one for each share
1513 that was found to be corrupt (before any repair).
1514 Each share identifier is a list of (serverid,
1515 storage_index, sharenum).
1516 list-remaining-corrupt-shares: like list-corrupt-shares, but mutable shares
1517 that were successfully repaired are not
1518 included. These are shares that need
1519 manual processing. Since immutable shares
1520 cannot be modified by clients, all corruption
1521 in immutable shares will be listed here.
1522 list-unhealthy-files: a list of (pathname, check-results) tuples, for
1523 each file that was not fully healthy. 'pathname' is
1524 relative to the directory on which deep-check was
1525 invoked. The 'check-results' field is the same as
1526 that returned by t=check&repair=true&output=JSON,
1528 stats: a dictionary with the same keys as the t=start-deep-stats command
1531 ``POST $URL?t=stream-deep-check&repair=true``
1533 This triggers a recursive walk of all files and directories, performing a
1534 t=check&repair=true on each one. For each unique object (duplicates are
1535 skipped), a single line of JSON is emitted to the HTTP response channel (or
1536 an error indication). When the walk is complete, a final line of JSON is
1537 emitted which contains the accumulated file-size/count "deep-stats" data.
1539 This emits the same data as t=stream-deep-check (without the repair=true),
1540 except that the "check-results" field is replaced with a
1541 "check-and-repair-results" field, which contains the keys returned by
1542 t=check&repair=true&output=json (i.e. repair-attempted, repair-successful,
1543 pre-repair-results, and post-repair-results). The output does not contain
1544 the summary dictionary that is provied by t=start-deep-check&repair=true
1545 (the one with count-objects-checked and list-unhealthy-files), since the
1546 receiving client is expected to calculate those values itself from the
1547 stream of per-object check-and-repair-results.
1549 Note that the "ERROR:" indication will only be emitted if traversal stops,
1550 which will only occur if an unrecoverable directory is encountered. If a
1551 file or directory repair fails, the traversal will continue, and the repair
1552 failure will be indicated in the JSON data (in the "repair-successful" key).
1554 ``POST $DIRURL?t=start-manifest`` (must add &ophandle=XYZ)
1556 This operation generates a "manfest" of the given directory tree, mostly
1557 for debugging. This is a table of (path, filecap/dircap), for every object
1558 reachable from the starting directory. The path will be slash-joined, and
1559 the filecap/dircap will contain a link to the object in question. This page
1560 gives immediate access to every object in the virtual filesystem subtree.
1562 This operation uses the same ophandle= mechanism as deep-check. The
1563 corresponding /operations/$HANDLE page has three different forms. The
1564 default is output=HTML.
1566 If output=text is added to the query args, the results will be a text/plain
1567 list. The first line is special: it is either "finished: yes" or "finished:
1568 no"; if the operation is not finished, you must periodically reload the
1569 page until it completes. The rest of the results are a plaintext list, with
1570 one file/dir per line, slash-separated, with the filecap/dircap separated
1573 If output=JSON is added to the queryargs, then the results will be a
1574 JSON-formatted dictionary with six keys. Note that because large directory
1575 structures can result in very large JSON results, the full results will not
1576 be available until the operation is complete (i.e. until output["finished"]
1579 finished (bool): if False then you must reload the page until True
1580 origin_si (base32 str): the storage index of the starting point
1581 manifest: list of (path, cap) tuples, where path is a list of strings.
1582 verifycaps: list of (printable) verify cap strings
1583 storage-index: list of (base32) storage index strings
1584 stats: a dictionary with the same keys as the t=start-deep-stats command
1587 ``POST $DIRURL?t=start-deep-size`` (must add &ophandle=XYZ)
1589 This operation generates a number (in bytes) containing the sum of the
1590 filesize of all directories and immutable files reachable from the given
1591 directory. This is a rough lower bound of the total space consumed by this
1592 subtree. It does not include space consumed by mutable files, nor does it
1593 take expansion or encoding overhead into account. Later versions of the
1594 code may improve this estimate upwards.
1596 The /operations/$HANDLE status output consists of two lines of text::
1601 ``POST $DIRURL?t=start-deep-stats`` (must add &ophandle=XYZ)
1603 This operation performs a recursive walk of all files and directories
1604 reachable from the given directory, and generates a collection of
1605 statistics about those objects.
1607 The result (obtained from the /operations/$OPHANDLE page) is a
1608 JSON-serialized dictionary with the following keys (note that some of these
1609 keys may be missing until 'finished' is True)::
1611 finished: (bool) True if the operation has finished, else False
1612 count-immutable-files: count of how many CHK files are in the set
1613 count-mutable-files: same, for mutable files (does not include directories)
1614 count-literal-files: same, for LIT files (data contained inside the URI)
1615 count-files: sum of the above three
1616 count-directories: count of directories
1617 count-unknown: count of unrecognized objects (perhaps from the future)
1618 size-immutable-files: total bytes for all CHK files in the set, =deep-size
1619 size-mutable-files (TODO): same, for current version of all mutable files
1620 size-literal-files: same, for LIT files
1621 size-directories: size of directories (includes size-literal-files)
1622 size-files-histogram: list of (minsize, maxsize, count) buckets,
1623 with a histogram of filesizes, 5dB/bucket,
1624 for both literal and immutable files
1625 largest-directory: number of children in the largest directory
1626 largest-immutable-file: number of bytes in the largest CHK file
1628 size-mutable-files is not implemented, because it would require extra
1629 queries to each mutable file to get their size. This may be implemented in
1632 Assuming no sharing, the basic space consumed by a single root directory is
1633 the sum of size-immutable-files, size-mutable-files, and size-directories.
1634 The actual disk space used by the shares is larger, because of the
1635 following sources of overhead::
1638 expansion due to erasure coding
1639 share management data (leases)
1640 backend (ext3) minimum block size
1642 ``POST $URL?t=stream-manifest``
1644 This operation performs a recursive walk of all files and directories
1645 reachable from the given starting point. For each such unique object
1646 (duplicates are skipped), a single line of JSON is emitted to the HTTP
1647 response channel (or an error indication, see below). When the walk is
1648 complete, a final line of JSON is emitted which contains the accumulated
1649 file-size/count "deep-stats" data.
1651 A CLI tool can split the response stream on newlines into "response units",
1652 and parse each response unit as JSON. Each such parsed unit will be a
1653 dictionary, and will contain at least the "type" key: a string, one of
1654 "file", "directory", or "stats".
1656 For all units that have a type of "file" or "directory", the dictionary will
1657 contain the following keys::
1659 "path": a list of strings, with the path that is traversed to reach the
1661 "cap": a write-cap URI for the file or directory, if available, else a
1663 "verifycap": a verify-cap URI for the file or directory
1664 "repaircap": an URI for the weakest cap that can still be used to repair
1666 "storage-index": a base32 storage index for the object
1668 Note that non-distributed files (i.e. LIT files) will have values of None
1669 for verifycap, repaircap, and storage-index, since these files can neither
1670 be verified nor repaired, and are not stored on the storage servers.
1672 The last unit in the stream will have a type of "stats", and will contain
1673 the keys described in the "start-deep-stats" operation, below.
1675 If any errors occur during the traversal (specifically if a directory is
1676 unrecoverable, such that further traversal is not possible), an error
1677 indication is written to the response body, instead of the usual line of
1678 JSON. This error indication line will begin with the string "ERROR:" (in all
1679 caps), and contain a summary of the error on the rest of the line. The
1680 remaining lines of the response body will be a python exception. The client
1681 application should look for the ERROR: and stop processing JSON as soon as
1682 it is seen. The line just before the ERROR: will describe the directory that
1683 was untraversable, since the manifest entry is emitted to the HTTP response
1684 body before the child is traversed.
1690 The portion of the web namespace that begins with "/uri" (and "/named") is
1691 dedicated to giving users (both humans and programs) access to the Tahoe
1692 virtual filesystem. The rest of the namespace provides status information
1693 about the state of the Tahoe node.
1695 ``GET /`` (the root page)
1697 This is the "Welcome Page", and contains a few distinct sections::
1699 Node information: library versions, local nodeid, services being provided.
1701 Filesystem Access Forms: create a new directory, view a file/directory by
1702 URI, upload a file (unlinked), download a file by
1705 Grid Status: introducer information, helper information, connected storage
1710 This page lists all active uploads and downloads, and contains a short list
1711 of recent upload/download operations. Each operation has a link to a page
1712 that describes file sizes, servers that were involved, and the time consumed
1713 in each phase of the operation.
1715 A GET of /status/?t=json will contain a machine-readable subset of the same
1716 data. It returns a JSON-encoded dictionary. The only key defined at this
1717 time is "active", with a value that is a list of operation dictionaries, one
1718 for each active operation. Once an operation is completed, it will no longer
1719 appear in data["active"] .
1721 Each op-dict contains a "type" key, one of "upload", "download",
1722 "mapupdate", "publish", or "retrieve" (the first two are for immutable
1723 files, while the latter three are for mutable files and directories).
1725 The "upload" op-dict will contain the following keys::
1727 type (string): "upload"
1728 storage-index-string (string): a base32-encoded storage index
1729 total-size (int): total size of the file
1730 status (string): current status of the operation
1731 progress-hash (float): 1.0 when the file has been hashed
1732 progress-ciphertext (float): 1.0 when the file has been encrypted.
1733 progress-encode-push (float): 1.0 when the file has been encoded and
1734 pushed to the storage servers. For helper
1735 uploads, the ciphertext value climbs to 1.0
1736 first, then encoding starts. For unassisted
1737 uploads, ciphertext and encode-push progress
1738 will climb at the same pace.
1740 The "download" op-dict will contain the following keys::
1742 type (string): "download"
1743 storage-index-string (string): a base32-encoded storage index
1744 total-size (int): total size of the file
1745 status (string): current status of the operation
1746 progress (float): 1.0 when the file has been fully downloaded
1748 Front-ends which want to report progress information are advised to simply
1749 average together all the progress-* indicators. A slightly more accurate
1750 value can be found by ignoring the progress-hash value (since the current
1751 implementation hashes synchronously, so clients will probably never see
1752 progress-hash!=1.0).
1754 ``GET /provisioning/``
1756 This page provides a basic tool to predict the likely storage and bandwidth
1757 requirements of a large Tahoe grid. It provides forms to input things like
1758 total number of users, number of files per user, average file size, number
1759 of servers, expansion ratio, hard drive failure rate, etc. It then provides
1760 numbers like how many disks per server will be needed, how many read
1761 operations per second should be expected, and the likely MTBF for files in
1762 the grid. This information is very preliminary, and the model upon which it
1763 is based still needs a lot of work.
1765 ``GET /helper_status/``
1767 If the node is running a helper (i.e. if [helper]enabled is set to True in
1768 tahoe.cfg), then this page will provide a list of all the helper operations
1769 currently in progress. If "?t=json" is added to the URL, it will return a
1770 JSON-formatted list of helper statistics, which can then be used to produce
1771 graphs to indicate how busy the helper is.
1773 ``GET /statistics/``
1775 This page provides "node statistics", which are collected from a variety of
1778 load_monitor: every second, the node schedules a timer for one second in
1779 the future, then measures how late the subsequent callback
1780 is. The "load_average" is this tardiness, measured in
1781 seconds, averaged over the last minute. It is an indication
1782 of a busy node, one which is doing more work than can be
1783 completed in a timely fashion. The "max_load" value is the
1784 highest value that has been seen in the last 60 seconds.
1786 cpu_monitor: every minute, the node uses time.clock() to measure how much
1787 CPU time it has used, and it uses this value to produce
1788 1min/5min/15min moving averages. These values range from 0%
1789 (0.0) to 100% (1.0), and indicate what fraction of the CPU
1790 has been used by the Tahoe node. Not all operating systems
1791 provide meaningful data to time.clock(): they may report 100%
1792 CPU usage at all times.
1794 uploader: this counts how many immutable files (and bytes) have been
1795 uploaded since the node was started
1797 downloader: this counts how many immutable files have been downloaded
1798 since the node was started
1800 publishes: this counts how many mutable files (including directories) have
1801 been modified since the node was started
1803 retrieves: this counts how many mutable files (including directories) have
1804 been read since the node was started
1806 There are other statistics that are tracked by the node. The "raw stats"
1807 section shows a formatted dump of all of them.
1809 By adding "?t=json" to the URL, the node will return a JSON-formatted
1810 dictionary of stats values, which can be used by other tools to produce
1811 graphs of node behavior. The misc/munin/ directory in the source
1812 distribution provides some tools to produce these graphs.
1814 ``GET /`` (introducer status)
1816 For Introducer nodes, the welcome page displays information about both
1817 clients and servers which are connected to the introducer. Servers make
1818 "service announcements", and these are listed in a table. Clients will
1819 subscribe to hear about service announcements, and these subscriptions are
1820 listed in a separate table. Both tables contain information about what
1821 version of Tahoe is being run by the remote node, their advertised and
1822 outbound IP addresses, their nodeid and nickname, and how long they have
1825 By adding "?t=json" to the URL, the node will return a JSON-formatted
1826 dictionary of stats values, which can be used to produce graphs of connected
1827 clients over time. This dictionary has the following keys::
1829 ["subscription_summary"] : a dictionary mapping service name (like
1830 "storage") to an integer with the number of
1831 clients that have subscribed to hear about that
1833 ["announcement_summary"] : a dictionary mapping service name to an integer
1834 with the number of servers which are announcing
1836 ["announcement_distinct_hosts"] : a dictionary mapping service name to an
1837 integer which represents the number of
1838 distinct hosts that are providing that
1839 service. If two servers have announced
1840 FURLs which use the same hostnames (but
1841 different ports and tubids), they are
1842 considered to be on the same host.
1845 Static Files in /public_html
1846 ============================
1848 The web-API server will take any request for a URL that starts with /static
1849 and serve it from a configurable directory which defaults to
1850 $BASEDIR/public_html . This is configured by setting the "[node]web.static"
1851 value in $BASEDIR/tahoe.cfg . If this is left at the default value of
1852 "public_html", then http://localhost:3456/static/subdir/foo.html will be
1853 served with the contents of the file $BASEDIR/public_html/subdir/foo.html .
1855 This can be useful to serve a javascript application which provides a
1856 prettier front-end to the rest of the Tahoe web-API.
1859 Safety and Security Issues -- Names vs. URIs
1860 ============================================
1862 Summary: use explicit file- and dir- caps whenever possible, to reduce the
1863 potential for surprises when the filesystem structure is changed.
1865 Tahoe provides a mutable filesystem, but the ways that the filesystem can
1866 change are limited. The only thing that can change is that the mapping from
1867 child names to child objects that each directory contains can be changed by
1868 adding a new child name pointing to an object, removing an existing child name,
1869 or changing an existing child name to point to a different object.
1871 Obviously if you query Tahoe for information about the filesystem and then act
1872 to change the filesystem (such as by getting a listing of the contents of a
1873 directory and then adding a file to the directory), then the filesystem might
1874 have been changed after you queried it and before you acted upon it. However,
1875 if you use the URI instead of the pathname of an object when you act upon the
1876 object, then the only change that can happen is if the object is a directory
1877 then the set of child names it has might be different. If, on the other hand,
1878 you act upon the object using its pathname, then a different object might be in
1879 that place, which can result in more kinds of surprises.
1881 For example, suppose you are writing code which recursively downloads the
1882 contents of a directory. The first thing your code does is fetch the listing
1883 of the contents of the directory. For each child that it fetched, if that
1884 child is a file then it downloads the file, and if that child is a directory
1885 then it recurses into that directory. Now, if the download and the recurse
1886 actions are performed using the child's name, then the results might be
1887 wrong, because for example a child name that pointed to a sub-directory when
1888 you listed the directory might have been changed to point to a file (in which
1889 case your attempt to recurse into it would result in an error and the file
1890 would be skipped), or a child name that pointed to a file when you listed the
1891 directory might now point to a sub-directory (in which case your attempt to
1892 download the child would result in a file containing HTML text describing the
1895 If your recursive algorithm uses the uri of the child instead of the name of
1896 the child, then those kinds of mistakes just can't happen. Note that both the
1897 child's name and the child's URI are included in the results of listing the
1898 parent directory, so it isn't any harder to use the URI for this purpose.
1900 The read and write caps in a given directory node are separate URIs, and
1901 can't be assumed to point to the same object even if they were retrieved in
1902 the same operation (although the web-API server attempts to ensure this
1903 in most cases). If you need to rely on that property, you should explicitly
1904 verify it. More generally, you should not make assumptions about the
1905 internal consistency of the contents of mutable directories. As a result
1906 of the signatures on mutable object versions, it is guaranteed that a given
1907 version was written in a single update, but -- as in the case of a file --
1908 the contents may have been chosen by a malicious writer in a way that is
1909 designed to confuse applications that rely on their consistency.
1911 In general, use names if you want "whatever object (whether file or
1912 directory) is found by following this name (or sequence of names) when my
1913 request reaches the server". Use URIs if you want "this particular object".
1919 Tahoe uses both mutable and immutable files. Mutable files can be created
1920 explicitly by doing an upload with ?mutable=true added, or implicitly by
1921 creating a new directory (since a directory is just a special way to
1922 interpret a given mutable file).
1924 Mutable files suffer from the same consistency-vs-availability tradeoff that
1925 all distributed data storage systems face. It is not possible to
1926 simultaneously achieve perfect consistency and perfect availability in the
1927 face of network partitions (servers being unreachable or faulty).
1929 Tahoe tries to achieve a reasonable compromise, but there is a basic rule in
1930 place, known as the Prime Coordination Directive: "Don't Do That". What this
1931 means is that if write-access to a mutable file is available to several
1932 parties, then those parties are responsible for coordinating their activities
1933 to avoid multiple simultaneous updates. This could be achieved by having
1934 these parties talk to each other and using some sort of locking mechanism, or
1935 by serializing all changes through a single writer.
1937 The consequences of performing uncoordinated writes can vary. Some of the
1938 writers may lose their changes, as somebody else wins the race condition. In
1939 many cases the file will be left in an "unhealthy" state, meaning that there
1940 are not as many redundant shares as we would like (reducing the reliability
1941 of the file against server failures). In the worst case, the file can be left
1942 in such an unhealthy state that no version is recoverable, even the old ones.
1943 It is this small possibility of data loss that prompts us to issue the Prime
1944 Coordination Directive.
1946 Tahoe nodes implement internal serialization to make sure that a single Tahoe
1947 node cannot conflict with itself. For example, it is safe to issue two
1948 directory modification requests to a single tahoe node's web-API server at the
1949 same time, because the Tahoe node will internally delay one of them until
1950 after the other has finished being applied. (This feature was introduced in
1951 Tahoe-1.1; back with Tahoe-1.0 the web client was responsible for serializing
1952 web requests themselves).
1954 For more details, please see the "Consistency vs Availability" and "The Prime
1955 Coordination Directive" sections of `mutable.rst <../specifications/mutable.rst>`_.
1958 .. [1] URLs and HTTP and UTF-8, Oh My
1960 HTTP does not provide a mechanism to specify the character set used to
1961 encode non-ASCII names in URLs
1962 (`RFC3986#2.1 <http://tools.ietf.org/html/rfc3986#section-2.1>`_).
1963 We prefer the convention that the ``filename=`` argument shall be a
1964 URL-escaped UTF-8 encoded Unicode string.
1965 For example, suppose we want to provoke the server into using a filename of
1966 "f i a n c e-acute e" (i.e. f i a n c U+00E9 e). The UTF-8 encoding of this
1967 is 0x66 0x69 0x61 0x6e 0x63 0xc3 0xa9 0x65 (or "fianc\\xC3\\xA9e", as python's
1968 ``repr()`` function would show). To encode this into a URL, the non-printable
1969 characters must be escaped with the urlencode ``%XX`` mechanism, giving
1970 us "fianc%C3%A9e". Thus, the first line of the HTTP request will be
1971 "``GET /uri/CAP...?save=true&filename=fianc%C3%A9e HTTP/1.1``". Not all
1972 browsers provide this: IE7 by default uses the Latin-1 encoding, which is
1973 "fianc%E9e" (although it has a configuration option to send URLs as UTF-8).
1975 The response header will need to indicate a non-ASCII filename. The actual
1976 mechanism to do this is not clear. For ASCII filenames, the response header
1979 Content-Disposition: attachment; filename="english.txt"
1981 If Tahoe were to enforce the UTF-8 convention, it would need to decode the
1982 URL argument into a Unicode string, and then encode it back into a sequence
1983 of bytes when creating the response header. One possibility would be to use
1984 unencoded UTF-8. Developers suggest that IE7 might accept this::
1986 #1: Content-Disposition: attachment; filename="fianc\xC3\xA9e"
1987 (note, the last four bytes of that line, not including the newline, are
1988 0xC3 0xA9 0x65 0x22)
1990 `RFC2231#4 <http://tools.ietf.org/html/rfc2231#section-4>`_
1991 (dated 1997): suggests that the following might work, and
1992 `some developers have reported <http://markmail.org/message/dsjyokgl7hv64ig3>`_
1993 that it is supported by Firefox (but not IE7)::
1995 #2: Content-Disposition: attachment; filename*=utf-8''fianc%C3%A9e
1997 My reading of `RFC2616#19.5.1 <http://tools.ietf.org/html/rfc2616#section-19.5.1>`_
1998 (which defines Content-Disposition) says that the filename= parameter is
1999 defined to be wrapped in quotes (presumably to allow spaces without breaking
2000 the parsing of subsequent parameters), which would give us::
2002 #3: Content-Disposition: attachment; filename*=utf-8''"fianc%C3%A9e"
2004 However this is contrary to the examples in the email thread listed above.
2006 Developers report that IE7 (when it is configured for UTF-8 URL encoding,
2007 which is not the default in Asian countries), will accept::
2009 #4: Content-Disposition: attachment; filename=fianc%C3%A9e
2011 However, for maximum compatibility, Tahoe simply copies bytes from the URL
2012 into the response header, rather than enforcing the UTF-8 convention. This
2013 means it does not try to decode the filename from the URL argument, nor does
2014 it encode the filename into the response header.