1 .. -*- coding: utf-8-with-signature -*-
3 ==========================
4 The Tahoe REST-ful Web API
5 ==========================
7 1. `Enabling the web-API port`_
8 2. `Basic Concepts: GET, PUT, DELETE, POST`_
13 4. `Slow Operations, Progress, and Cancelling`_
14 5. `Programmatic Operations`_
17 2. `Writing/Uploading a File`_
18 3. `Creating a New Directory`_
19 4. `Getting Information About a File Or Directory (as JSON)`_
20 5. `Attaching an Existing File or Directory by its read- or write-cap`_
21 6. `Adding Multiple Files or Directories to a Parent Directory at Once`_
22 7. `Unlinking a File or Directory`_
24 6. `Browser Operations: Human-Oriented Interfaces`_
26 1. `Viewing a Directory (as HTML)`_
27 2. `Viewing/Downloading a File`_
28 3. `Getting Information About a File Or Directory (as HTML)`_
29 4. `Creating a Directory`_
30 5. `Uploading a File`_
31 6. `Attaching an Existing File Or Directory (by URI)`_
32 7. `Unlinking a Child`_
33 8. `Renaming a Child`_
34 9. `Relinking ("Moving") a Child`_
35 10. `Other Utilities`_
36 11. `Debugging and Testing Features`_
38 7. `Other Useful Pages`_
39 8. `Static Files in /public_html`_
40 9. `Safety and Security Issues -- Names vs. URIs`_
41 10. `Concurrency Issues`_
42 11. `Access Blacklist`_
45 Enabling the web-API port
46 =========================
48 Every Tahoe node is capable of running a built-in HTTP server. To enable
49 this, just write a port number into the "[node]web.port" line of your node's
50 tahoe.cfg file. For example, writing "web.port = 3456" into the "[node]"
51 section of $NODEDIR/tahoe.cfg will cause the node to run a webserver on port
54 This string is actually a Twisted "strports" specification, meaning you can
55 get more control over the interface to which the server binds by supplying
56 additional arguments. For more details, see the documentation on
57 `twisted.application.strports`_.
59 Writing "tcp:3456:interface=127.0.0.1" into the web.port line does the same
60 but binds to the loopback interface, ensuring that only the programs on the
61 local host can connect. Using "ssl:3456:privateKey=mykey.pem:certKey=cert.pem"
64 This webport can be set when the node is created by passing a --webport
65 option to the 'tahoe create-node' command. By default, the node listens on
66 port 3456, on the loopback (127.0.0.1) interface.
68 .. _twisted.application.strports: https://twistedmatrix.com/documents/current/api/twisted.application.strports.html
71 Basic Concepts: GET, PUT, DELETE, POST
72 ======================================
74 As described in `docs/architecture.rst`_, each file and directory in a Tahoe
75 virtual filesystem is referenced by an identifier that combines the
76 designation of the object with the authority to do something with it (such as
77 read or modify the contents). This identifier is called a "read-cap" or
78 "write-cap", depending upon whether it enables read-only or read-write
79 access. These "caps" are also referred to as URIs (which may be confusing
80 because they are not currently RFC3986_-compliant URIs).
82 The Tahoe web-based API is "REST-ful", meaning it implements the concepts of
83 "REpresentational State Transfer": the original scheme by which the World
84 Wide Web was intended to work. Each object (file or directory) is referenced
85 by a URL that includes the read- or write- cap. HTTP methods (GET, PUT, and
86 DELETE) are used to manipulate these objects. You can think of the URL as a
87 noun, and the method as a verb.
89 In REST, the GET method is used to retrieve information about an object, or
90 to retrieve some representation of the object itself. When the object is a
91 file, the basic GET method will simply return the contents of that file.
92 Other variations (generally implemented by adding query parameters to the
93 URL) will return information about the object, such as metadata. GET
94 operations are required to have no side-effects.
96 PUT is used to upload new objects into the filesystem, or to replace an
97 existing link or the contents of a mutable file. DELETE is used to unlink
98 objects from directories. Both PUT and DELETE are required to be idempotent:
99 performing the same operation multiple times must have the same side-effects
100 as only performing it once.
102 POST is used for more complicated actions that cannot be expressed as a GET,
103 PUT, or DELETE. POST operations can be thought of as a method call: sending
104 some message to the object referenced by the URL. In Tahoe, POST is also used
105 for operations that must be triggered by an HTML form (including upload and
106 unlinking), because otherwise a regular web browser has no way to accomplish
107 these tasks. In general, everything that can be done with a PUT or DELETE can
108 also be done with a POST.
110 Tahoe's web API is designed for two different kinds of consumer. The first is
111 a program that needs to manipulate the virtual file system. Such programs are
112 expected to use the RESTful interface described above. The second is a human
113 using a standard web browser to work with the filesystem. This user is given
114 a series of HTML pages with links to download files, and forms that use POST
115 actions to upload, rename, and unlink files.
117 When an error occurs, the HTTP response code will be set to an appropriate
118 400-series code (like 404 Not Found for an unknown childname, or 400 Bad Request
119 when the parameters to a web-API operation are invalid), and the HTTP response
120 body will usually contain a few lines of explanation as to the cause of the
121 error and possible responses. Unusual exceptions may result in a 500 Internal
122 Server Error as a catch-all, with a default response body containing
123 a Nevow-generated HTML-ized representation of the Python exception stack trace
124 that caused the problem. CLI programs which want to copy the response body to
125 stderr should provide an "Accept: text/plain" header to their requests to get
126 a plain text stack trace instead. If the Accept header contains ``*/*``, or
127 ``text/*``, or text/html (or if there is no Accept header), HTML tracebacks will
130 .. _RFC3986: https://tools.ietf.org/html/rfc3986
131 .. _docs/architecture.rst: ../architecture.rst
137 Tahoe uses a variety of read- and write- caps to identify files and
138 directories. The most common of these is the "immutable file read-cap", which
139 is used for most uploaded files. These read-caps look like the following::
141 URI:CHK:ime6pvkaxuetdfah2p2f35pe54:4btz54xk3tew6nd4y2ojpxj4m6wxjqqlwnztgre6gnjgtucd5r4a:3:10:202
143 The next most common is a "directory write-cap", which provides both read and
144 write access to a directory, and look like this::
146 URI:DIR2:djrdkfawoqihigoett4g6auz6a:jx5mplfpwexnoqff7y5e4zjus4lidm76dcuarpct7cckorh2dpgq
148 There are also "directory read-caps", which start with "URI:DIR2-RO:", and
149 give read-only access to a directory. Finally there are also mutable file
150 read- and write- caps, which start with "URI:SSK", and give access to mutable
153 (Later versions of Tahoe will make these strings shorter, and will remove the
154 unfortunate colons, which must be escaped when these caps are embedded in
157 To refer to any Tahoe object through the web API, you simply need to combine
158 a prefix (which indicates the HTTP server to use) with the cap (which
159 indicates which object inside that server to access). Since the default Tahoe
160 webport is 3456, the most common prefix is one that will use a local node
161 listening on this port::
163 http://127.0.0.1:3456/uri/ + $CAP
165 So, to access the directory named above, the URL would be::
167 http://127.0.0.1:3456/uri/URI%3ADIR2%3Adjrdkfawoqihigoett4g6auz6a%3Ajx5mplfpwexnoqff7y5e4zjus4lidm76dcuarpct7cckorh2dpgq/
169 (note that the colons in the directory-cap are url-encoded into "%3A"
172 Likewise, to access the file named above, use::
174 http://127.0.0.1:3456/uri/URI%3ACHK%3Aime6pvkaxuetdfah2p2f35pe54%3A4btz54xk3tew6nd4y2ojpxj4m6wxjqqlwnztgre6gnjgtucd5r4a%3A3%3A10%3A202
176 In the rest of this document, we'll use "$DIRCAP" as shorthand for a read-cap
177 or write-cap that refers to a directory, and "$FILECAP" to abbreviate a cap
178 that refers to a file (whether mutable or immutable). So those URLs above can
181 http://127.0.0.1:3456/uri/$DIRCAP/
182 http://127.0.0.1:3456/uri/$FILECAP
184 The operation summaries below will abbreviate these further, by eliding the
185 server prefix. They will be displayed like this::
190 /cap can be used as a synonym for /uri. If interoperability with older web-API
191 servers is required, /uri should be used.
196 Tahoe directories contain named child entries, just like directories in a regular
197 local filesystem. These child entries, called "dirnodes", consist of a name,
198 metadata, a write slot, and a read slot. The write and read slots normally contain
199 a write-cap and read-cap referring to the same object, which can be either a file
200 or a subdirectory. The write slot may be empty (actually, both may be empty,
201 but that is unusual).
203 If you have a Tahoe URL that refers to a directory, and want to reference a
204 named child inside it, just append the child name to the URL. For example, if
205 our sample directory contains a file named "welcome.txt", we can refer to
208 http://127.0.0.1:3456/uri/$DIRCAP/welcome.txt
210 (or http://127.0.0.1:3456/uri/URI%3ADIR2%3Adjrdkfawoqihigoett4g6auz6a%3Ajx5mplfpwexnoqff7y5e4zjus4lidm76dcuarpct7cckorh2dpgq/welcome.txt)
212 Multiple levels of subdirectories can be handled this way::
214 http://127.0.0.1:3456/uri/$DIRCAP/tahoe-source/docs/architecture.rst
216 In this document, when we need to refer to a URL that references a file using
217 this child-of-some-directory format, we'll use the following string::
219 /uri/$DIRCAP/[SUBDIRS../]FILENAME
221 The "[SUBDIRS../]" part means that there are zero or more (optional)
222 subdirectory names in the middle of the URL. The "FILENAME" at the end means
223 that this whole URL refers to a file of some sort, rather than to a
226 When we need to refer specifically to a directory in this way, we'll write::
228 /uri/$DIRCAP/[SUBDIRS../]SUBDIR
231 Note that all components of pathnames in URLs are required to be UTF-8
232 encoded, so "resume.doc" (with an acute accent on both E's) would be accessed
235 http://127.0.0.1:3456/uri/$DIRCAP/r%C3%A9sum%C3%A9.doc
237 Also note that the filenames inside upload POST forms are interpreted using
238 whatever character set was provided in the conventional '_charset' field, and
239 defaults to UTF-8 if not otherwise specified. The JSON representation of each
240 directory contains native Unicode strings. Tahoe directories are specified to
241 contain Unicode filenames, and cannot contain binary strings that are not
242 representable as such.
244 All Tahoe operations that refer to existing files or directories must include
245 a suitable read- or write- cap in the URL: the web-API server won't add one
246 for you. If you don't know the cap, you can't access the file. This allows
247 the security properties of Tahoe caps to be extended across the web-API
251 Slow Operations, Progress, and Cancelling
252 =========================================
254 Certain operations can be expected to take a long time. The "t=deep-check",
255 described below, will recursively visit every file and directory reachable
256 from a given starting point, which can take minutes or even hours for
257 extremely large directory structures. A single long-running HTTP request is a
258 fragile thing: proxies, NAT boxes, browsers, and users may all grow impatient
259 with waiting and give up on the connection.
261 For this reason, long-running operations have an "operation handle", which
262 can be used to poll for status/progress messages while the operation
263 proceeds. This handle can also be used to cancel the operation. These handles
264 are created by the client, and passed in as a an "ophandle=" query argument
265 to the POST or PUT request which starts the operation. The following
266 operations can then be used to retrieve status:
268 ``GET /operations/$HANDLE?output=HTML (with or without t=status)``
270 ``GET /operations/$HANDLE?output=JSON (same)``
272 These two retrieve the current status of the given operation. Each operation
273 presents a different sort of information, but in general the page retrieved
276 * whether the operation is complete, or if it is still running
277 * how much of the operation is complete, and how much is left, if possible
279 Note that the final status output can be quite large: a deep-manifest of a
280 directory structure with 300k directories and 200k unique files is about
281 275MB of JSON, and might take two minutes to generate. For this reason, the
282 full status is not provided until the operation has completed.
284 The HTML form will include a meta-refresh tag, which will cause a regular
285 web browser to reload the status page about 60 seconds later. This tag will
286 be removed once the operation has completed.
288 There may be more status information available under
289 /operations/$HANDLE/$ETC : i.e., the handle forms the root of a URL space.
291 ``POST /operations/$HANDLE?t=cancel``
293 This terminates the operation, and returns an HTML page explaining what was
294 cancelled. If the operation handle has already expired (see below), this
295 POST will return a 404, which indicates that the operation is no longer
296 running (either it was completed or terminated). The response body will be
297 the same as a GET /operations/$HANDLE on this operation handle, and the
298 handle will be expired immediately afterwards.
300 The operation handle will eventually expire, to avoid consuming an unbounded
301 amount of memory. The handle's time-to-live can be reset at any time, by
302 passing a retain-for= argument (with a count of seconds) to either the
303 initial POST that starts the operation, or the subsequent GET request which
304 asks about the operation. For example, if a 'GET
305 /operations/$HANDLE?output=JSON&retain-for=600' query is performed, the
306 handle will remain active for 600 seconds (10 minutes) after the GET was
309 In addition, if the GET includes a release-after-complete=True argument, and
310 the operation has completed, the operation handle will be released
313 If a retain-for= argument is not used, the default handle lifetimes are:
315 * handles will remain valid at least until their operation finishes
316 * uncollected handles for finished operations (i.e. handles for
317 operations that have finished but for which the GET page has not been
318 accessed since completion) will remain valid for four days, or for
319 the total time consumed by the operation, whichever is greater.
320 * collected handles (i.e. the GET page has been retrieved at least once
321 since the operation completed) will remain valid for one day.
323 Many "slow" operations can begin to use unacceptable amounts of memory when
324 operating on large directory structures. The memory usage increases when the
325 ophandle is polled, as the results must be copied into a JSON string, sent
326 over the wire, then parsed by a client. So, as an alternative, many "slow"
327 operations have streaming equivalents. These equivalents do not use operation
328 handles. Instead, they emit line-oriented status results immediately. Client
329 code can cancel the operation by simply closing the HTTP connection.
332 Programmatic Operations
333 =======================
335 Now that we know how to build URLs that refer to files and directories in a
336 Tahoe virtual filesystem, what sorts of operations can we do with those URLs?
337 This section contains a catalog of GET, PUT, DELETE, and POST operations that
338 can be performed on these URLs. This set of operations are aimed at programs
339 that use HTTP to communicate with a Tahoe node. A later section describes
340 operations that are intended for web browsers.
346 ``GET /uri/$FILECAP``
348 ``GET /uri/$DIRCAP/[SUBDIRS../]FILENAME``
350 This will retrieve the contents of the given file. The HTTP response body
351 will contain the sequence of bytes that make up the file. The "Range:"
352 header can be used to restrict which portions of the file are returned (see
353 RFC 2616 section 14.35.1 "Byte Ranges"), however Tahoe only supports a
354 single "bytes" range and never provides a `multipart/byteranges` response.
356 To view files in a web browser, you may want more control over the
357 Content-Type and Content-Disposition headers. Please see the next section
358 "Browser Operations", for details on how to modify these URLs for that
362 Writing/Uploading a File
363 ------------------------
365 ``PUT /uri/$FILECAP``
367 ``PUT /uri/$DIRCAP/[SUBDIRS../]FILENAME``
369 Upload a file, using the data from the HTTP request body, and add whatever
370 child links and subdirectories are necessary to make the file available at
371 the given location. Once this operation succeeds, a GET on the same URL will
372 retrieve the same contents that were just uploaded. This will create any
373 necessary intermediate subdirectories.
375 To use the /uri/$FILECAP form, $FILECAP must be a write-cap for a mutable file.
377 In the /uri/$DIRCAP/[SUBDIRS../]FILENAME form, if the target file is a
378 writeable mutable file, that file's contents will be overwritten
379 in-place. If it is a read-cap for a mutable file, an error will occur.
380 If it is an immutable file, the old file will be discarded, and a new
381 one will be put in its place. If the target file is a writable mutable
382 file, you may also specify an "offset" parameter -- a byte offset that
383 determines where in the mutable file the data from the HTTP request
384 body is placed. This operation is relatively efficient for MDMF mutable
385 files, and is relatively inefficient (but still supported) for SDMF
386 mutable files. If no offset parameter is specified, then the entire
387 file is replaced with the data from the HTTP request body. For an
388 immutable file, the "offset" parameter is not valid.
390 When creating a new file, you can control the type of file created by
391 specifying a format= argument in the query string. format=MDMF creates an
392 MDMF mutable file. format=SDMF creates an SDMF mutable file. format=CHK
393 creates an immutable file. The value of the format argument is
394 case-insensitive. If no format is specified, the newly-created file will be
395 immutable (but see below).
397 For compatibility with previous versions of Tahoe-LAFS, the web-API will
398 also accept a mutable=true argument in the query string. If mutable=true is
399 given, then the new file will be mutable, and its format will be the default
400 mutable file format, as configured by the [client]mutable.format option of
401 tahoe.cfg on the Tahoe-LAFS node hosting the webapi server. Use of
402 mutable=true is discouraged; new code should use format= instead of
403 mutable=true (unless it needs to be compatible with web-API servers older
404 than v1.9.0). If neither format= nor mutable=true are given, the
405 newly-created file will be immutable.
407 This returns the file-cap of the resulting file. If a new file was created
408 by this method, the HTTP response code (as dictated by rfc2616) will be set
409 to 201 CREATED. If an existing file was replaced or modified, the response
412 Note that the 'curl -T localfile http://127.0.0.1:3456/uri/$DIRCAP/foo.txt'
413 command can be used to invoke this operation.
417 This uploads a file, and produces a file-cap for the contents, but does not
418 attach the file into the filesystem. No directories will be modified by
419 this operation. The file-cap is returned as the body of the HTTP response.
421 This method accepts format= and mutable=true as query string arguments, and
422 interprets those arguments in the same way as the linked forms of PUT
423 described immediately above.
425 Creating a New Directory
426 ------------------------
428 ``POST /uri?t=mkdir``
432 Create a new empty directory and return its write-cap as the HTTP response
433 body. This does not make the newly created directory visible from the
434 filesystem. The "PUT" operation is provided for backwards compatibility:
435 new code should use POST.
437 This supports a format= argument in the query string. The format=
438 argument, if specified, controls the format of the directory. format=MDMF
439 indicates that the directory should be stored as an MDMF file; format=SDMF
440 indicates that the directory should be stored as an SDMF file. The value of
441 the format= argument is case-insensitive. If no format= argument is
442 given, the directory's format is determined by the default mutable file
443 format, as configured on the Tahoe-LAFS node responding to the request.
445 ``POST /uri?t=mkdir-with-children``
447 Create a new directory, populated with a set of child nodes, and return its
448 write-cap as the HTTP response body. The new directory is not attached to
449 any other directory: the returned write-cap is the only reference to it.
451 The format of the directory can be controlled with the format= argument in
452 the query string, as described above.
454 Initial children are provided as the body of the POST form (this is more
455 efficient than doing separate mkdir and set_children operations). If the
456 body is empty, the new directory will be empty. If not empty, the body will
457 be interpreted as a UTF-8 JSON-encoded dictionary of children with which the
458 new directory should be populated, using the same format as would be
459 returned in the 'children' value of the t=json GET request, described below.
460 Each dictionary key should be a child name, and each value should be a list
461 of [TYPE, PROPDICT], where PROPDICT contains "rw_uri", "ro_uri", and
462 "metadata" keys (all others are ignored). For example, the PUT request body
466 "Fran\u00e7ais": [ "filenode", {
467 "ro_uri": "URI:CHK:...",
469 "ctime": 1202777696.7564139,
470 "mtime": 1202777696.7564139,
472 "linkcrtime": 1202777696.7564139,
473 "linkmotime": 1202777696.7564139
475 "subdir": [ "dirnode", {
476 "rw_uri": "URI:DIR2:...",
477 "ro_uri": "URI:DIR2-RO:...",
479 "ctime": 1202778102.7589991,
480 "mtime": 1202778111.2160511,
482 "linkcrtime": 1202777696.7564139,
483 "linkmotime": 1202777696.7564139
487 For forward-compatibility, a mutable directory can also contain caps in
488 a format that is unknown to the web-API server. When such caps are retrieved
489 from a mutable directory in a "ro_uri" field, they will be prefixed with
490 the string "ro.", indicating that they must not be decoded without
491 checking that they are read-only. The "ro." prefix must not be stripped
492 off without performing this check. (Future versions of the web-API server
493 will perform it where necessary.)
495 If both the "rw_uri" and "ro_uri" fields are present in a given PROPDICT,
496 and the web-API server recognizes the rw_uri as a write cap, then it will
497 reset the ro_uri to the corresponding read cap and discard the original
498 contents of ro_uri (in order to ensure that the two caps correspond to the
499 same object and that the ro_uri is in fact read-only). However this may not
500 happen for caps in a format unknown to the web-API server. Therefore, when
501 writing a directory the web-API client should ensure that the contents
502 of "rw_uri" and "ro_uri" for a given PROPDICT are a consistent
503 (write cap, read cap) pair if possible. If the web-API client only has
504 one cap and does not know whether it is a write cap or read cap, then
505 it is acceptable to set "rw_uri" to that cap and omit "ro_uri". The
506 client must not put a write cap into a "ro_uri" field.
508 The metadata may have a "no-write" field. If this is set to true in the
509 metadata of a link, it will not be possible to open that link for writing
510 via the SFTP frontend; see FTP-and-SFTP.rst_ for details. Also, if the
511 "no-write" field is set to true in the metadata of a link to a mutable
512 child, it will cause the link to be diminished to read-only.
514 Note that the web-API-using client application must not provide the
515 "Content-Type: multipart/form-data" header that usually accompanies HTML
516 form submissions, since the body is not formatted this way. Doing so will
517 cause a server error as the lower-level code misparses the request body.
519 Child file names should each be expressed as a Unicode string, then used as
520 keys of the dictionary. The dictionary should then be converted into JSON,
521 and the resulting string encoded into UTF-8. This UTF-8 bytestring should
522 then be used as the POST body.
524 ``POST /uri?t=mkdir-immutable``
526 Like t=mkdir-with-children above, but the new directory will be
527 deep-immutable. This means that the directory itself is immutable, and that
528 it can only contain objects that are treated as being deep-immutable, like
529 immutable files, literal files, and deep-immutable directories.
531 For forward-compatibility, a deep-immutable directory can also contain caps
532 in a format that is unknown to the web-API server. When such caps are retrieved
533 from a deep-immutable directory in a "ro_uri" field, they will be prefixed
534 with the string "imm.", indicating that they must not be decoded without
535 checking that they are immutable. The "imm." prefix must not be stripped
536 off without performing this check. (Future versions of the web-API server
537 will perform it where necessary.)
539 The cap for each child may be given either in the "rw_uri" or "ro_uri"
540 field of the PROPDICT (not both). If a cap is given in the "rw_uri" field,
541 then the web-API server will check that it is an immutable read-cap of a
542 *known* format, and give an error if it is not. If a cap is given in the
543 "ro_uri" field, then the web-API server will still check whether known
544 caps are immutable, but for unknown caps it will simply assume that the
545 cap can be stored, as described above. Note that an attacker would be
546 able to store any cap in an immutable directory, so this check when
547 creating the directory is only to help non-malicious clients to avoid
548 accidentally giving away more authority than intended.
550 A non-empty request body is mandatory, since after the directory is created,
551 it will not be possible to add more children to it.
553 ``POST /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=mkdir``
555 ``PUT /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=mkdir``
557 Create new directories as necessary to make sure that the named target
558 ($DIRCAP/SUBDIRS../SUBDIR) is a directory. This will create additional
559 intermediate mutable directories as necessary. If the named target directory
560 already exists, this will make no changes to it.
562 If the final directory is created, it will be empty.
564 This accepts a format= argument in the query string, which controls the
565 format of the named target directory, if it does not already exist. format=
566 is interpreted in the same way as in the POST /uri?t=mkdir form. Note that
567 format= only controls the format of the named target directory;
568 intermediate directories, if created, are created based on the default
569 mutable type, as configured on the Tahoe-LAFS server responding to the
572 This operation will return an error if a blocking file is present at any of
573 the parent names, preventing the server from creating the necessary parent
574 directory; or if it would require changing an immutable directory.
576 The write-cap of the new directory will be returned as the HTTP response
579 ``POST /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=mkdir-with-children``
581 Like /uri?t=mkdir-with-children, but the final directory is created as a
582 child of an existing mutable directory. This will create additional
583 intermediate mutable directories as necessary. If the final directory is
584 created, it will be populated with initial children from the POST request
585 body, as described above.
587 This accepts a format= argument in the query string, which controls the
588 format of the target directory, if the target directory is created as part
589 of the operation. format= is interpreted in the same way as in the POST/
590 uri?t=mkdir-with-children operation. Note that format= only controls the
591 format of the named target directory; intermediate directories, if created,
592 are created using the default mutable type setting, as configured on the
593 Tahoe-LAFS server responding to the request.
595 This operation will return an error if a blocking file is present at any of
596 the parent names, preventing the server from creating the necessary parent
597 directory; or if it would require changing an immutable directory; or if
598 the immediate parent directory already has a a child named SUBDIR.
600 ``POST /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=mkdir-immutable``
602 Like /uri?t=mkdir-immutable, but the final directory is created as a child
603 of an existing mutable directory. The final directory will be deep-immutable,
604 and will be populated with the children specified as a JSON dictionary in
605 the POST request body.
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 SUBDIR.
613 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=mkdir&name=NAME``
615 Create a new empty mutable directory and attach it to the given existing
616 directory. This will create additional intermediate directories as necessary.
618 This accepts a format= argument in the query string, which controls the
619 format of the named target directory, if it does not already exist. format=
620 is interpreted in the same way as in the POST /uri?t=mkdir form. Note that
621 format= only controls the format of the named target directory;
622 intermediate directories, if created, are created based on the default
623 mutable type, as configured on the Tahoe-LAFS server responding to the
626 This operation will return an error if a blocking file is present at any of
627 the parent names, preventing the server from creating the necessary parent
628 directory, or if it would require changing any immutable directory.
630 The URL of this operation points to the parent of the bottommost new directory,
631 whereas the /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=mkdir operation above has a URL
632 that points directly to the bottommost new directory.
634 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=mkdir-with-children&name=NAME``
636 Like /uri/$DIRCAP/[SUBDIRS../]?t=mkdir&name=NAME, but the new directory will
637 be populated with initial children via the POST request body. This command
638 will create additional intermediate mutable directories as necessary.
640 This accepts a format= argument in the query string, which controls the
641 format of the target directory, if the target directory is created as part
642 of the operation. format= is interpreted in the same way as in the POST/
643 uri?t=mkdir-with-children operation. Note that format= only controls the
644 format of the named target directory; intermediate directories, if created,
645 are created using the default mutable type setting, as configured on the
646 Tahoe-LAFS server responding to the request.
648 This operation will return an error if a blocking file is present at any of
649 the parent names, preventing the server from creating the necessary parent
650 directory; or if it would require changing an immutable directory; or if
651 the immediate parent directory already has a a child named NAME.
653 Note that the name= argument must be passed as a queryarg, because the POST
654 request body is used for the initial children JSON.
656 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=mkdir-immutable&name=NAME``
658 Like /uri/$DIRCAP/[SUBDIRS../]?t=mkdir-with-children&name=NAME, but the
659 final directory will be deep-immutable. The children are specified as a
660 JSON dictionary in the POST request body. Again, the name= argument must be
661 passed as a queryarg.
663 In Tahoe 1.6 this operation creates intermediate mutable directories if
664 necessary, but that behaviour should not be relied on; see ticket #920.
666 This operation will return an error if the parent directory is immutable,
667 or already has a child named NAME.
669 .. _FTP-and-SFTP.rst: FTP-and-SFTP.rst
672 Getting Information About a File Or Directory (as JSON)
673 -------------------------------------------------------
675 ``GET /uri/$FILECAP?t=json``
677 ``GET /uri/$DIRCAP?t=json``
679 ``GET /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=json``
681 ``GET /uri/$DIRCAP/[SUBDIRS../]FILENAME?t=json``
683 This returns a machine-parseable JSON-encoded description of the given
684 object. The JSON always contains a list, and the first element of the list is
685 always a flag that indicates whether the referenced object is a file or a
686 directory. If it is a capability to a file, then the information includes
687 file size and URI, like this::
689 GET /uri/$FILECAP?t=json :
693 "verify_uri": verify_uri,
699 If it is a capability to a directory followed by a path from that directory
700 to a file, then the information also includes metadata from the link to the
701 file in the parent directory, like this::
703 GET /uri/$DIRCAP/[SUBDIRS../]FILENAME?t=json
707 "verify_uri": verify_uri,
712 "ctime": 1202777696.7564139,
713 "mtime": 1202777696.7564139,
715 "linkcrtime": 1202777696.7564139,
716 "linkmotime": 1202777696.7564139
719 If it is a directory, then it includes information about the children of
720 this directory, as a mapping from child name to a set of data about the
721 child (the same data that would appear in a corresponding GET?t=json of the
722 child itself). The child entries also include metadata about each child,
723 including link-creation- and link-change- timestamps. The output looks like
726 GET /uri/$DIRCAP?t=json :
727 GET /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=json :
730 "rw_uri": read_write_uri,
731 "ro_uri": read_only_uri,
732 "verify_uri": verify_uri,
736 "foo.txt": [ "filenode",
741 "ctime": 1202777696.7564139,
742 "mtime": 1202777696.7564139,
744 "linkcrtime": 1202777696.7564139,
745 "linkmotime": 1202777696.7564139
747 "subdir": [ "dirnode",
752 "ctime": 1202778102.7589991,
753 "mtime": 1202778111.2160511,
755 "linkcrtime": 1202777696.7564139,
756 "linkmotime": 1202777696.7564139
760 In the above example, note how 'children' is a dictionary in which the keys
761 are child names and the values depend upon whether the child is a file or a
762 directory. The value is mostly the same as the JSON representation of the
763 child object (except that directories do not recurse -- the "children"
764 entry of the child is omitted, and the directory view includes the metadata
765 that is stored on the directory edge).
767 The rw_uri field will be present in the information about a directory
768 if and only if you have read-write access to that directory. The verify_uri
769 field will be present if and only if the object has a verify-cap
770 (non-distributed LIT files do not have verify-caps).
772 If the cap is of an unknown format, then the file size and verify_uri will
775 GET /uri/$UNKNOWNCAP?t=json :
778 "ro_uri": unknown_read_uri
781 GET /uri/$DIRCAP/[SUBDIRS../]UNKNOWNCHILDNAME?t=json :
784 "rw_uri": unknown_write_uri,
785 "ro_uri": unknown_read_uri,
788 "ctime": 1202777696.7564139,
789 "mtime": 1202777696.7564139,
791 "linkcrtime": 1202777696.7564139,
792 "linkmotime": 1202777696.7564139
795 As in the case of file nodes, the metadata will only be present when the
796 capability is to a directory followed by a path. The "mutable" field is also
797 not always present; when it is absent, the mutability of the object is not
803 The value of the 'tahoe':'linkmotime' key is updated whenever a link to a
804 child is set. The value of the 'tahoe':'linkcrtime' key is updated whenever
805 a link to a child is created -- i.e. when there was not previously a link
808 Note however, that if the edge in the Tahoe filesystem points to a mutable
809 file and the contents of that mutable file is changed, then the
810 'tahoe':'linkmotime' value on that edge will *not* be updated, since the
811 edge itself wasn't updated -- only the mutable file was.
813 The timestamps are represented as a number of seconds since the UNIX epoch
814 (1970-01-01 00:00:00 UTC), with leap seconds not being counted in the long
817 In Tahoe earlier than v1.4.0, 'mtime' and 'ctime' keys were populated
818 instead of the 'tahoe':'linkmotime' and 'tahoe':'linkcrtime' keys. Starting
819 in Tahoe v1.4.0, the 'linkmotime'/'linkcrtime' keys in the 'tahoe' sub-dict
820 are populated. However, prior to Tahoe v1.7beta, a bug caused the 'tahoe'
821 sub-dict to be deleted by web-API requests in which new metadata is
822 specified, and not to be added to existing child links that lack it.
824 From Tahoe v1.7.0 onward, the 'mtime' and 'ctime' fields are no longer
825 populated or updated (see ticket #924), except by "tahoe backup" as
826 explained below. For backward compatibility, when an existing link is
827 updated and 'tahoe':'linkcrtime' is not present in the previous metadata
828 but 'ctime' is, the old value of 'ctime' is used as the new value of
829 'tahoe':'linkcrtime'.
831 The reason we added the new fields in Tahoe v1.4.0 is that there is a
832 "set_children" API (described below) which you can use to overwrite the
833 values of the 'mtime'/'ctime' pair, and this API is used by the
834 "tahoe backup" command (in Tahoe v1.3.0 and later) to set the 'mtime' and
835 'ctime' values when backing up files from a local filesystem into the
836 Tahoe filesystem. As of Tahoe v1.4.0, the set_children API cannot be used
837 to set anything under the 'tahoe' key of the metadata dict -- if you
838 include 'tahoe' keys in your 'metadata' arguments then it will silently
841 Therefore, if the 'tahoe' sub-dict is present, you can rely on the
842 'linkcrtime' and 'linkmotime' values therein to have the semantics described
843 above. (This is assuming that only official Tahoe clients have been used to
844 write those links, and that their system clocks were set to what you expected
845 -- there is nothing preventing someone from editing their Tahoe client or
846 writing their own Tahoe client which would overwrite those values however
847 they like, and there is nothing to constrain their system clock from taking
850 When an edge is created or updated by "tahoe backup", the 'mtime' and
851 'ctime' keys on that edge are set as follows:
853 * 'mtime' is set to the timestamp read from the local filesystem for the
854 "mtime" of the local file in question, which means the last time the
855 contents of that file were changed.
857 * On Windows, 'ctime' is set to the creation timestamp for the file
858 read from the local filesystem. On other platforms, 'ctime' is set to
859 the UNIX "ctime" of the local file, which means the last time that
860 either the contents or the metadata of the local file was changed.
862 There are several ways that the 'ctime' field could be confusing:
864 1. You might be confused about whether it reflects the time of the creation
865 of a link in the Tahoe filesystem (by a version of Tahoe < v1.7.0) or a
866 timestamp copied in by "tahoe backup" from a local filesystem.
868 2. You might be confused about whether it is a copy of the file creation
869 time (if "tahoe backup" was run on a Windows system) or of the last
870 contents-or-metadata change (if "tahoe backup" was run on a different
873 3. You might be confused by the fact that changing the contents of a
874 mutable file in Tahoe doesn't have any effect on any links pointing at
875 that file in any directories, although "tahoe backup" sets the link
876 'ctime'/'mtime' to reflect timestamps about the local file corresponding
877 to the Tahoe file to which the link points.
879 4. Also, quite apart from Tahoe, you might be confused about the meaning
880 of the "ctime" in UNIX local filesystems, which people sometimes think
881 means file creation time, but which actually means, in UNIX local
882 filesystems, the most recent time that the file contents or the file
883 metadata (such as owner, permission bits, extended attributes, etc.)
884 has changed. Note that although "ctime" does not mean file creation time
885 in UNIX, links created by a version of Tahoe prior to v1.7.0, and never
886 written by "tahoe backup", will have 'ctime' set to the link creation
890 Attaching an Existing File or Directory by its read- or write-cap
891 -----------------------------------------------------------------
893 ``PUT /uri/$DIRCAP/[SUBDIRS../]CHILDNAME?t=uri``
895 This attaches a child object (either a file or directory) to a specified
896 location in the virtual filesystem. The child object is referenced by its
897 read- or write- cap, as provided in the HTTP request body. This will create
898 intermediate directories as necessary.
900 This is similar to a UNIX hardlink: by referencing a previously-uploaded file
901 (or previously-created directory) instead of uploading/creating a new one,
902 you can create two references to the same object.
904 The read- or write- cap of the child is provided in the body of the HTTP
905 request, and this same cap is returned in the response body.
907 The default behavior is to overwrite any existing object at the same
908 location. To prevent this (and make the operation return an error instead
909 of overwriting), add a "replace=false" argument, as "?t=uri&replace=false".
910 With replace=false, this operation will return an HTTP 409 "Conflict" error
911 if there is already an object at the given location, rather than
912 overwriting the existing object. To allow the operation to overwrite a
913 file, but return an error when trying to overwrite a directory, use
914 "replace=only-files" (this behavior is closer to the traditional UNIX "mv"
915 command). Note that "true", "t", and "1" are all synonyms for "True", and
916 "false", "f", and "0" are synonyms for "False", and the parameter is
919 Note that this operation does not take its child cap in the form of
920 separate "rw_uri" and "ro_uri" fields. Therefore, it cannot accept a
921 child cap in a format unknown to the web-API server, unless its URI
922 starts with "ro." or "imm.". This restriction is necessary because the
923 server is not able to attenuate an unknown write cap to a read cap.
924 Unknown URIs starting with "ro." or "imm.", on the other hand, are
925 assumed to represent read caps. The client should not prefix a write
926 cap with "ro." or "imm." and pass it to this operation, since that
927 would result in granting the cap's write authority to holders of the
931 Adding Multiple Files or Directories to a Parent Directory at Once
932 ------------------------------------------------------------------
934 ``POST /uri/$DIRCAP/[SUBDIRS..]?t=set_children``
936 ``POST /uri/$DIRCAP/[SUBDIRS..]?t=set-children`` (Tahoe >= v1.6)
938 This command adds multiple children to a directory in a single operation.
939 It reads the request body and interprets it as a JSON-encoded description
940 of the child names and read/write-caps that should be added.
942 The body should be a JSON-encoded dictionary, in the same format as the
943 "children" value returned by the "GET /uri/$DIRCAP?t=json" operation
944 described above. In this format, each key is a child names, and the
945 corresponding value is a tuple of (type, childinfo). "type" is ignored, and
946 "childinfo" is a dictionary that contains "rw_uri", "ro_uri", and
947 "metadata" keys. You can take the output of "GET /uri/$DIRCAP1?t=json" and
948 use it as the input to "POST /uri/$DIRCAP2?t=set_children" to make DIR2
949 look very much like DIR1 (except for any existing children of DIR2 that
950 were not overwritten, and any existing "tahoe" metadata keys as described
953 When the set_children request contains a child name that already exists in
954 the target directory, this command defaults to overwriting that child with
955 the new value (both child cap and metadata, but if the JSON data does not
956 contain a "metadata" key, the old child's metadata is preserved). The
957 command takes a boolean "overwrite=" query argument to control this
958 behavior. If you use "?t=set_children&overwrite=false", then an attempt to
959 replace an existing child will instead cause an error.
961 Any "tahoe" key in the new child's "metadata" value is ignored. Any
962 existing "tahoe" metadata is preserved. The metadata["tahoe"] value is
963 reserved for metadata generated by the tahoe node itself. The only two keys
964 currently placed here are "linkcrtime" and "linkmotime". For details, see
965 the section above entitled "Getting Information About a File Or Directory (as
966 JSON)", in the "About the metadata" subsection.
968 Note that this command was introduced with the name "set_children", which
969 uses an underscore rather than a hyphen as other multi-word command names
970 do. The variant with a hyphen is now accepted, but clients that desire
971 backward compatibility should continue to use "set_children".
974 Unlinking a File or Directory
975 -----------------------------
977 ``DELETE /uri/$DIRCAP/[SUBDIRS../]CHILDNAME``
979 This removes the given name from its parent directory. CHILDNAME is the
980 name to be removed, and $DIRCAP/SUBDIRS.. indicates the directory that will
983 Note that this does not actually delete the file or directory that the name
984 points to from the tahoe grid -- it only unlinks the named reference from
985 this directory. If there are other names in this directory or in other
986 directories that point to the resource, then it will remain accessible
987 through those paths. Even if all names pointing to this object are removed
988 from their parent directories, then someone with possession of its read-cap
989 can continue to access the object through that cap.
991 The object will only become completely unreachable once 1: there are no
992 reachable directories that reference it, and 2: nobody is holding a read-
993 or write- cap to the object. (This behavior is very similar to the way
994 hardlinks and anonymous files work in traditional UNIX filesystems).
996 This operation will not modify more than a single directory. Intermediate
997 directories which were implicitly created by PUT or POST methods will *not*
998 be automatically removed by DELETE.
1000 This method returns the file- or directory- cap of the object that was just
1004 Browser Operations: Human-oriented interfaces
1005 =============================================
1007 This section describes the HTTP operations that provide support for humans
1008 running a web browser. Most of these operations use HTML forms that use POST
1009 to drive the Tahoe node. This section is intended for HTML authors who want
1010 to write web pages that contain forms and buttons which manipulate the Tahoe
1013 Note that for all POST operations, the arguments listed can be provided
1014 either as URL query arguments or as form body fields. URL query arguments are
1015 separated from the main URL by "?", and from each other by "&". For example,
1016 "POST /uri/$DIRCAP?t=upload&mutable=true". Form body fields are usually
1017 specified by using <input type="hidden"> elements. For clarity, the
1018 descriptions below display the most significant arguments as URL query args.
1021 Viewing a Directory (as HTML)
1022 -----------------------------
1024 ``GET /uri/$DIRCAP/[SUBDIRS../]``
1026 This returns an HTML page, intended to be displayed to a human by a web
1027 browser, which contains HREF links to all files and directories reachable
1028 from this directory. These HREF links do not have a t= argument, meaning
1029 that a human who follows them will get pages also meant for a human. It also
1030 contains forms to upload new files, and to unlink files and directories
1031 from their parent directory. Those forms use POST methods to do their job.
1034 Viewing/Downloading a File
1035 --------------------------
1037 ``GET /uri/$FILECAP``
1039 ``GET /uri/$DIRCAP/[SUBDIRS../]FILENAME``
1041 This will retrieve the contents of the given file. The HTTP response body
1042 will contain the sequence of bytes that make up the file.
1044 If you want the HTTP response to include a useful Content-Type header,
1045 either use the second form (which starts with a $DIRCAP), or add a
1046 "filename=foo" query argument, like "GET /uri/$FILECAP?filename=foo.jpg".
1047 The bare "GET /uri/$FILECAP" does not give the Tahoe node enough information
1048 to determine a Content-Type (since Tahoe immutable files are merely
1049 sequences of bytes, not typed+named file objects).
1051 If the URL has both filename= and "save=true" in the query arguments, then
1052 the server to add a "Content-Disposition: attachment" header, along with a
1053 filename= parameter. When a user clicks on such a link, most browsers will
1054 offer to let the user save the file instead of displaying it inline (indeed,
1055 most browsers will refuse to display it inline). "true", "t", "1", and other
1056 case-insensitive equivalents are all treated the same.
1058 Character-set handling in URLs and HTTP headers is a dubious art [1]_. For
1059 maximum compatibility, Tahoe simply copies the bytes from the filename=
1060 argument into the Content-Disposition header's filename= parameter, without
1061 trying to interpret them in any particular way.
1064 ``GET /named/$FILECAP/FILENAME``
1066 This is an alternate download form which makes it easier to get the correct
1067 filename. The Tahoe server will provide the contents of the given file, with
1068 a Content-Type header derived from the given filename. This form is used to
1069 get browsers to use the "Save Link As" feature correctly, and also helps
1070 command-line tools like "wget" and "curl" use the right filename. Note that
1071 this form can *only* be used with file caps; it is an error to use a
1072 directory cap after the /named/ prefix.
1074 URLs may also use /file/$FILECAP/FILENAME as a synonym for
1075 /named/$FILECAP/FILENAME.
1077 Getting Information About a File Or Directory (as HTML)
1078 -------------------------------------------------------
1080 ``GET /uri/$FILECAP?t=info``
1082 ``GET /uri/$DIRCAP/?t=info``
1084 ``GET /uri/$DIRCAP/[SUBDIRS../]SUBDIR/?t=info``
1086 ``GET /uri/$DIRCAP/[SUBDIRS../]FILENAME?t=info``
1088 This returns a human-oriented HTML page with more detail about the selected
1089 file or directory object. This page contains the following items:
1093 * JSON representation
1094 * raw contents (text/plain)
1095 * access caps (URIs): verify-cap, read-cap, write-cap (for mutable objects)
1096 * check/verify/repair form
1097 * deep-check/deep-size/deep-stats/manifest (for directories)
1098 * replace-contents form (for mutable files)
1101 Creating a Directory
1102 --------------------
1104 ``POST /uri?t=mkdir``
1106 This creates a new empty directory, but does not attach it to the virtual
1109 If a "redirect_to_result=true" argument is provided, then the HTTP response
1110 will cause the web browser to be redirected to a /uri/$DIRCAP page that
1111 gives access to the newly-created directory. If you bookmark this page,
1112 you'll be able to get back to the directory again in the future. This is the
1113 recommended way to start working with a Tahoe server: create a new unlinked
1114 directory (using redirect_to_result=true), then bookmark the resulting
1115 /uri/$DIRCAP page. There is a "create directory" button on the Welcome page
1116 to invoke this action.
1118 This accepts a format= argument in the query string. Refer to the
1119 documentation of the PUT /uri?t=mkdir operation in `Creating A
1120 New Directory`_ for information on the behavior of the format= argument.
1122 If "redirect_to_result=true" is not provided (or is given a value of
1123 "false"), then the HTTP response body will simply be the write-cap of the
1126 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=mkdir&name=CHILDNAME``
1128 This creates a new empty directory as a child of the designated SUBDIR. This
1129 will create additional intermediate directories as necessary.
1131 This accepts a format= argument in the query string. Refer to the
1132 documentation of POST /uri/$DIRCAP/[SUBDIRS../]?t=mkdir&name=CHILDNAME in
1133 `Creating a New Directory`_ for information on the behavior of the format=
1136 If a "when_done=URL" argument is provided, the HTTP response will cause the
1137 web browser to redirect to the given URL. This provides a convenient way to
1138 return the browser to the directory that was just modified. Without a
1139 when_done= argument, the HTTP response will simply contain the write-cap of
1140 the directory that was just created.
1146 ``POST /uri?t=upload``
1148 This uploads a file, and produces a file-cap for the contents, but does not
1149 attach the file into the filesystem. No directories will be modified by
1152 The file must be provided as the "file" field of an HTML encoded form body,
1153 produced in response to an HTML form like this::
1155 <form action="/uri" method="POST" enctype="multipart/form-data">
1156 <input type="hidden" name="t" value="upload" />
1157 <input type="file" name="file" />
1158 <input type="submit" value="Upload Unlinked" />
1161 If a "when_done=URL" argument is provided, the response body will cause the
1162 browser to redirect to the given URL. If the when_done= URL has the string
1163 "%(uri)s" in it, that string will be replaced by a URL-escaped form of the
1164 newly created file-cap. (Note that without this substitution, there is no
1165 way to access the file that was just uploaded).
1167 The default (in the absence of when_done=) is to return an HTML page that
1168 describes the results of the upload. This page will contain information
1169 about which storage servers were used for the upload, how long each
1170 operation took, etc.
1172 This accepts format= and mutable=true query string arguments. Refer to
1173 `Writing/Uploading a File`_ for information on the behavior of format= and
1176 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=upload``
1178 This uploads a file, and attaches it as a new child of the given directory,
1179 which must be mutable. The file must be provided as the "file" field of an
1180 HTML-encoded form body, produced in response to an HTML form like this::
1182 <form action="." method="POST" enctype="multipart/form-data">
1183 <input type="hidden" name="t" value="upload" />
1184 <input type="file" name="file" />
1185 <input type="submit" value="Upload" />
1188 A "name=" argument can be provided to specify the new child's name,
1189 otherwise it will be taken from the "filename" field of the upload form
1190 (most web browsers will copy the last component of the original file's
1191 pathname into this field). To avoid confusion, name= is not allowed to
1194 If there is already a child with that name, and it is a mutable file, then
1195 its contents are replaced with the data being uploaded. If it is not a
1196 mutable file, the default behavior is to remove the existing child before
1197 creating a new one. To prevent this (and make the operation return an error
1198 instead of overwriting the old child), add a "replace=false" argument, as
1199 "?t=upload&replace=false". With replace=false, this operation will return an
1200 HTTP 409 "Conflict" error if there is already an object at the given
1201 location, rather than overwriting the existing object. Note that "true",
1202 "t", and "1" are all synonyms for "True", and "false", "f", and "0" are
1203 synonyms for "False". the parameter is case-insensitive.
1205 This will create additional intermediate directories as necessary, although
1206 since it is expected to be triggered by a form that was retrieved by "GET
1207 /uri/$DIRCAP/[SUBDIRS../]", it is likely that the parent directory will
1210 This accepts format= and mutable=true query string arguments. Refer to
1211 `Writing/Uploading a File`_ for information on the behavior of format= and
1214 If a "when_done=URL" argument is provided, the HTTP response will cause the
1215 web browser to redirect to the given URL. This provides a convenient way to
1216 return the browser to the directory that was just modified. Without a
1217 when_done= argument, the HTTP response will simply contain the file-cap of
1218 the file that was just uploaded (a write-cap for mutable files, or a
1219 read-cap for immutable files).
1221 ``POST /uri/$DIRCAP/[SUBDIRS../]FILENAME?t=upload``
1223 This also uploads a file and attaches it as a new child of the given
1224 directory, which must be mutable. It is a slight variant of the previous
1225 operation, as the URL refers to the target file rather than the parent
1226 directory. It is otherwise identical: this accepts mutable= and when_done=
1229 ``POST /uri/$FILECAP?t=upload``
1231 This modifies the contents of an existing mutable file in-place. An error is
1232 signalled if $FILECAP does not refer to a mutable file. It behaves just like
1233 the "PUT /uri/$FILECAP" form, but uses a POST for the benefit of HTML forms
1237 Attaching An Existing File Or Directory (by URI)
1238 ------------------------------------------------
1240 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=uri&name=CHILDNAME&uri=CHILDCAP``
1242 This attaches a given read- or write- cap "CHILDCAP" to the designated
1243 directory, with a specified child name. This behaves much like the PUT t=uri
1244 operation, and is a lot like a UNIX hardlink. It is subject to the same
1245 restrictions as that operation on the use of cap formats unknown to the
1248 This will create additional intermediate directories as necessary, although
1249 since it is expected to be triggered by a form that was retrieved by "GET
1250 /uri/$DIRCAP/[SUBDIRS../]", it is likely that the parent directory will
1253 This accepts the same replace= argument as POST t=upload.
1259 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=delete&name=CHILDNAME``
1261 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=unlink&name=CHILDNAME`` (Tahoe >= v1.9)
1263 This instructs the node to remove a child object (file or subdirectory) from
1264 the given directory, which must be mutable. Note that the entire subtree is
1265 unlinked from the parent. Unlike deleting a subdirectory in a UNIX local
1266 filesystem, the subtree need not be empty; if it isn't, then other references
1267 into the subtree will see that the child subdirectories are not modified by
1268 this operation. Only the link from the given directory to its child is severed.
1270 In Tahoe-LAFS v1.9.0 and later, t=unlink can be used as a synonym for t=delete.
1271 If interoperability with older web-API servers is required, t=delete should
1278 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=rename&from_name=OLD&to_name=NEW``
1280 This instructs the node to rename a child of the given directory, which must
1281 be mutable. This has a similar effect to removing the child, then adding the
1282 same child-cap under the new name, except that it preserves metadata. This
1283 operation cannot move the child to a different directory.
1285 The default behavior is to overwrite any existing link at the destination
1286 (replace=true). To prevent this (and make the operation return an error
1287 instead of overwriting), add a "replace=false" argument. With replace=false,
1288 this operation will return an HTTP 409 "Conflict" error if the destination
1289 is not the same link as the source and there is already a link at the
1290 destination, rather than overwriting the existing link. To allow the
1291 operation to overwrite a link to a file, but return an HTTP 409 error when
1292 trying to overwrite a link to a directory, use "replace=only-files" (this
1293 behavior is closer to the traditional UNIX "mv" command). Note that "true",
1294 "t", and "1" are all synonyms for "True"; "false", "f", and "0" are synonyms
1295 for "False"; and the parameter is case-insensitive.
1298 Relinking ("Moving") a Child
1299 ----------------------------
1301 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=relink&from_name=OLD&to_dir=$NEWDIRCAP/[NEWSUBDIRS../]&to_name=NEW``
1302 ``[&replace=true|false|only-files]`` (Tahoe >= v1.10)
1304 This instructs the node to move a child of the given source directory, into
1305 a different directory and/or to a different name. The command is named
1306 ``relink`` because what it does is add a new link to the child from the new
1307 location, then remove the old link. Nothing is actually "moved": the child
1308 is still reachable through any path from which it was formerly reachable,
1309 and the storage space occupied by its ciphertext is not affected.
1311 The source and destination directories must be writeable. If {{{to_dir}}} is
1312 not present, the child link is renamed within the same directory. If
1313 {{{to_name}}} is not present then it defaults to {{{from_name}}}. If the
1314 destination link (directory and name) is the same as the source link, the
1315 operation has no effect.
1317 Metadata from the source directory entry is preserved. Multiple levels of
1318 descent in the source and destination paths are supported.
1320 This operation will return an HTTP 404 "Not Found" error if
1321 ``$DIRCAP/[SUBDIRS../]``, the child being moved, or the destination
1322 directory does not exist. It will return an HTTP 400 "Bad Request" error
1323 if any entry in the source or destination paths is not a directory.
1325 The default behavior is to overwrite any existing link at the destination
1326 (replace=true). To prevent this (and make the operation return an error
1327 instead of overwriting), add a "replace=false" argument. With replace=false,
1328 this operation will return an HTTP 409 "Conflict" error if the destination
1329 is not the same link as the source and there is already a link at the
1330 destination, rather than overwriting the existing link. To allow the
1331 operation to overwrite a link to a file, but return an HTTP 409 error when
1332 trying to overwrite a link to a directory, use "replace=only-files" (this
1333 behavior is closer to the traditional UNIX "mv" command). Note that "true",
1334 "t", and "1" are all synonyms for "True"; "false", "f", and "0" are synonyms
1335 for "False"; and the parameter is case-insensitive.
1337 When relinking into a different directory, for safety, the child link is
1338 not removed from the old directory until it has been successfully added to
1339 the new directory. This implies that in case of a crash or failure, the
1340 link to the child will not be lost, but it could be linked at both the old
1343 The source link should not be the same as any link (directory and child name)
1344 in the ``to_dir`` path. This restriction is not enforced, but it may be
1345 enforced in a future version. If it were violated then the result would be
1346 to create a cycle in the directory structure that is not necessarily reachable
1347 from the root of the destination path (``$NEWDIRCAP``), which could result in
1348 data loss, as described in ticket `#943`_.
1350 .. _`#943`: https://tahoe-lafs.org/trac/tahoe-lafs/ticket/943
1356 ``GET /uri?uri=$CAP``
1358 This causes a redirect to /uri/$CAP, and retains any additional query
1359 arguments (like filename= or save=). This is for the convenience of web
1360 forms which allow the user to paste in a read- or write- cap (obtained
1361 through some out-of-band channel, like IM or email).
1363 Note that this form merely redirects to the specific file or directory
1364 indicated by the $CAP: unlike the GET /uri/$DIRCAP form, you cannot
1365 traverse to children by appending additional path segments to the URL.
1367 ``GET /uri/$DIRCAP/[SUBDIRS../]?t=rename-form&name=$CHILDNAME``
1369 This provides a useful facility to browser-based user interfaces. It
1370 returns a page containing a form targetting the "POST $DIRCAP t=rename"
1371 functionality described above, with the provided $CHILDNAME present in the
1372 'from_name' field of that form. I.e. this presents a form offering to
1373 rename $CHILDNAME, requesting the new name, and submitting POST rename.
1374 This same URL format can also be used with "move-form" with the expected
1377 ``GET /uri/$DIRCAP/[SUBDIRS../]CHILDNAME?t=uri``
1379 This returns the file- or directory- cap for the specified object.
1381 ``GET /uri/$DIRCAP/[SUBDIRS../]CHILDNAME?t=readonly-uri``
1383 This returns a read-only file- or directory- cap for the specified object.
1384 If the object is an immutable file, this will return the same value as
1388 Debugging and Testing Features
1389 ------------------------------
1391 These URLs are less-likely to be helpful to the casual Tahoe user, and are
1392 mainly intended for developers.
1394 ``POST $URL?t=check``
1396 This triggers the FileChecker to determine the current "health" of the
1397 given file or directory, by counting how many shares are available. The
1398 page that is returned will display the results. This can be used as a "show
1399 me detailed information about this file" page.
1401 If a verify=true argument is provided, the node will perform a more
1402 intensive check, downloading and verifying every single bit of every share.
1404 If an add-lease=true argument is provided, the node will also add (or
1405 renew) a lease to every share it encounters. Each lease will keep the share
1406 alive for a certain period of time (one month by default). Once the last
1407 lease expires or is explicitly cancelled, the storage server is allowed to
1410 If an output=JSON argument is provided, the response will be
1411 machine-readable JSON instead of human-oriented HTML. The data is a
1412 dictionary with the following keys::
1414 storage-index: a base32-encoded string with the objects's storage index,
1415 or an empty string for LIT files
1416 summary: a string, with a one-line summary of the stats of the file
1417 results: a dictionary that describes the state of the file. For LIT files,
1418 this dictionary has only the 'healthy' key, which will always be
1419 True. For distributed files, this dictionary has the following
1421 count-happiness: the servers-of-happiness level of the file, as
1422 defined in `docs/specifications/servers-of-happiness.rst`_.
1423 count-shares-good: the number of good shares that were found
1424 count-shares-needed: 'k', the number of shares required for recovery
1425 count-shares-expected: 'N', the number of total shares generated
1426 count-good-share-hosts: the number of distinct storage servers with
1427 good shares. Note that a high value does not
1428 necessarily imply good share distribution,
1429 because some of these servers may only hold
1431 count-wrong-shares: for mutable files, the number of shares for
1432 versions other than the 'best' one (highest
1433 sequence number, highest roothash). These are
1434 either old, or created by an uncoordinated or
1435 not fully successful write.
1436 count-recoverable-versions: for mutable files, the number of
1437 recoverable versions of the file. For
1438 a healthy file, this will equal 1.
1439 count-unrecoverable-versions: for mutable files, the number of
1440 unrecoverable versions of the file.
1441 For a healthy file, this will be 0.
1442 count-corrupt-shares: the number of shares with integrity failures
1443 list-corrupt-shares: a list of "share locators", one for each share
1444 that was found to be corrupt. Each share locator
1445 is a list of (serverid, storage_index, sharenum).
1446 servers-responding: list of base32-encoded storage server identifiers,
1447 one for each server which responded to the share
1449 healthy: (bool) True if the file is completely healthy, False otherwise.
1450 Healthy files have at least N good shares. Overlapping shares
1451 do not currently cause a file to be marked unhealthy. If there
1452 are at least N good shares, then corrupt shares do not cause the
1453 file to be marked unhealthy, although the corrupt shares will be
1454 listed in the results (list-corrupt-shares) and should be manually
1455 removed to wasting time in subsequent downloads (as the
1456 downloader rediscovers the corruption and uses alternate shares).
1457 Future compatibility: the meaning of this field may change to
1458 reflect whether the servers-of-happiness criterion is met
1460 sharemap: dict mapping share identifier to list of serverids
1461 (base32-encoded strings). This indicates which servers are
1462 holding which shares. For immutable files, the shareid is
1463 an integer (the share number, from 0 to N-1). For
1464 immutable files, it is a string of the form
1465 'seq%d-%s-sh%d', containing the sequence number, the
1466 roothash, and the share number.
1468 Before Tahoe-LAFS v1.11, the `results` dictionary also had a `needs-rebalancing`
1469 field, but that has been removed since it was computed incorrectly.
1471 .. _`docs/specifications/servers-of-happiness.rst`: ../specifications/servers-of-happiness.rst
1474 ``POST $URL?t=start-deep-check`` (must add &ophandle=XYZ)
1476 This initiates a recursive walk of all files and directories reachable from
1477 the target, performing a check on each one just like t=check. The result
1478 page will contain a summary of the results, including details on any
1479 file/directory that was not fully healthy.
1481 t=start-deep-check can only be invoked on a directory. An error (400
1482 BAD_REQUEST) will be signalled if it is invoked on a file. The recursive
1483 walker will deal with loops safely.
1485 This accepts the same verify= and add-lease= arguments as t=check.
1487 Since this operation can take a long time (perhaps a second per object),
1488 the ophandle= argument is required (see "Slow Operations, Progress, and
1489 Cancelling" above). The response to this POST will be a redirect to the
1490 corresponding /operations/$HANDLE page (with output=HTML or output=JSON to
1491 match the output= argument given to the POST). The deep-check operation
1492 will continue to run in the background, and the /operations page should be
1493 used to find out when the operation is done.
1495 Detailed check results for non-healthy files and directories will be
1496 available under /operations/$HANDLE/$STORAGEINDEX, and the HTML status will
1497 contain links to these detailed results.
1499 The HTML /operations/$HANDLE page for incomplete operations will contain a
1500 meta-refresh tag, set to 60 seconds, so that a browser which uses
1501 deep-check will automatically poll until the operation has completed.
1503 The JSON page (/options/$HANDLE?output=JSON) will contain a
1504 machine-readable JSON dictionary with the following keys::
1506 finished: a boolean, True if the operation is complete, else False. Some
1507 of the remaining keys may not be present until the operation
1509 root-storage-index: a base32-encoded string with the storage index of the
1510 starting point of the deep-check operation
1511 count-objects-checked: count of how many objects were checked. Note that
1512 non-distributed objects (i.e. small immutable LIT
1513 files) are not checked, since for these objects,
1514 the data is contained entirely in the URI.
1515 count-objects-healthy: how many of those objects were completely healthy
1516 count-objects-unhealthy: how many were damaged in some way
1517 count-corrupt-shares: how many shares were found to have corruption,
1518 summed over all objects examined
1519 list-corrupt-shares: a list of "share identifiers", one for each share
1520 that was found to be corrupt. Each share identifier
1521 is a list of (serverid, storage_index, sharenum).
1522 list-unhealthy-files: a list of (pathname, check-results) tuples, for
1523 each file that was not fully healthy. 'pathname' is
1524 a list of strings (which can be joined by "/"
1525 characters to turn it into a single string),
1526 relative to the directory on which deep-check was
1527 invoked. The 'check-results' field is the same as
1528 that returned by t=check&output=JSON, described
1530 stats: a dictionary with the same keys as the t=start-deep-stats command
1533 ``POST $URL?t=stream-deep-check``
1535 This initiates a recursive walk of all files and directories reachable from
1536 the target, performing a check on each one just like t=check. For each
1537 unique object (duplicates are skipped), a single line of JSON is emitted to
1538 the HTTP response channel (or an error indication, see below). When the walk
1539 is complete, a final line of JSON is emitted which contains the accumulated
1540 file-size/count "deep-stats" data.
1542 This command takes the same arguments as t=start-deep-check.
1544 A CLI tool can split the response stream on newlines into "response units",
1545 and parse each response unit as JSON. Each such parsed unit will be a
1546 dictionary, and will contain at least the "type" key: a string, one of
1547 "file", "directory", or "stats".
1549 For all units that have a type of "file" or "directory", the dictionary will
1550 contain the following keys::
1552 "path": a list of strings, with the path that is traversed to reach the
1554 "cap": a write-cap URI for the file or directory, if available, else a
1556 "verifycap": a verify-cap URI for the file or directory
1557 "repaircap": an URI for the weakest cap that can still be used to repair
1559 "storage-index": a base32 storage index for the object
1560 "check-results": a copy of the dictionary which would be returned by
1561 t=check&output=json, with three top-level keys:
1562 "storage-index", "summary", and "results", and a variety
1563 of counts and sharemaps in the "results" value.
1565 Note that non-distributed files (i.e. LIT files) will have values of None
1566 for verifycap, repaircap, and storage-index, since these files can neither
1567 be verified nor repaired, and are not stored on the storage servers.
1568 Likewise the check-results dictionary will be limited: an empty string for
1569 storage-index, and a results dictionary with only the "healthy" key.
1571 The last unit in the stream will have a type of "stats", and will contain
1572 the keys described in the "start-deep-stats" operation, below.
1574 If any errors occur during the traversal (specifically if a directory is
1575 unrecoverable, such that further traversal is not possible), an error
1576 indication is written to the response body, instead of the usual line of
1577 JSON. This error indication line will begin with the string "ERROR:" (in all
1578 caps), and contain a summary of the error on the rest of the line. The
1579 remaining lines of the response body will be a python exception. The client
1580 application should look for the ERROR: and stop processing JSON as soon as
1581 it is seen. Note that neither a file being unrecoverable nor a directory
1582 merely being unhealthy will cause traversal to stop. The line just before
1583 the ERROR: will describe the directory that was untraversable, since the
1584 unit is emitted to the HTTP response body before the child is traversed.
1587 ``POST $URL?t=check&repair=true``
1589 This performs a health check of the given file or directory, and if the
1590 checker determines that the object is not healthy (some shares are missing
1591 or corrupted), it will perform a "repair". During repair, any missing
1592 shares will be regenerated and uploaded to new servers.
1594 This accepts the same verify=true and add-lease= arguments as t=check. When
1595 an output=JSON argument is provided, the machine-readable JSON response
1596 will contain the following keys::
1598 storage-index: a base32-encoded string with the objects's storage index,
1599 or an empty string for LIT files
1600 repair-attempted: (bool) True if repair was attempted
1601 repair-successful: (bool) True if repair was attempted and the file was
1602 fully healthy afterwards. False if no repair was
1603 attempted, or if a repair attempt failed.
1604 pre-repair-results: a dictionary that describes the state of the file
1605 before any repair was performed. This contains exactly
1606 the same keys as the 'results' value of the t=check
1607 response, described above.
1608 post-repair-results: a dictionary that describes the state of the file
1609 after any repair was performed. If no repair was
1610 performed, post-repair-results and pre-repair-results
1611 will be the same. This contains exactly the same keys
1612 as the 'results' value of the t=check response,
1615 ``POST $URL?t=start-deep-check&repair=true`` (must add &ophandle=XYZ)
1617 This triggers a recursive walk of all files and directories, performing a
1618 t=check&repair=true on each one.
1620 Like t=start-deep-check without the repair= argument, this can only be
1621 invoked on a directory. An error (400 BAD_REQUEST) will be signalled if it
1622 is invoked on a file. The recursive walker will deal with loops safely.
1624 This accepts the same verify= and add-lease= arguments as
1625 t=start-deep-check. It uses the same ophandle= mechanism as
1626 start-deep-check. When an output=JSON argument is provided, the response
1627 will contain the following keys::
1629 finished: (bool) True if the operation has completed, else False
1630 root-storage-index: a base32-encoded string with the storage index of the
1631 starting point of the deep-check operation
1632 count-objects-checked: count of how many objects were checked
1634 count-objects-healthy-pre-repair: how many of those objects were completely
1635 healthy, before any repair
1636 count-objects-unhealthy-pre-repair: how many were damaged in some way
1637 count-objects-healthy-post-repair: how many of those objects were completely
1638 healthy, after any repair
1639 count-objects-unhealthy-post-repair: how many were damaged in some way
1641 count-repairs-attempted: repairs were attempted on this many objects.
1642 count-repairs-successful: how many repairs resulted in healthy objects
1643 count-repairs-unsuccessful: how many repairs resulted did not results in
1644 completely healthy objects
1645 count-corrupt-shares-pre-repair: how many shares were found to have
1646 corruption, summed over all objects
1647 examined, before any repair
1648 count-corrupt-shares-post-repair: how many shares were found to have
1649 corruption, summed over all objects
1650 examined, after any repair
1651 list-corrupt-shares: a list of "share identifiers", one for each share
1652 that was found to be corrupt (before any repair).
1653 Each share identifier is a list of (serverid,
1654 storage_index, sharenum).
1655 list-remaining-corrupt-shares: like list-corrupt-shares, but mutable shares
1656 that were successfully repaired are not
1657 included. These are shares that need
1658 manual processing. Since immutable shares
1659 cannot be modified by clients, all corruption
1660 in immutable shares will be listed here.
1661 list-unhealthy-files: a list of (pathname, check-results) tuples, for
1662 each file that was not fully healthy. 'pathname' is
1663 relative to the directory on which deep-check was
1664 invoked. The 'check-results' field is the same as
1665 that returned by t=check&repair=true&output=JSON,
1667 stats: a dictionary with the same keys as the t=start-deep-stats command
1670 ``POST $URL?t=stream-deep-check&repair=true``
1672 This triggers a recursive walk of all files and directories, performing a
1673 t=check&repair=true on each one. For each unique object (duplicates are
1674 skipped), a single line of JSON is emitted to the HTTP response channel (or
1675 an error indication). When the walk is complete, a final line of JSON is
1676 emitted which contains the accumulated file-size/count "deep-stats" data.
1678 This emits the same data as t=stream-deep-check (without the repair=true),
1679 except that the "check-results" field is replaced with a
1680 "check-and-repair-results" field, which contains the keys returned by
1681 t=check&repair=true&output=json (i.e. repair-attempted, repair-successful,
1682 pre-repair-results, and post-repair-results). The output does not contain
1683 the summary dictionary that is provied by t=start-deep-check&repair=true
1684 (the one with count-objects-checked and list-unhealthy-files), since the
1685 receiving client is expected to calculate those values itself from the
1686 stream of per-object check-and-repair-results.
1688 Note that the "ERROR:" indication will only be emitted if traversal stops,
1689 which will only occur if an unrecoverable directory is encountered. If a
1690 file or directory repair fails, the traversal will continue, and the repair
1691 failure will be indicated in the JSON data (in the "repair-successful" key).
1693 ``POST $DIRURL?t=start-manifest`` (must add &ophandle=XYZ)
1695 This operation generates a "manfest" of the given directory tree, mostly
1696 for debugging. This is a table of (path, filecap/dircap), for every object
1697 reachable from the starting directory. The path will be slash-joined, and
1698 the filecap/dircap will contain a link to the object in question. This page
1699 gives immediate access to every object in the virtual filesystem subtree.
1701 This operation uses the same ophandle= mechanism as deep-check. The
1702 corresponding /operations/$HANDLE page has three different forms. The
1703 default is output=HTML.
1705 If output=text is added to the query args, the results will be a text/plain
1706 list. The first line is special: it is either "finished: yes" or "finished:
1707 no"; if the operation is not finished, you must periodically reload the
1708 page until it completes. The rest of the results are a plaintext list, with
1709 one file/dir per line, slash-separated, with the filecap/dircap separated
1712 If output=JSON is added to the queryargs, then the results will be a
1713 JSON-formatted dictionary with six keys. Note that because large directory
1714 structures can result in very large JSON results, the full results will not
1715 be available until the operation is complete (i.e. until output["finished"]
1718 finished (bool): if False then you must reload the page until True
1719 origin_si (base32 str): the storage index of the starting point
1720 manifest: list of (path, cap) tuples, where path is a list of strings.
1721 verifycaps: list of (printable) verify cap strings
1722 storage-index: list of (base32) storage index strings
1723 stats: a dictionary with the same keys as the t=start-deep-stats command
1726 ``POST $DIRURL?t=start-deep-size`` (must add &ophandle=XYZ)
1728 This operation generates a number (in bytes) containing the sum of the
1729 filesize of all directories and immutable files reachable from the given
1730 directory. This is a rough lower bound of the total space consumed by this
1731 subtree. It does not include space consumed by mutable files, nor does it
1732 take expansion or encoding overhead into account. Later versions of the
1733 code may improve this estimate upwards.
1735 The /operations/$HANDLE status output consists of two lines of text::
1740 ``POST $DIRURL?t=start-deep-stats`` (must add &ophandle=XYZ)
1742 This operation performs a recursive walk of all files and directories
1743 reachable from the given directory, and generates a collection of
1744 statistics about those objects.
1746 The result (obtained from the /operations/$OPHANDLE page) is a
1747 JSON-serialized dictionary with the following keys (note that some of these
1748 keys may be missing until 'finished' is True)::
1750 finished: (bool) True if the operation has finished, else False
1751 count-immutable-files: count of how many CHK files are in the set
1752 count-mutable-files: same, for mutable files (does not include directories)
1753 count-literal-files: same, for LIT files (data contained inside the URI)
1754 count-files: sum of the above three
1755 count-directories: count of directories
1756 count-unknown: count of unrecognized objects (perhaps from the future)
1757 size-immutable-files: total bytes for all CHK files in the set, =deep-size
1758 size-mutable-files (TODO): same, for current version of all mutable files
1759 size-literal-files: same, for LIT files
1760 size-directories: size of directories (includes size-literal-files)
1761 size-files-histogram: list of (minsize, maxsize, count) buckets,
1762 with a histogram of filesizes, 5dB/bucket,
1763 for both literal and immutable files
1764 largest-directory: number of children in the largest directory
1765 largest-immutable-file: number of bytes in the largest CHK file
1767 size-mutable-files is not implemented, because it would require extra
1768 queries to each mutable file to get their size. This may be implemented in
1771 Assuming no sharing, the basic space consumed by a single root directory is
1772 the sum of size-immutable-files, size-mutable-files, and size-directories.
1773 The actual disk space used by the shares is larger, because of the
1774 following sources of overhead::
1777 expansion due to erasure coding
1778 share management data (leases)
1779 backend (ext3) minimum block size
1781 ``POST $URL?t=stream-manifest``
1783 This operation performs a recursive walk of all files and directories
1784 reachable from the given starting point. For each such unique object
1785 (duplicates are skipped), a single line of JSON is emitted to the HTTP
1786 response channel (or an error indication, see below). When the walk is
1787 complete, a final line of JSON is emitted which contains the accumulated
1788 file-size/count "deep-stats" data.
1790 A CLI tool can split the response stream on newlines into "response units",
1791 and parse each response unit as JSON. Each such parsed unit will be a
1792 dictionary, and will contain at least the "type" key: a string, one of
1793 "file", "directory", or "stats".
1795 For all units that have a type of "file" or "directory", the dictionary will
1796 contain the following keys::
1798 "path": a list of strings, with the path that is traversed to reach the
1800 "cap": a write-cap URI for the file or directory, if available, else a
1802 "verifycap": a verify-cap URI for the file or directory
1803 "repaircap": an URI for the weakest cap that can still be used to repair
1805 "storage-index": a base32 storage index for the object
1807 Note that non-distributed files (i.e. LIT files) will have values of None
1808 for verifycap, repaircap, and storage-index, since these files can neither
1809 be verified nor repaired, and are not stored on the storage servers.
1811 The last unit in the stream will have a type of "stats", and will contain
1812 the keys described in the "start-deep-stats" operation, below.
1814 If any errors occur during the traversal (specifically if a directory is
1815 unrecoverable, such that further traversal is not possible), an error
1816 indication is written to the response body, instead of the usual line of
1817 JSON. This error indication line will begin with the string "ERROR:" (in all
1818 caps), and contain a summary of the error on the rest of the line. The
1819 remaining lines of the response body will be a python exception. The client
1820 application should look for the ERROR: and stop processing JSON as soon as
1821 it is seen. The line just before the ERROR: will describe the directory that
1822 was untraversable, since the manifest entry is emitted to the HTTP response
1823 body before the child is traversed.
1829 The portion of the web namespace that begins with "/uri" (and "/named") is
1830 dedicated to giving users (both humans and programs) access to the Tahoe
1831 virtual filesystem. The rest of the namespace provides status information
1832 about the state of the Tahoe node.
1834 ``GET /`` (the root page)
1836 This is the "Welcome Page", and contains a few distinct sections::
1838 Node information: library versions, local nodeid, services being provided.
1840 Filesystem Access Forms: create a new directory, view a file/directory by
1841 URI, upload a file (unlinked), download a file by
1844 Grid Status: introducer information, helper information, connected storage
1849 This page lists all active uploads and downloads, and contains a short list
1850 of recent upload/download operations. Each operation has a link to a page
1851 that describes file sizes, servers that were involved, and the time consumed
1852 in each phase of the operation.
1854 A GET of /status/?t=json will contain a machine-readable subset of the same
1855 data. It returns a JSON-encoded dictionary. The only key defined at this
1856 time is "active", with a value that is a list of operation dictionaries, one
1857 for each active operation. Once an operation is completed, it will no longer
1858 appear in data["active"] .
1860 Each op-dict contains a "type" key, one of "upload", "download",
1861 "mapupdate", "publish", or "retrieve" (the first two are for immutable
1862 files, while the latter three are for mutable files and directories).
1864 The "upload" op-dict will contain the following keys::
1866 type (string): "upload"
1867 storage-index-string (string): a base32-encoded storage index
1868 total-size (int): total size of the file
1869 status (string): current status of the operation
1870 progress-hash (float): 1.0 when the file has been hashed
1871 progress-ciphertext (float): 1.0 when the file has been encrypted.
1872 progress-encode-push (float): 1.0 when the file has been encoded and
1873 pushed to the storage servers. For helper
1874 uploads, the ciphertext value climbs to 1.0
1875 first, then encoding starts. For unassisted
1876 uploads, ciphertext and encode-push progress
1877 will climb at the same pace.
1879 The "download" op-dict will contain the following keys::
1881 type (string): "download"
1882 storage-index-string (string): a base32-encoded storage index
1883 total-size (int): total size of the file
1884 status (string): current status of the operation
1885 progress (float): 1.0 when the file has been fully downloaded
1887 Front-ends which want to report progress information are advised to simply
1888 average together all the progress-* indicators. A slightly more accurate
1889 value can be found by ignoring the progress-hash value (since the current
1890 implementation hashes synchronously, so clients will probably never see
1891 progress-hash!=1.0).
1893 ``GET /helper_status/``
1895 If the node is running a helper (i.e. if [helper]enabled is set to True in
1896 tahoe.cfg), then this page will provide a list of all the helper operations
1897 currently in progress. If "?t=json" is added to the URL, it will return a
1898 JSON-formatted list of helper statistics, which can then be used to produce
1899 graphs to indicate how busy the helper is.
1901 ``GET /statistics/``
1903 This page provides "node statistics", which are collected from a variety of
1906 load_monitor: every second, the node schedules a timer for one second in
1907 the future, then measures how late the subsequent callback
1908 is. The "load_average" is this tardiness, measured in
1909 seconds, averaged over the last minute. It is an indication
1910 of a busy node, one which is doing more work than can be
1911 completed in a timely fashion. The "max_load" value is the
1912 highest value that has been seen in the last 60 seconds.
1914 cpu_monitor: every minute, the node uses time.clock() to measure how much
1915 CPU time it has used, and it uses this value to produce
1916 1min/5min/15min moving averages. These values range from 0%
1917 (0.0) to 100% (1.0), and indicate what fraction of the CPU
1918 has been used by the Tahoe node. Not all operating systems
1919 provide meaningful data to time.clock(): they may report 100%
1920 CPU usage at all times.
1922 uploader: this counts how many immutable files (and bytes) have been
1923 uploaded since the node was started
1925 downloader: this counts how many immutable files have been downloaded
1926 since the node was started
1928 publishes: this counts how many mutable files (including directories) have
1929 been modified since the node was started
1931 retrieves: this counts how many mutable files (including directories) have
1932 been read since the node was started
1934 There are other statistics that are tracked by the node. The "raw stats"
1935 section shows a formatted dump of all of them.
1937 By adding "?t=json" to the URL, the node will return a JSON-formatted
1938 dictionary of stats values, which can be used by other tools to produce
1939 graphs of node behavior. The misc/munin/ directory in the source
1940 distribution provides some tools to produce these graphs.
1942 ``GET /`` (introducer status)
1944 For Introducer nodes, the welcome page displays information about both
1945 clients and servers which are connected to the introducer. Servers make
1946 "service announcements", and these are listed in a table. Clients will
1947 subscribe to hear about service announcements, and these subscriptions are
1948 listed in a separate table. Both tables contain information about what
1949 version of Tahoe is being run by the remote node, their advertised and
1950 outbound IP addresses, their nodeid and nickname, and how long they have
1953 By adding "?t=json" to the URL, the node will return a JSON-formatted
1954 dictionary of stats values, which can be used to produce graphs of connected
1955 clients over time. This dictionary has the following keys::
1957 ["subscription_summary"] : a dictionary mapping service name (like
1958 "storage") to an integer with the number of
1959 clients that have subscribed to hear about that
1961 ["announcement_summary"] : a dictionary mapping service name to an integer
1962 with the number of servers which are announcing
1964 ["announcement_distinct_hosts"] : a dictionary mapping service name to an
1965 integer which represents the number of
1966 distinct hosts that are providing that
1967 service. If two servers have announced
1968 FURLs which use the same hostnames (but
1969 different ports and tubids), they are
1970 considered to be on the same host.
1973 Static Files in /public_html
1974 ============================
1976 The web-API server will take any request for a URL that starts with /static
1977 and serve it from a configurable directory which defaults to
1978 $BASEDIR/public_html . This is configured by setting the "[node]web.static"
1979 value in $BASEDIR/tahoe.cfg . If this is left at the default value of
1980 "public_html", then http://127.0.0.1:3456/static/subdir/foo.html will be
1981 served with the contents of the file $BASEDIR/public_html/subdir/foo.html .
1983 This can be useful to serve a javascript application which provides a
1984 prettier front-end to the rest of the Tahoe web-API.
1987 Safety and Security Issues -- Names vs. URIs
1988 ============================================
1990 Summary: use explicit file- and dir- caps whenever possible, to reduce the
1991 potential for surprises when the filesystem structure is changed.
1993 Tahoe provides a mutable filesystem, but the ways that the filesystem can
1994 change are limited. The only thing that can change is that the mapping from
1995 child names to child objects that each directory contains can be changed by
1996 adding a new child name pointing to an object, removing an existing child name,
1997 or changing an existing child name to point to a different object.
1999 Obviously if you query Tahoe for information about the filesystem and then act
2000 to change the filesystem (such as by getting a listing of the contents of a
2001 directory and then adding a file to the directory), then the filesystem might
2002 have been changed after you queried it and before you acted upon it. However,
2003 if you use the URI instead of the pathname of an object when you act upon the
2004 object, then the only change that can happen is if the object is a directory
2005 then the set of child names it has might be different. If, on the other hand,
2006 you act upon the object using its pathname, then a different object might be in
2007 that place, which can result in more kinds of surprises.
2009 For example, suppose you are writing code which recursively downloads the
2010 contents of a directory. The first thing your code does is fetch the listing
2011 of the contents of the directory. For each child that it fetched, if that
2012 child is a file then it downloads the file, and if that child is a directory
2013 then it recurses into that directory. Now, if the download and the recurse
2014 actions are performed using the child's name, then the results might be
2015 wrong, because for example a child name that pointed to a sub-directory when
2016 you listed the directory might have been changed to point to a file (in which
2017 case your attempt to recurse into it would result in an error and the file
2018 would be skipped), or a child name that pointed to a file when you listed the
2019 directory might now point to a sub-directory (in which case your attempt to
2020 download the child would result in a file containing HTML text describing the
2023 If your recursive algorithm uses the uri of the child instead of the name of
2024 the child, then those kinds of mistakes just can't happen. Note that both the
2025 child's name and the child's URI are included in the results of listing the
2026 parent directory, so it isn't any harder to use the URI for this purpose.
2028 The read and write caps in a given directory node are separate URIs, and
2029 can't be assumed to point to the same object even if they were retrieved in
2030 the same operation (although the web-API server attempts to ensure this
2031 in most cases). If you need to rely on that property, you should explicitly
2032 verify it. More generally, you should not make assumptions about the
2033 internal consistency of the contents of mutable directories. As a result
2034 of the signatures on mutable object versions, it is guaranteed that a given
2035 version was written in a single update, but -- as in the case of a file --
2036 the contents may have been chosen by a malicious writer in a way that is
2037 designed to confuse applications that rely on their consistency.
2039 In general, use names if you want "whatever object (whether file or
2040 directory) is found by following this name (or sequence of names) when my
2041 request reaches the server". Use URIs if you want "this particular object".
2047 Tahoe uses both mutable and immutable files. Mutable files can be created
2048 explicitly by doing an upload with ?mutable=true added, or implicitly by
2049 creating a new directory (since a directory is just a special way to
2050 interpret a given mutable file).
2052 Mutable files suffer from the same consistency-vs-availability tradeoff that
2053 all distributed data storage systems face. It is not possible to
2054 simultaneously achieve perfect consistency and perfect availability in the
2055 face of network partitions (servers being unreachable or faulty).
2057 Tahoe tries to achieve a reasonable compromise, but there is a basic rule in
2058 place, known as the Prime Coordination Directive: "Don't Do That". What this
2059 means is that if write-access to a mutable file is available to several
2060 parties, then those parties are responsible for coordinating their activities
2061 to avoid multiple simultaneous updates. This could be achieved by having
2062 these parties talk to each other and using some sort of locking mechanism, or
2063 by serializing all changes through a single writer.
2065 The consequences of performing uncoordinated writes can vary. Some of the
2066 writers may lose their changes, as somebody else wins the race condition. In
2067 many cases the file will be left in an "unhealthy" state, meaning that there
2068 are not as many redundant shares as we would like (reducing the reliability
2069 of the file against server failures). In the worst case, the file can be left
2070 in such an unhealthy state that no version is recoverable, even the old ones.
2071 It is this small possibility of data loss that prompts us to issue the Prime
2072 Coordination Directive.
2074 Tahoe nodes implement internal serialization to make sure that a single Tahoe
2075 node cannot conflict with itself. For example, it is safe to issue two
2076 directory modification requests to a single tahoe node's web-API server at the
2077 same time, because the Tahoe node will internally delay one of them until
2078 after the other has finished being applied. (This feature was introduced in
2079 Tahoe-1.1; back with Tahoe-1.0 the web client was responsible for serializing
2080 web requests themselves).
2082 For more details, please see the "Consistency vs Availability" and "The Prime
2083 Coordination Directive" sections of mutable.rst_.
2085 .. _mutable.rst: ../specifications/mutable.rst
2091 Gateway nodes may find it necessary to prohibit access to certain files. The
2092 web-API has a facility to block access to filecaps by their storage index,
2093 returning a 403 "Forbidden" error instead of the original file.
2095 This blacklist is recorded in $NODEDIR/access.blacklist, and contains one
2096 blocked file per line. Comment lines (starting with ``#``) are ignored. Each
2097 line consists of the storage-index (in the usual base32 format as displayed
2098 by the "More Info" page, or by the "tahoe debug dump-cap" command), followed
2099 by whitespace, followed by a reason string, which will be included in the 403
2100 error message. This could hold a URL to a page that explains why the file is
2101 blocked, for example.
2103 So for example, if you found a need to block access to a file with filecap
2104 ``URI:CHK:n7r3m6wmomelk4sep3kw5cvduq:os7ijw5c3maek7pg65e5254k2fzjflavtpejjyhshpsxuqzhcwwq:3:20:14861``,
2105 you could do the following::
2107 tahoe debug dump-cap URI:CHK:n7r3m6wmomelk4sep3kw5cvduq:os7ijw5c3maek7pg65e5254k2fzjflavtpejjyhshpsxuqzhcwwq:3:20:14861
2108 -> storage index: whpepioyrnff7orecjolvbudeu
2109 echo "whpepioyrnff7orecjolvbudeu my puppy told me to" >>$NODEDIR/access.blacklist
2110 tahoe restart $NODEDIR
2111 tahoe get URI:CHK:n7r3m6wmomelk4sep3kw5cvduq:os7ijw5c3maek7pg65e5254k2fzjflavtpejjyhshpsxuqzhcwwq:3:20:14861
2112 -> error, 403 Access Prohibited: my puppy told me to
2114 The ``access.blacklist`` file will be checked each time a file or directory
2115 is accessed: the file's ``mtime`` is used to decide whether it need to be
2116 reloaded. Therefore no node restart is necessary when creating the initial
2117 blacklist, nor when adding second, third, or additional entries to the list.
2118 When modifying the file, be careful to update it atomically, otherwise a
2119 request may arrive while the file is only halfway written, and the partial
2120 file may be incorrectly parsed.
2122 The blacklist is applied to all access paths (including SFTP, FTP, and CLI
2123 operations), not just the web-API. The blacklist also applies to directories.
2124 If a directory is blacklisted, the gateway will refuse access to both that
2125 directory and any child files/directories underneath it, when accessed via
2126 "DIRCAP/SUBDIR/FILENAME" -style URLs. Users who go directly to the child
2127 file/dir will bypass the blacklist.
2129 The node will log the SI of the file being blocked, and the reason code, into
2130 the ``logs/twistd.log`` file.
2133 .. [1] URLs and HTTP and UTF-8, Oh My
2135 HTTP does not provide a mechanism to specify the character set used to
2136 encode non-ASCII names in URLs (`RFC3986#2.1`_). We prefer the convention
2137 that the ``filename=`` argument shall be a URL-escaped UTF-8 encoded Unicode
2138 string. For example, suppose we want to provoke the server into using a
2139 filename of "f i a n c e-acute e" (i.e. f i a n c U+00E9 e). The UTF-8
2140 encoding of this is 0x66 0x69 0x61 0x6e 0x63 0xc3 0xa9 0x65 (or
2141 "fianc\\xC3\\xA9e", as python's ``repr()`` function would show). To encode
2142 this into a URL, the non-printable characters must be escaped with the
2143 urlencode ``%XX`` mechanism, giving us "fianc%C3%A9e". Thus, the first line
2144 of the HTTP request will be "``GET
2145 /uri/CAP...?save=true&filename=fianc%C3%A9e HTTP/1.1``". Not all browsers
2146 provide this: IE7 by default uses the Latin-1 encoding, which is "fianc%E9e"
2147 (although it has a configuration option to send URLs as UTF-8).
2149 The response header will need to indicate a non-ASCII filename. The actual
2150 mechanism to do this is not clear. For ASCII filenames, the response header
2153 Content-Disposition: attachment; filename="english.txt"
2155 If Tahoe were to enforce the UTF-8 convention, it would need to decode the
2156 URL argument into a Unicode string, and then encode it back into a sequence
2157 of bytes when creating the response header. One possibility would be to use
2158 unencoded UTF-8. Developers suggest that IE7 might accept this::
2160 #1: Content-Disposition: attachment; filename="fianc\xC3\xA9e"
2161 (note, the last four bytes of that line, not including the newline, are
2162 0xC3 0xA9 0x65 0x22)
2164 `RFC2231#4`_ (dated 1997): suggests that the following might work, and `some
2165 developers have reported`_ that it is supported by Firefox (but not IE7)::
2167 #2: Content-Disposition: attachment; filename*=utf-8''fianc%C3%A9e
2169 My reading of `RFC2616#19.5.1`_ (which defines Content-Disposition) says
2170 that the filename= parameter is defined to be wrapped in quotes (presumably
2171 to allow spaces without breaking the parsing of subsequent parameters),
2172 which would give us::
2174 #3: Content-Disposition: attachment; filename*=utf-8''"fianc%C3%A9e"
2176 However this is contrary to the examples in the email thread listed above.
2178 Developers report that IE7 (when it is configured for UTF-8 URL encoding,
2179 which is not the default in Asian countries), will accept::
2181 #4: Content-Disposition: attachment; filename=fianc%C3%A9e
2183 However, for maximum compatibility, Tahoe simply copies bytes from the URL
2184 into the response header, rather than enforcing the UTF-8 convention. This
2185 means it does not try to decode the filename from the URL argument, nor does
2186 it encode the filename into the response header.
2188 .. _RFC3986#2.1: https://tools.ietf.org/html/rfc3986#section-2.1
2189 .. _RFC2231#4: https://tools.ietf.org/html/rfc2231#section-4
2190 .. _some developers have reported: http://markmail.org/message/dsjyokgl7hv64ig3
2191 .. _RFC2616#19.5.1: https://tools.ietf.org/html/rfc2616#section-19.5.1