1 ==========================
2 The Tahoe REST-ful Web API
3 ==========================
5 1. `Enabling the web-API port`_
6 2. `Basic Concepts: GET, PUT, DELETE, POST`_
11 4. `Slow Operations, Progress, and Cancelling`_
12 5. `Programmatic Operations`_
15 2. `Writing/Uploading a File`_
16 3. `Creating a New Directory`_
17 4. `Getting Information About a File Or Directory (as JSON)`_
18 5. `Attaching an Existing File or Directory by its read- or write-cap`_
19 6. `Adding Multiple Files or Directories to a Parent Directory at Once`_
20 7. `Unlinking a File or Directory`_
22 6. `Browser Operations: Human-Oriented Interfaces`_
24 1. `Viewing a Directory (as HTML)`_
25 2. `Viewing/Downloading a File`_
26 3. `Getting Information About a File Or Directory (as HTML)`_
27 4. `Creating a Directory`_
28 5. `Uploading a File`_
29 6. `Attaching an Existing File Or Directory (by URI)`_
30 7. `Unlinking a Child`_
31 8. `Renaming a Child`_
32 9. `Relinking ("Moving") a Child`_
33 10. `Other Utilities`_
34 11. `Debugging and Testing Features`_
36 7. `Other Useful Pages`_
37 8. `Static Files in /public_html`_
38 9. `Safety and Security Issues -- Names vs. URIs`_
39 10. `Concurrency Issues`_
40 11. `Access Blacklist`_
43 Enabling the web-API port
44 =========================
46 Every Tahoe node is capable of running a built-in HTTP server. To enable
47 this, just write a port number into the "[node]web.port" line of your node's
48 tahoe.cfg file. For example, writing "web.port = 3456" into the "[node]"
49 section of $NODEDIR/tahoe.cfg will cause the node to run a webserver on port
52 This string is actually a Twisted "strports" specification, meaning you can
53 get more control over the interface to which the server binds by supplying
54 additional arguments. For more details, see the documentation on
55 `twisted.application.strports
56 <https://twistedmatrix.com/documents/current/api/twisted.application.strports.html>`_.
58 Writing "tcp:3456:interface=127.0.0.1" into the web.port line does the same
59 but binds to the loopback interface, ensuring that only the programs on the
60 local host can connect. Using "ssl:3456:privateKey=mykey.pem:certKey=cert.pem"
63 This webport can be set when the node is created by passing a --webport
64 option to the 'tahoe create-node' command. By default, the node listens on
65 port 3456, on the loopback (127.0.0.1) interface.
68 Basic Concepts: GET, PUT, DELETE, POST
69 ======================================
71 As described in `docs/architecture.rst <../architecture.rst>`_, each file
72 and directory in a Tahoe virtual filesystem is referenced by an identifier
73 that combines the designation of the object with the authority to do something
74 with it (such as read or modify the contents). This identifier is called a
75 "read-cap" or "write-cap", depending upon whether it enables read-only or
76 read-write access. These "caps" are also referred to as URIs (which may be
77 confusing because they are not currently `RFC3986
78 <https://tools.ietf.org/html/rfc3986>`_-compliant URIs).
80 The Tahoe web-based API is "REST-ful", meaning it implements the concepts of
81 "REpresentational State Transfer": the original scheme by which the World
82 Wide Web was intended to work. Each object (file or directory) is referenced
83 by a URL that includes the read- or write- cap. HTTP methods (GET, PUT, and
84 DELETE) are used to manipulate these objects. You can think of the URL as a
85 noun, and the method as a verb.
87 In REST, the GET method is used to retrieve information about an object, or
88 to retrieve some representation of the object itself. When the object is a
89 file, the basic GET method will simply return the contents of that file.
90 Other variations (generally implemented by adding query parameters to the
91 URL) will return information about the object, such as metadata. GET
92 operations are required to have no side-effects.
94 PUT is used to upload new objects into the filesystem, or to replace an
95 existing link or the contents of a mutable file. DELETE is used to unlink
96 objects from directories. Both PUT and DELETE are required to be idempotent:
97 performing the same operation multiple times must have the same side-effects
98 as only performing it once.
100 POST is used for more complicated actions that cannot be expressed as a GET,
101 PUT, or DELETE. POST operations can be thought of as a method call: sending
102 some message to the object referenced by the URL. In Tahoe, POST is also used
103 for operations that must be triggered by an HTML form (including upload and
104 unlinking), because otherwise a regular web browser has no way to accomplish
105 these tasks. In general, everything that can be done with a PUT or DELETE can
106 also be done with a POST.
108 Tahoe's web API is designed for two different kinds of consumer. The first is
109 a program that needs to manipulate the virtual file system. Such programs are
110 expected to use the RESTful interface described above. The second is a human
111 using a standard web browser to work with the filesystem. This user is given
112 a series of HTML pages with links to download files, and forms that use POST
113 actions to upload, rename, and unlink files.
115 When an error occurs, the HTTP response code will be set to an appropriate
116 400-series code (like 404 Not Found for an unknown childname, or 400 Bad Request
117 when the parameters to a web-API operation are invalid), and the HTTP response
118 body will usually contain a few lines of explanation as to the cause of the
119 error and possible responses. Unusual exceptions may result in a 500 Internal
120 Server Error as a catch-all, with a default response body containing
121 a Nevow-generated HTML-ized representation of the Python exception stack trace
122 that caused the problem. CLI programs which want to copy the response body to
123 stderr should provide an "Accept: text/plain" header to their requests to get
124 a plain text stack trace instead. If the Accept header contains ``*/*``, or
125 ``text/*``, or text/html (or if there is no Accept header), HTML tracebacks will
132 Tahoe uses a variety of read- and write- caps to identify files and
133 directories. The most common of these is the "immutable file read-cap", which
134 is used for most uploaded files. These read-caps look like the following::
136 URI:CHK:ime6pvkaxuetdfah2p2f35pe54:4btz54xk3tew6nd4y2ojpxj4m6wxjqqlwnztgre6gnjgtucd5r4a:3:10:202
138 The next most common is a "directory write-cap", which provides both read and
139 write access to a directory, and look like this::
141 URI:DIR2:djrdkfawoqihigoett4g6auz6a:jx5mplfpwexnoqff7y5e4zjus4lidm76dcuarpct7cckorh2dpgq
143 There are also "directory read-caps", which start with "URI:DIR2-RO:", and
144 give read-only access to a directory. Finally there are also mutable file
145 read- and write- caps, which start with "URI:SSK", and give access to mutable
148 (Later versions of Tahoe will make these strings shorter, and will remove the
149 unfortunate colons, which must be escaped when these caps are embedded in
152 To refer to any Tahoe object through the web API, you simply need to combine
153 a prefix (which indicates the HTTP server to use) with the cap (which
154 indicates which object inside that server to access). Since the default Tahoe
155 webport is 3456, the most common prefix is one that will use a local node
156 listening on this port::
158 http://127.0.0.1:3456/uri/ + $CAP
160 So, to access the directory named above, the URL would be::
162 http://127.0.0.1:3456/uri/URI%3ADIR2%3Adjrdkfawoqihigoett4g6auz6a%3Ajx5mplfpwexnoqff7y5e4zjus4lidm76dcuarpct7cckorh2dpgq/
164 (note that the colons in the directory-cap are url-encoded into "%3A"
167 Likewise, to access the file named above, use::
169 http://127.0.0.1:3456/uri/URI%3ACHK%3Aime6pvkaxuetdfah2p2f35pe54%3A4btz54xk3tew6nd4y2ojpxj4m6wxjqqlwnztgre6gnjgtucd5r4a%3A3%3A10%3A202
171 In the rest of this document, we'll use "$DIRCAP" as shorthand for a read-cap
172 or write-cap that refers to a directory, and "$FILECAP" to abbreviate a cap
173 that refers to a file (whether mutable or immutable). So those URLs above can
176 http://127.0.0.1:3456/uri/$DIRCAP/
177 http://127.0.0.1:3456/uri/$FILECAP
179 The operation summaries below will abbreviate these further, by eliding the
180 server prefix. They will be displayed like this::
185 /cap can be used as a synonym for /uri. If interoperability with older web-API
186 servers is required, /uri should be used.
191 Tahoe directories contain named child entries, just like directories in a regular
192 local filesystem. These child entries, called "dirnodes", consist of a name,
193 metadata, a write slot, and a read slot. The write and read slots normally contain
194 a write-cap and read-cap referring to the same object, which can be either a file
195 or a subdirectory. The write slot may be empty (actually, both may be empty,
196 but that is unusual).
198 If you have a Tahoe URL that refers to a directory, and want to reference a
199 named child inside it, just append the child name to the URL. For example, if
200 our sample directory contains a file named "welcome.txt", we can refer to
203 http://127.0.0.1:3456/uri/$DIRCAP/welcome.txt
205 (or http://127.0.0.1:3456/uri/URI%3ADIR2%3Adjrdkfawoqihigoett4g6auz6a%3Ajx5mplfpwexnoqff7y5e4zjus4lidm76dcuarpct7cckorh2dpgq/welcome.txt)
207 Multiple levels of subdirectories can be handled this way::
209 http://127.0.0.1:3456/uri/$DIRCAP/tahoe-source/docs/architecture.rst
211 In this document, when we need to refer to a URL that references a file using
212 this child-of-some-directory format, we'll use the following string::
214 /uri/$DIRCAP/[SUBDIRS../]FILENAME
216 The "[SUBDIRS../]" part means that there are zero or more (optional)
217 subdirectory names in the middle of the URL. The "FILENAME" at the end means
218 that this whole URL refers to a file of some sort, rather than to a
221 When we need to refer specifically to a directory in this way, we'll write::
223 /uri/$DIRCAP/[SUBDIRS../]SUBDIR
226 Note that all components of pathnames in URLs are required to be UTF-8
227 encoded, so "resume.doc" (with an acute accent on both E's) would be accessed
230 http://127.0.0.1:3456/uri/$DIRCAP/r%C3%A9sum%C3%A9.doc
232 Also note that the filenames inside upload POST forms are interpreted using
233 whatever character set was provided in the conventional '_charset' field, and
234 defaults to UTF-8 if not otherwise specified. The JSON representation of each
235 directory contains native Unicode strings. Tahoe directories are specified to
236 contain Unicode filenames, and cannot contain binary strings that are not
237 representable as such.
239 All Tahoe operations that refer to existing files or directories must include
240 a suitable read- or write- cap in the URL: the web-API server won't add one
241 for you. If you don't know the cap, you can't access the file. This allows
242 the security properties of Tahoe caps to be extended across the web-API
246 Slow Operations, Progress, and Cancelling
247 =========================================
249 Certain operations can be expected to take a long time. The "t=deep-check",
250 described below, will recursively visit every file and directory reachable
251 from a given starting point, which can take minutes or even hours for
252 extremely large directory structures. A single long-running HTTP request is a
253 fragile thing: proxies, NAT boxes, browsers, and users may all grow impatient
254 with waiting and give up on the connection.
256 For this reason, long-running operations have an "operation handle", which
257 can be used to poll for status/progress messages while the operation
258 proceeds. This handle can also be used to cancel the operation. These handles
259 are created by the client, and passed in as a an "ophandle=" query argument
260 to the POST or PUT request which starts the operation. The following
261 operations can then be used to retrieve status:
263 ``GET /operations/$HANDLE?output=HTML (with or without t=status)``
265 ``GET /operations/$HANDLE?output=JSON (same)``
267 These two retrieve the current status of the given operation. Each operation
268 presents a different sort of information, but in general the page retrieved
271 * whether the operation is complete, or if it is still running
272 * how much of the operation is complete, and how much is left, if possible
274 Note that the final status output can be quite large: a deep-manifest of a
275 directory structure with 300k directories and 200k unique files is about
276 275MB of JSON, and might take two minutes to generate. For this reason, the
277 full status is not provided until the operation has completed.
279 The HTML form will include a meta-refresh tag, which will cause a regular
280 web browser to reload the status page about 60 seconds later. This tag will
281 be removed once the operation has completed.
283 There may be more status information available under
284 /operations/$HANDLE/$ETC : i.e., the handle forms the root of a URL space.
286 ``POST /operations/$HANDLE?t=cancel``
288 This terminates the operation, and returns an HTML page explaining what was
289 cancelled. If the operation handle has already expired (see below), this
290 POST will return a 404, which indicates that the operation is no longer
291 running (either it was completed or terminated). The response body will be
292 the same as a GET /operations/$HANDLE on this operation handle, and the
293 handle will be expired immediately afterwards.
295 The operation handle will eventually expire, to avoid consuming an unbounded
296 amount of memory. The handle's time-to-live can be reset at any time, by
297 passing a retain-for= argument (with a count of seconds) to either the
298 initial POST that starts the operation, or the subsequent GET request which
299 asks about the operation. For example, if a 'GET
300 /operations/$HANDLE?output=JSON&retain-for=600' query is performed, the
301 handle will remain active for 600 seconds (10 minutes) after the GET was
304 In addition, if the GET includes a release-after-complete=True argument, and
305 the operation has completed, the operation handle will be released
308 If a retain-for= argument is not used, the default handle lifetimes are:
310 * handles will remain valid at least until their operation finishes
311 * uncollected handles for finished operations (i.e. handles for
312 operations that have finished but for which the GET page has not been
313 accessed since completion) will remain valid for four days, or for
314 the total time consumed by the operation, whichever is greater.
315 * collected handles (i.e. the GET page has been retrieved at least once
316 since the operation completed) will remain valid for one day.
318 Many "slow" operations can begin to use unacceptable amounts of memory when
319 operating on large directory structures. The memory usage increases when the
320 ophandle is polled, as the results must be copied into a JSON string, sent
321 over the wire, then parsed by a client. So, as an alternative, many "slow"
322 operations have streaming equivalents. These equivalents do not use operation
323 handles. Instead, they emit line-oriented status results immediately. Client
324 code can cancel the operation by simply closing the HTTP connection.
327 Programmatic Operations
328 =======================
330 Now that we know how to build URLs that refer to files and directories in a
331 Tahoe virtual filesystem, what sorts of operations can we do with those URLs?
332 This section contains a catalog of GET, PUT, DELETE, and POST operations that
333 can be performed on these URLs. This set of operations are aimed at programs
334 that use HTTP to communicate with a Tahoe node. A later section describes
335 operations that are intended for web browsers.
341 ``GET /uri/$FILECAP``
343 ``GET /uri/$DIRCAP/[SUBDIRS../]FILENAME``
345 This will retrieve the contents of the given file. The HTTP response body
346 will contain the sequence of bytes that make up the file.
348 To view files in a web browser, you may want more control over the
349 Content-Type and Content-Disposition headers. Please see the next section
350 "Browser Operations", for details on how to modify these URLs for that
354 Writing/Uploading a File
355 ------------------------
357 ``PUT /uri/$FILECAP``
359 ``PUT /uri/$DIRCAP/[SUBDIRS../]FILENAME``
361 Upload a file, using the data from the HTTP request body, and add whatever
362 child links and subdirectories are necessary to make the file available at
363 the given location. Once this operation succeeds, a GET on the same URL will
364 retrieve the same contents that were just uploaded. This will create any
365 necessary intermediate subdirectories.
367 To use the /uri/$FILECAP form, $FILECAP must be a write-cap for a mutable file.
369 In the /uri/$DIRCAP/[SUBDIRS../]FILENAME form, if the target file is a
370 writeable mutable file, that file's contents will be overwritten
371 in-place. If it is a read-cap for a mutable file, an error will occur.
372 If it is an immutable file, the old file will be discarded, and a new
373 one will be put in its place. If the target file is a writable mutable
374 file, you may also specify an "offset" parameter -- a byte offset that
375 determines where in the mutable file the data from the HTTP request
376 body is placed. This operation is relatively efficient for MDMF mutable
377 files, and is relatively inefficient (but still supported) for SDMF
378 mutable files. If no offset parameter is specified, then the entire
379 file is replaced with the data from the HTTP request body. For an
380 immutable file, the "offset" parameter is not valid.
382 When creating a new file, you can control the type of file created by
383 specifying a format= argument in the query string. format=MDMF creates an
384 MDMF mutable file. format=SDMF creates an SDMF mutable file. format=CHK
385 creates an immutable file. The value of the format argument is
386 case-insensitive. If no format is specified, the newly-created file will be
387 immutable (but see below).
389 For compatibility with previous versions of Tahoe-LAFS, the web-API will
390 also accept a mutable=true argument in the query string. If mutable=true is
391 given, then the new file will be mutable, and its format will be the default
392 mutable file format, as configured by the [client]mutable.format option of
393 tahoe.cfg on the Tahoe-LAFS node hosting the webapi server. Use of
394 mutable=true is discouraged; new code should use format= instead of
395 mutable=true (unless it needs to be compatible with web-API servers older
396 than v1.9.0). If neither format= nor mutable=true are given, the
397 newly-created file will be immutable.
399 This returns the file-cap of the resulting file. If a new file was created
400 by this method, the HTTP response code (as dictated by rfc2616) will be set
401 to 201 CREATED. If an existing file was replaced or modified, the response
404 Note that the 'curl -T localfile http://127.0.0.1:3456/uri/$DIRCAP/foo.txt'
405 command can be used to invoke this operation.
409 This uploads a file, and produces a file-cap for the contents, but does not
410 attach the file into the filesystem. No directories will be modified by
411 this operation. The file-cap is returned as the body of the HTTP response.
413 This method accepts format= and mutable=true as query string arguments, and
414 interprets those arguments in the same way as the linked forms of PUT
415 described immediately above.
417 Creating a New Directory
418 ------------------------
420 ``POST /uri?t=mkdir``
424 Create a new empty directory and return its write-cap as the HTTP response
425 body. This does not make the newly created directory visible from the
426 filesystem. The "PUT" operation is provided for backwards compatibility:
427 new code should use POST.
429 This supports a format= argument in the query string. The format=
430 argument, if specified, controls the format of the directory. format=MDMF
431 indicates that the directory should be stored as an MDMF file; format=SDMF
432 indicates that the directory should be stored as an SDMF file. The value of
433 the format= argument is case-insensitive. If no format= argument is
434 given, the directory's format is determined by the default mutable file
435 format, as configured on the Tahoe-LAFS node responding to the request.
437 ``POST /uri?t=mkdir-with-children``
439 Create a new directory, populated with a set of child nodes, and return its
440 write-cap as the HTTP response body. The new directory is not attached to
441 any other directory: the returned write-cap is the only reference to it.
443 The format of the directory can be controlled with the format= argument in
444 the query string, as described above.
446 Initial children are provided as the body of the POST form (this is more
447 efficient than doing separate mkdir and set_children operations). If the
448 body is empty, the new directory will be empty. If not empty, the body will
449 be interpreted as a UTF-8 JSON-encoded dictionary of children with which the
450 new directory should be populated, using the same format as would be
451 returned in the 'children' value of the t=json GET request, described below.
452 Each dictionary key should be a child name, and each value should be a list
453 of [TYPE, PROPDICT], where PROPDICT contains "rw_uri", "ro_uri", and
454 "metadata" keys (all others are ignored). For example, the PUT request body
458 "Fran\u00e7ais": [ "filenode", {
459 "ro_uri": "URI:CHK:...",
461 "ctime": 1202777696.7564139,
462 "mtime": 1202777696.7564139,
464 "linkcrtime": 1202777696.7564139,
465 "linkmotime": 1202777696.7564139
467 "subdir": [ "dirnode", {
468 "rw_uri": "URI:DIR2:...",
469 "ro_uri": "URI:DIR2-RO:...",
471 "ctime": 1202778102.7589991,
472 "mtime": 1202778111.2160511,
474 "linkcrtime": 1202777696.7564139,
475 "linkmotime": 1202777696.7564139
479 For forward-compatibility, a mutable directory can also contain caps in
480 a format that is unknown to the web-API server. When such caps are retrieved
481 from a mutable directory in a "ro_uri" field, they will be prefixed with
482 the string "ro.", indicating that they must not be decoded without
483 checking that they are read-only. The "ro." prefix must not be stripped
484 off without performing this check. (Future versions of the web-API server
485 will perform it where necessary.)
487 If both the "rw_uri" and "ro_uri" fields are present in a given PROPDICT,
488 and the web-API server recognizes the rw_uri as a write cap, then it will
489 reset the ro_uri to the corresponding read cap and discard the original
490 contents of ro_uri (in order to ensure that the two caps correspond to the
491 same object and that the ro_uri is in fact read-only). However this may not
492 happen for caps in a format unknown to the web-API server. Therefore, when
493 writing a directory the web-API client should ensure that the contents
494 of "rw_uri" and "ro_uri" for a given PROPDICT are a consistent
495 (write cap, read cap) pair if possible. If the web-API client only has
496 one cap and does not know whether it is a write cap or read cap, then
497 it is acceptable to set "rw_uri" to that cap and omit "ro_uri". The
498 client must not put a write cap into a "ro_uri" field.
500 The metadata may have a "no-write" field. If this is set to true in the
501 metadata of a link, it will not be possible to open that link for writing
502 via the SFTP frontend; see `<FTP-and-SFTP.rst>`_ for details.
503 Also, if the "no-write" field is set to true in the metadata of a link to
504 a mutable child, it will cause the link to be diminished to read-only.
506 Note that the web-API-using client application must not provide the
507 "Content-Type: multipart/form-data" header that usually accompanies HTML
508 form submissions, since the body is not formatted this way. Doing so will
509 cause a server error as the lower-level code misparses the request body.
511 Child file names should each be expressed as a Unicode string, then used as
512 keys of the dictionary. The dictionary should then be converted into JSON,
513 and the resulting string encoded into UTF-8. This UTF-8 bytestring should
514 then be used as the POST body.
516 ``POST /uri?t=mkdir-immutable``
518 Like t=mkdir-with-children above, but the new directory will be
519 deep-immutable. This means that the directory itself is immutable, and that
520 it can only contain objects that are treated as being deep-immutable, like
521 immutable files, literal files, and deep-immutable directories.
523 For forward-compatibility, a deep-immutable directory can also contain caps
524 in a format that is unknown to the web-API server. When such caps are retrieved
525 from a deep-immutable directory in a "ro_uri" field, they will be prefixed
526 with the string "imm.", indicating that they must not be decoded without
527 checking that they are immutable. The "imm." prefix must not be stripped
528 off without performing this check. (Future versions of the web-API server
529 will perform it where necessary.)
531 The cap for each child may be given either in the "rw_uri" or "ro_uri"
532 field of the PROPDICT (not both). If a cap is given in the "rw_uri" field,
533 then the web-API server will check that it is an immutable read-cap of a
534 *known* format, and give an error if it is not. If a cap is given in the
535 "ro_uri" field, then the web-API server will still check whether known
536 caps are immutable, but for unknown caps it will simply assume that the
537 cap can be stored, as described above. Note that an attacker would be
538 able to store any cap in an immutable directory, so this check when
539 creating the directory is only to help non-malicious clients to avoid
540 accidentally giving away more authority than intended.
542 A non-empty request body is mandatory, since after the directory is created,
543 it will not be possible to add more children to it.
545 ``POST /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=mkdir``
547 ``PUT /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=mkdir``
549 Create new directories as necessary to make sure that the named target
550 ($DIRCAP/SUBDIRS../SUBDIR) is a directory. This will create additional
551 intermediate mutable directories as necessary. If the named target directory
552 already exists, this will make no changes to it.
554 If the final directory is created, it will be empty.
556 This accepts a format= argument in the query string, which controls the
557 format of the named target directory, if it does not already exist. format=
558 is interpreted in the same way as in the POST /uri?t=mkdir form. Note that
559 format= only controls the format of the named target directory;
560 intermediate directories, if created, are created based on the default
561 mutable type, as configured on the Tahoe-LAFS server responding to the
564 This operation will return an error if a blocking file is present at any of
565 the parent names, preventing the server from creating the necessary parent
566 directory; or if it would require changing an immutable directory.
568 The write-cap of the new directory will be returned as the HTTP response
571 ``POST /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=mkdir-with-children``
573 Like /uri?t=mkdir-with-children, but the final directory is created as a
574 child of an existing mutable directory. This will create additional
575 intermediate mutable directories as necessary. If the final directory is
576 created, it will be populated with initial children from the POST request
577 body, as described above.
579 This accepts a format= argument in the query string, which controls the
580 format of the target directory, if the target directory is created as part
581 of the operation. format= is interpreted in the same way as in the POST/
582 uri?t=mkdir-with-children operation. Note that format= only controls the
583 format of the named target directory; intermediate directories, if created,
584 are created using the default mutable type setting, as configured on the
585 Tahoe-LAFS server responding to the request.
587 This operation will return an error if a blocking file is present at any of
588 the parent names, preventing the server from creating the necessary parent
589 directory; or if it would require changing an immutable directory; or if
590 the immediate parent directory already has a a child named SUBDIR.
592 ``POST /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=mkdir-immutable``
594 Like /uri?t=mkdir-immutable, but the final directory is created as a child
595 of an existing mutable directory. The final directory will be deep-immutable,
596 and will be populated with the children specified as a JSON dictionary in
597 the POST request body.
599 In Tahoe 1.6 this operation creates intermediate mutable directories if
600 necessary, but that behaviour should not be relied on; see ticket #920.
602 This operation will return an error if the parent directory is immutable,
603 or already has a child named SUBDIR.
605 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=mkdir&name=NAME``
607 Create a new empty mutable directory and attach it to the given existing
608 directory. This will create additional intermediate directories as necessary.
610 This accepts a format= argument in the query string, which controls the
611 format of the named target directory, if it does not already exist. format=
612 is interpreted in the same way as in the POST /uri?t=mkdir form. Note that
613 format= only controls the format of the named target directory;
614 intermediate directories, if created, are created based on the default
615 mutable type, as configured on the Tahoe-LAFS server responding to the
618 This operation will return an error if a blocking file is present at any of
619 the parent names, preventing the server from creating the necessary parent
620 directory, or if it would require changing any immutable directory.
622 The URL of this operation points to the parent of the bottommost new directory,
623 whereas the /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=mkdir operation above has a URL
624 that points directly to the bottommost new directory.
626 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=mkdir-with-children&name=NAME``
628 Like /uri/$DIRCAP/[SUBDIRS../]?t=mkdir&name=NAME, but the new directory will
629 be populated with initial children via the POST request body. This command
630 will create additional intermediate mutable directories as necessary.
632 This accepts a format= argument in the query string, which controls the
633 format of the target directory, if the target directory is created as part
634 of the operation. format= is interpreted in the same way as in the POST/
635 uri?t=mkdir-with-children operation. Note that format= only controls the
636 format of the named target directory; intermediate directories, if created,
637 are created using the default mutable type setting, as configured on the
638 Tahoe-LAFS server responding to the request.
640 This operation will return an error if a blocking file is present at any of
641 the parent names, preventing the server from creating the necessary parent
642 directory; or if it would require changing an immutable directory; or if
643 the immediate parent directory already has a a child named NAME.
645 Note that the name= argument must be passed as a queryarg, because the POST
646 request body is used for the initial children JSON.
648 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=mkdir-immutable&name=NAME``
650 Like /uri/$DIRCAP/[SUBDIRS../]?t=mkdir-with-children&name=NAME, but the
651 final directory will be deep-immutable. The children are specified as a
652 JSON dictionary in the POST request body. Again, the name= argument must be
653 passed as a queryarg.
655 In Tahoe 1.6 this operation creates intermediate mutable directories if
656 necessary, but that behaviour should not be relied on; see ticket #920.
658 This operation will return an error if the parent directory is immutable,
659 or already has a child named NAME.
662 Getting Information About a File Or Directory (as JSON)
663 -------------------------------------------------------
665 ``GET /uri/$FILECAP?t=json``
667 ``GET /uri/$DIRCAP?t=json``
669 ``GET /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=json``
671 ``GET /uri/$DIRCAP/[SUBDIRS../]FILENAME?t=json``
673 This returns a machine-parseable JSON-encoded description of the given
674 object. The JSON always contains a list, and the first element of the list is
675 always a flag that indicates whether the referenced object is a file or a
676 directory. If it is a capability to a file, then the information includes
677 file size and URI, like this::
679 GET /uri/$FILECAP?t=json :
683 "verify_uri": verify_uri,
689 If it is a capability to a directory followed by a path from that directory
690 to a file, then the information also includes metadata from the link to the
691 file in the parent directory, like this::
693 GET /uri/$DIRCAP/[SUBDIRS../]FILENAME?t=json
697 "verify_uri": verify_uri,
702 "ctime": 1202777696.7564139,
703 "mtime": 1202777696.7564139,
705 "linkcrtime": 1202777696.7564139,
706 "linkmotime": 1202777696.7564139
709 If it is a directory, then it includes information about the children of
710 this directory, as a mapping from child name to a set of data about the
711 child (the same data that would appear in a corresponding GET?t=json of the
712 child itself). The child entries also include metadata about each child,
713 including link-creation- and link-change- timestamps. The output looks like
716 GET /uri/$DIRCAP?t=json :
717 GET /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=json :
720 "rw_uri": read_write_uri,
721 "ro_uri": read_only_uri,
722 "verify_uri": verify_uri,
726 "foo.txt": [ "filenode",
731 "ctime": 1202777696.7564139,
732 "mtime": 1202777696.7564139,
734 "linkcrtime": 1202777696.7564139,
735 "linkmotime": 1202777696.7564139
737 "subdir": [ "dirnode",
742 "ctime": 1202778102.7589991,
743 "mtime": 1202778111.2160511,
745 "linkcrtime": 1202777696.7564139,
746 "linkmotime": 1202777696.7564139
750 In the above example, note how 'children' is a dictionary in which the keys
751 are child names and the values depend upon whether the child is a file or a
752 directory. The value is mostly the same as the JSON representation of the
753 child object (except that directories do not recurse -- the "children"
754 entry of the child is omitted, and the directory view includes the metadata
755 that is stored on the directory edge).
757 The rw_uri field will be present in the information about a directory
758 if and only if you have read-write access to that directory. The verify_uri
759 field will be present if and only if the object has a verify-cap
760 (non-distributed LIT files do not have verify-caps).
762 If the cap is of an unknown format, then the file size and verify_uri will
765 GET /uri/$UNKNOWNCAP?t=json :
768 "ro_uri": unknown_read_uri
771 GET /uri/$DIRCAP/[SUBDIRS../]UNKNOWNCHILDNAME?t=json :
774 "rw_uri": unknown_write_uri,
775 "ro_uri": unknown_read_uri,
778 "ctime": 1202777696.7564139,
779 "mtime": 1202777696.7564139,
781 "linkcrtime": 1202777696.7564139,
782 "linkmotime": 1202777696.7564139
785 As in the case of file nodes, the metadata will only be present when the
786 capability is to a directory followed by a path. The "mutable" field is also
787 not always present; when it is absent, the mutability of the object is not
793 The value of the 'tahoe':'linkmotime' key is updated whenever a link to a
794 child is set. The value of the 'tahoe':'linkcrtime' key is updated whenever
795 a link to a child is created -- i.e. when there was not previously a link
798 Note however, that if the edge in the Tahoe filesystem points to a mutable
799 file and the contents of that mutable file is changed, then the
800 'tahoe':'linkmotime' value on that edge will *not* be updated, since the
801 edge itself wasn't updated -- only the mutable file was.
803 The timestamps are represented as a number of seconds since the UNIX epoch
804 (1970-01-01 00:00:00 UTC), with leap seconds not being counted in the long
807 In Tahoe earlier than v1.4.0, 'mtime' and 'ctime' keys were populated
808 instead of the 'tahoe':'linkmotime' and 'tahoe':'linkcrtime' keys. Starting
809 in Tahoe v1.4.0, the 'linkmotime'/'linkcrtime' keys in the 'tahoe' sub-dict
810 are populated. However, prior to Tahoe v1.7beta, a bug caused the 'tahoe'
811 sub-dict to be deleted by web-API requests in which new metadata is
812 specified, and not to be added to existing child links that lack it.
814 From Tahoe v1.7.0 onward, the 'mtime' and 'ctime' fields are no longer
815 populated or updated (see ticket #924), except by "tahoe backup" as
816 explained below. For backward compatibility, when an existing link is
817 updated and 'tahoe':'linkcrtime' is not present in the previous metadata
818 but 'ctime' is, the old value of 'ctime' is used as the new value of
819 'tahoe':'linkcrtime'.
821 The reason we added the new fields in Tahoe v1.4.0 is that there is a
822 "set_children" API (described below) which you can use to overwrite the
823 values of the 'mtime'/'ctime' pair, and this API is used by the
824 "tahoe backup" command (in Tahoe v1.3.0 and later) to set the 'mtime' and
825 'ctime' values when backing up files from a local filesystem into the
826 Tahoe filesystem. As of Tahoe v1.4.0, the set_children API cannot be used
827 to set anything under the 'tahoe' key of the metadata dict -- if you
828 include 'tahoe' keys in your 'metadata' arguments then it will silently
831 Therefore, if the 'tahoe' sub-dict is present, you can rely on the
832 'linkcrtime' and 'linkmotime' values therein to have the semantics described
833 above. (This is assuming that only official Tahoe clients have been used to
834 write those links, and that their system clocks were set to what you expected
835 -- there is nothing preventing someone from editing their Tahoe client or
836 writing their own Tahoe client which would overwrite those values however
837 they like, and there is nothing to constrain their system clock from taking
840 When an edge is created or updated by "tahoe backup", the 'mtime' and
841 'ctime' keys on that edge are set as follows:
843 * 'mtime' is set to the timestamp read from the local filesystem for the
844 "mtime" of the local file in question, which means the last time the
845 contents of that file were changed.
847 * On Windows, 'ctime' is set to the creation timestamp for the file
848 read from the local filesystem. On other platforms, 'ctime' is set to
849 the UNIX "ctime" of the local file, which means the last time that
850 either the contents or the metadata of the local file was changed.
852 There are several ways that the 'ctime' field could be confusing:
854 1. You might be confused about whether it reflects the time of the creation
855 of a link in the Tahoe filesystem (by a version of Tahoe < v1.7.0) or a
856 timestamp copied in by "tahoe backup" from a local filesystem.
858 2. You might be confused about whether it is a copy of the file creation
859 time (if "tahoe backup" was run on a Windows system) or of the last
860 contents-or-metadata change (if "tahoe backup" was run on a different
863 3. You might be confused by the fact that changing the contents of a
864 mutable file in Tahoe doesn't have any effect on any links pointing at
865 that file in any directories, although "tahoe backup" sets the link
866 'ctime'/'mtime' to reflect timestamps about the local file corresponding
867 to the Tahoe file to which the link points.
869 4. Also, quite apart from Tahoe, you might be confused about the meaning
870 of the "ctime" in UNIX local filesystems, which people sometimes think
871 means file creation time, but which actually means, in UNIX local
872 filesystems, the most recent time that the file contents or the file
873 metadata (such as owner, permission bits, extended attributes, etc.)
874 has changed. Note that although "ctime" does not mean file creation time
875 in UNIX, links created by a version of Tahoe prior to v1.7.0, and never
876 written by "tahoe backup", will have 'ctime' set to the link creation
880 Attaching an Existing File or Directory by its read- or write-cap
881 -----------------------------------------------------------------
883 ``PUT /uri/$DIRCAP/[SUBDIRS../]CHILDNAME?t=uri``
885 This attaches a child object (either a file or directory) to a specified
886 location in the virtual filesystem. The child object is referenced by its
887 read- or write- cap, as provided in the HTTP request body. This will create
888 intermediate directories as necessary.
890 This is similar to a UNIX hardlink: by referencing a previously-uploaded file
891 (or previously-created directory) instead of uploading/creating a new one,
892 you can create two references to the same object.
894 The read- or write- cap of the child is provided in the body of the HTTP
895 request, and this same cap is returned in the response body.
897 The default behavior is to overwrite any existing object at the same
898 location. To prevent this (and make the operation return an error instead
899 of overwriting), add a "replace=false" argument, as "?t=uri&replace=false".
900 With replace=false, this operation will return an HTTP 409 "Conflict" error
901 if there is already an object at the given location, rather than
902 overwriting the existing object. To allow the operation to overwrite a
903 file, but return an error when trying to overwrite a directory, use
904 "replace=only-files" (this behavior is closer to the traditional UNIX "mv"
905 command). Note that "true", "t", and "1" are all synonyms for "True", and
906 "false", "f", and "0" are synonyms for "False", and the parameter is
909 Note that this operation does not take its child cap in the form of
910 separate "rw_uri" and "ro_uri" fields. Therefore, it cannot accept a
911 child cap in a format unknown to the web-API server, unless its URI
912 starts with "ro." or "imm.". This restriction is necessary because the
913 server is not able to attenuate an unknown write cap to a read cap.
914 Unknown URIs starting with "ro." or "imm.", on the other hand, are
915 assumed to represent read caps. The client should not prefix a write
916 cap with "ro." or "imm." and pass it to this operation, since that
917 would result in granting the cap's write authority to holders of the
921 Adding Multiple Files or Directories to a Parent Directory at Once
922 ------------------------------------------------------------------
924 ``POST /uri/$DIRCAP/[SUBDIRS..]?t=set_children``
926 ``POST /uri/$DIRCAP/[SUBDIRS..]?t=set-children`` (Tahoe >= v1.6)
928 This command adds multiple children to a directory in a single operation.
929 It reads the request body and interprets it as a JSON-encoded description
930 of the child names and read/write-caps that should be added.
932 The body should be a JSON-encoded dictionary, in the same format as the
933 "children" value returned by the "GET /uri/$DIRCAP?t=json" operation
934 described above. In this format, each key is a child names, and the
935 corresponding value is a tuple of (type, childinfo). "type" is ignored, and
936 "childinfo" is a dictionary that contains "rw_uri", "ro_uri", and
937 "metadata" keys. You can take the output of "GET /uri/$DIRCAP1?t=json" and
938 use it as the input to "POST /uri/$DIRCAP2?t=set_children" to make DIR2
939 look very much like DIR1 (except for any existing children of DIR2 that
940 were not overwritten, and any existing "tahoe" metadata keys as described
943 When the set_children request contains a child name that already exists in
944 the target directory, this command defaults to overwriting that child with
945 the new value (both child cap and metadata, but if the JSON data does not
946 contain a "metadata" key, the old child's metadata is preserved). The
947 command takes a boolean "overwrite=" query argument to control this
948 behavior. If you use "?t=set_children&overwrite=false", then an attempt to
949 replace an existing child will instead cause an error.
951 Any "tahoe" key in the new child's "metadata" value is ignored. Any
952 existing "tahoe" metadata is preserved. The metadata["tahoe"] value is
953 reserved for metadata generated by the tahoe node itself. The only two keys
954 currently placed here are "linkcrtime" and "linkmotime". For details, see
955 the section above entitled "Getting Information About a File Or Directory (as
956 JSON)", in the "About the metadata" subsection.
958 Note that this command was introduced with the name "set_children", which
959 uses an underscore rather than a hyphen as other multi-word command names
960 do. The variant with a hyphen is now accepted, but clients that desire
961 backward compatibility should continue to use "set_children".
964 Unlinking a File or Directory
965 -----------------------------
967 ``DELETE /uri/$DIRCAP/[SUBDIRS../]CHILDNAME``
969 This removes the given name from its parent directory. CHILDNAME is the
970 name to be removed, and $DIRCAP/SUBDIRS.. indicates the directory that will
973 Note that this does not actually delete the file or directory that the name
974 points to from the tahoe grid -- it only unlinks the named reference from
975 this directory. If there are other names in this directory or in other
976 directories that point to the resource, then it will remain accessible
977 through those paths. Even if all names pointing to this object are removed
978 from their parent directories, then someone with possession of its read-cap
979 can continue to access the object through that cap.
981 The object will only become completely unreachable once 1: there are no
982 reachable directories that reference it, and 2: nobody is holding a read-
983 or write- cap to the object. (This behavior is very similar to the way
984 hardlinks and anonymous files work in traditional UNIX filesystems).
986 This operation will not modify more than a single directory. Intermediate
987 directories which were implicitly created by PUT or POST methods will *not*
988 be automatically removed by DELETE.
990 This method returns the file- or directory- cap of the object that was just
994 Browser Operations: Human-oriented interfaces
995 =============================================
997 This section describes the HTTP operations that provide support for humans
998 running a web browser. Most of these operations use HTML forms that use POST
999 to drive the Tahoe node. This section is intended for HTML authors who want
1000 to write web pages that contain forms and buttons which manipulate the Tahoe
1003 Note that for all POST operations, the arguments listed can be provided
1004 either as URL query arguments or as form body fields. URL query arguments are
1005 separated from the main URL by "?", and from each other by "&". For example,
1006 "POST /uri/$DIRCAP?t=upload&mutable=true". Form body fields are usually
1007 specified by using <input type="hidden"> elements. For clarity, the
1008 descriptions below display the most significant arguments as URL query args.
1011 Viewing a Directory (as HTML)
1012 -----------------------------
1014 ``GET /uri/$DIRCAP/[SUBDIRS../]``
1016 This returns an HTML page, intended to be displayed to a human by a web
1017 browser, which contains HREF links to all files and directories reachable
1018 from this directory. These HREF links do not have a t= argument, meaning
1019 that a human who follows them will get pages also meant for a human. It also
1020 contains forms to upload new files, and to unlink files and directories
1021 from their parent directory. Those forms use POST methods to do their job.
1024 Viewing/Downloading a File
1025 --------------------------
1027 ``GET /uri/$FILECAP``
1029 ``GET /uri/$DIRCAP/[SUBDIRS../]FILENAME``
1031 This will retrieve the contents of the given file. The HTTP response body
1032 will contain the sequence of bytes that make up the file.
1034 If you want the HTTP response to include a useful Content-Type header,
1035 either use the second form (which starts with a $DIRCAP), or add a
1036 "filename=foo" query argument, like "GET /uri/$FILECAP?filename=foo.jpg".
1037 The bare "GET /uri/$FILECAP" does not give the Tahoe node enough information
1038 to determine a Content-Type (since Tahoe immutable files are merely
1039 sequences of bytes, not typed+named file objects).
1041 If the URL has both filename= and "save=true" in the query arguments, then
1042 the server to add a "Content-Disposition: attachment" header, along with a
1043 filename= parameter. When a user clicks on such a link, most browsers will
1044 offer to let the user save the file instead of displaying it inline (indeed,
1045 most browsers will refuse to display it inline). "true", "t", "1", and other
1046 case-insensitive equivalents are all treated the same.
1048 Character-set handling in URLs and HTTP headers is a dubious art [1]_. For
1049 maximum compatibility, Tahoe simply copies the bytes from the filename=
1050 argument into the Content-Disposition header's filename= parameter, without
1051 trying to interpret them in any particular way.
1054 ``GET /named/$FILECAP/FILENAME``
1056 This is an alternate download form which makes it easier to get the correct
1057 filename. The Tahoe server will provide the contents of the given file, with
1058 a Content-Type header derived from the given filename. This form is used to
1059 get browsers to use the "Save Link As" feature correctly, and also helps
1060 command-line tools like "wget" and "curl" use the right filename. Note that
1061 this form can *only* be used with file caps; it is an error to use a
1062 directory cap after the /named/ prefix.
1064 URLs may also use /file/$FILECAP/FILENAME as a synonym for
1065 /named/$FILECAP/FILENAME.
1067 Getting Information About a File Or Directory (as HTML)
1068 -------------------------------------------------------
1070 ``GET /uri/$FILECAP?t=info``
1072 ``GET /uri/$DIRCAP/?t=info``
1074 ``GET /uri/$DIRCAP/[SUBDIRS../]SUBDIR/?t=info``
1076 ``GET /uri/$DIRCAP/[SUBDIRS../]FILENAME?t=info``
1078 This returns a human-oriented HTML page with more detail about the selected
1079 file or directory object. This page contains the following items:
1083 * JSON representation
1084 * raw contents (text/plain)
1085 * access caps (URIs): verify-cap, read-cap, write-cap (for mutable objects)
1086 * check/verify/repair form
1087 * deep-check/deep-size/deep-stats/manifest (for directories)
1088 * replace-contents form (for mutable files)
1091 Creating a Directory
1092 --------------------
1094 ``POST /uri?t=mkdir``
1096 This creates a new empty directory, but does not attach it to the virtual
1099 If a "redirect_to_result=true" argument is provided, then the HTTP response
1100 will cause the web browser to be redirected to a /uri/$DIRCAP page that
1101 gives access to the newly-created directory. If you bookmark this page,
1102 you'll be able to get back to the directory again in the future. This is the
1103 recommended way to start working with a Tahoe server: create a new unlinked
1104 directory (using redirect_to_result=true), then bookmark the resulting
1105 /uri/$DIRCAP page. There is a "create directory" button on the Welcome page
1106 to invoke this action.
1108 This accepts a format= argument in the query string. Refer to the
1109 documentation of the PUT /uri?t=mkdir operation in `Creating A
1110 New Directory`_ for information on the behavior of the format= argument.
1112 If "redirect_to_result=true" is not provided (or is given a value of
1113 "false"), then the HTTP response body will simply be the write-cap of the
1116 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=mkdir&name=CHILDNAME``
1118 This creates a new empty directory as a child of the designated SUBDIR. This
1119 will create additional intermediate directories as necessary.
1121 This accepts a format= argument in the query string. Refer to the
1122 documentation of POST /uri/$DIRCAP/[SUBDIRS../]?t=mkdir&name=CHILDNAME in
1123 `Creating a New Directory`_ for information on the behavior of the format=
1126 If a "when_done=URL" argument is provided, the HTTP response will cause the
1127 web browser to redirect to the given URL. This provides a convenient way to
1128 return the browser to the directory that was just modified. Without a
1129 when_done= argument, the HTTP response will simply contain the write-cap of
1130 the directory that was just created.
1136 ``POST /uri?t=upload``
1138 This uploads a file, and produces a file-cap for the contents, but does not
1139 attach the file into the filesystem. No directories will be modified by
1142 The file must be provided as the "file" field of an HTML encoded form body,
1143 produced in response to an HTML form like this::
1145 <form action="/uri" method="POST" enctype="multipart/form-data">
1146 <input type="hidden" name="t" value="upload" />
1147 <input type="file" name="file" />
1148 <input type="submit" value="Upload Unlinked" />
1151 If a "when_done=URL" argument is provided, the response body will cause the
1152 browser to redirect to the given URL. If the when_done= URL has the string
1153 "%(uri)s" in it, that string will be replaced by a URL-escaped form of the
1154 newly created file-cap. (Note that without this substitution, there is no
1155 way to access the file that was just uploaded).
1157 The default (in the absence of when_done=) is to return an HTML page that
1158 describes the results of the upload. This page will contain information
1159 about which storage servers were used for the upload, how long each
1160 operation took, etc.
1162 This accepts format= and mutable=true query string arguments. Refer to
1163 `Writing/Uploading a File`_ for information on the behavior of format= and
1166 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=upload``
1168 This uploads a file, and attaches it as a new child of the given directory,
1169 which must be mutable. The file must be provided as the "file" field of an
1170 HTML-encoded form body, produced in response to an HTML form like this::
1172 <form action="." method="POST" enctype="multipart/form-data">
1173 <input type="hidden" name="t" value="upload" />
1174 <input type="file" name="file" />
1175 <input type="submit" value="Upload" />
1178 A "name=" argument can be provided to specify the new child's name,
1179 otherwise it will be taken from the "filename" field of the upload form
1180 (most web browsers will copy the last component of the original file's
1181 pathname into this field). To avoid confusion, name= is not allowed to
1184 If there is already a child with that name, and it is a mutable file, then
1185 its contents are replaced with the data being uploaded. If it is not a
1186 mutable file, the default behavior is to remove the existing child before
1187 creating a new one. To prevent this (and make the operation return an error
1188 instead of overwriting the old child), add a "replace=false" argument, as
1189 "?t=upload&replace=false". With replace=false, this operation will return an
1190 HTTP 409 "Conflict" error if there is already an object at the given
1191 location, rather than overwriting the existing object. Note that "true",
1192 "t", and "1" are all synonyms for "True", and "false", "f", and "0" are
1193 synonyms for "False". the parameter is case-insensitive.
1195 This will create additional intermediate directories as necessary, although
1196 since it is expected to be triggered by a form that was retrieved by "GET
1197 /uri/$DIRCAP/[SUBDIRS../]", it is likely that the parent directory will
1200 This accepts format= and mutable=true query string arguments. Refer to
1201 `Writing/Uploading a File`_ for information on the behavior of format= and
1204 If a "when_done=URL" argument is provided, the HTTP response will cause the
1205 web browser to redirect to the given URL. This provides a convenient way to
1206 return the browser to the directory that was just modified. Without a
1207 when_done= argument, the HTTP response will simply contain the file-cap of
1208 the file that was just uploaded (a write-cap for mutable files, or a
1209 read-cap for immutable files).
1211 ``POST /uri/$DIRCAP/[SUBDIRS../]FILENAME?t=upload``
1213 This also uploads a file and attaches it as a new child of the given
1214 directory, which must be mutable. It is a slight variant of the previous
1215 operation, as the URL refers to the target file rather than the parent
1216 directory. It is otherwise identical: this accepts mutable= and when_done=
1219 ``POST /uri/$FILECAP?t=upload``
1221 This modifies the contents of an existing mutable file in-place. An error is
1222 signalled if $FILECAP does not refer to a mutable file. It behaves just like
1223 the "PUT /uri/$FILECAP" form, but uses a POST for the benefit of HTML forms
1227 Attaching An Existing File Or Directory (by URI)
1228 ------------------------------------------------
1230 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=uri&name=CHILDNAME&uri=CHILDCAP``
1232 This attaches a given read- or write- cap "CHILDCAP" to the designated
1233 directory, with a specified child name. This behaves much like the PUT t=uri
1234 operation, and is a lot like a UNIX hardlink. It is subject to the same
1235 restrictions as that operation on the use of cap formats unknown to the
1238 This will create additional intermediate directories as necessary, although
1239 since it is expected to be triggered by a form that was retrieved by "GET
1240 /uri/$DIRCAP/[SUBDIRS../]", it is likely that the parent directory will
1243 This accepts the same replace= argument as POST t=upload.
1249 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=delete&name=CHILDNAME``
1251 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=unlink&name=CHILDNAME`` (Tahoe >= v1.9)
1253 This instructs the node to remove a child object (file or subdirectory) from
1254 the given directory, which must be mutable. Note that the entire subtree is
1255 unlinked from the parent. Unlike deleting a subdirectory in a UNIX local
1256 filesystem, the subtree need not be empty; if it isn't, then other references
1257 into the subtree will see that the child subdirectories are not modified by
1258 this operation. Only the link from the given directory to its child is severed.
1260 In Tahoe-LAFS v1.9.0 and later, t=unlink can be used as a synonym for t=delete.
1261 If interoperability with older web-API servers is required, t=delete should
1268 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=rename&from_name=OLD&to_name=NEW``
1270 This instructs the node to rename a child of the given directory, which must
1271 be mutable. This has a similar effect to removing the child, then adding the
1272 same child-cap under the new name, except that it preserves metadata. This
1273 operation cannot move the child to a different directory.
1275 The default behavior is to overwrite any existing link at the destination
1276 (replace=true). To prevent this (and make the operation return an error
1277 instead of overwriting), add a "replace=false" argument. With replace=false,
1278 this operation will return an HTTP 409 "Conflict" error if the destination
1279 is not the same link as the source and there is already a link at the
1280 destination, rather than overwriting the existing link. To allow the
1281 operation to overwrite a link to a file, but return an HTTP 409 error when
1282 trying to overwrite a link to a directory, use "replace=only-files" (this
1283 behavior is closer to the traditional UNIX "mv" command). Note that "true",
1284 "t", and "1" are all synonyms for "True"; "false", "f", and "0" are synonyms
1285 for "False"; and the parameter is case-insensitive.
1288 Relinking ("Moving") a Child
1289 ----------------------------
1291 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=relink&from_name=OLD&to_dir=$NEWDIRCAP/[NEWSUBDIRS../]&to_name=NEW``
1292 ``[&replace=true|false|only-files]`` (Tahoe >= v1.10)
1294 This instructs the node to move a child of the given source directory, into
1295 a different directory and/or to a different name. The command is named
1296 ``relink`` because what it does is add a new link to the child from the new
1297 location, then remove the old link. Nothing is actually "moved": the child
1298 is still reachable through any path from which it was formerly reachable,
1299 and the storage space occupied by its ciphertext is not affected.
1301 The source and destination directories must be writeable. If {{{to_dir}}} is
1302 not present, the child link is renamed within the same directory. If
1303 {{{to_name}}} is not present then it defaults to {{{from_name}}}. If the
1304 destination link (directory and name) is the same as the source link, the
1305 operation has no effect.
1307 Metadata from the source directory entry is preserved. Multiple levels of
1308 descent in the source and destination paths are supported.
1310 This operation will return an HTTP 404 "Not Found" error if
1311 ``$DIRCAP/[SUBDIRS../]``, the child being moved, or the destination
1312 directory does not exist. It will return an HTTP 400 "Bad Request" error
1313 if any entry in the source or destination paths is not a directory.
1315 The default behavior is to overwrite any existing link at the destination
1316 (replace=true). To prevent this (and make the operation return an error
1317 instead of overwriting), add a "replace=false" argument. With replace=false,
1318 this operation will return an HTTP 409 "Conflict" error if the destination
1319 is not the same link as the source and there is already a link at the
1320 destination, rather than overwriting the existing link. To allow the
1321 operation to overwrite a link to a file, but return an HTTP 409 error when
1322 trying to overwrite a link to a directory, use "replace=only-files" (this
1323 behavior is closer to the traditional UNIX "mv" command). Note that "true",
1324 "t", and "1" are all synonyms for "True"; "false", "f", and "0" are synonyms
1325 for "False"; and the parameter is case-insensitive.
1327 When relinking into a different directory, for safety, the child link is
1328 not removed from the old directory until it has been successfully added to
1329 the new directory. This implies that in case of a crash or failure, the
1330 link to the child will not be lost, but it could be linked at both the old
1333 The source link should not be the same as any link (directory and child name)
1334 in the ``to_dir`` path. This restriction is not enforced, but it may be
1335 enforced in a future version. If it were violated then the result would be
1336 to create a cycle in the directory structure that is not necessarily reachable
1337 from the root of the destination path (``$NEWDIRCAP``), which could result in
1338 data loss, as described in ticket `#943`_.
1340 .. _`#943`: https://tahoe-lafs.org/trac/tahoe-lafs/ticket/943
1346 ``GET /uri?uri=$CAP``
1348 This causes a redirect to /uri/$CAP, and retains any additional query
1349 arguments (like filename= or save=). This is for the convenience of web
1350 forms which allow the user to paste in a read- or write- cap (obtained
1351 through some out-of-band channel, like IM or email).
1353 Note that this form merely redirects to the specific file or directory
1354 indicated by the $CAP: unlike the GET /uri/$DIRCAP form, you cannot
1355 traverse to children by appending additional path segments to the URL.
1357 ``GET /uri/$DIRCAP/[SUBDIRS../]?t=rename-form&name=$CHILDNAME``
1359 This provides a useful facility to browser-based user interfaces. It
1360 returns a page containing a form targetting the "POST $DIRCAP t=rename"
1361 functionality described above, with the provided $CHILDNAME present in the
1362 'from_name' field of that form. I.e. this presents a form offering to
1363 rename $CHILDNAME, requesting the new name, and submitting POST rename.
1364 This same URL format can also be used with "move-form" with the expected
1367 ``GET /uri/$DIRCAP/[SUBDIRS../]CHILDNAME?t=uri``
1369 This returns the file- or directory- cap for the specified object.
1371 ``GET /uri/$DIRCAP/[SUBDIRS../]CHILDNAME?t=readonly-uri``
1373 This returns a read-only file- or directory- cap for the specified object.
1374 If the object is an immutable file, this will return the same value as
1378 Debugging and Testing Features
1379 ------------------------------
1381 These URLs are less-likely to be helpful to the casual Tahoe user, and are
1382 mainly intended for developers.
1384 ``POST $URL?t=check``
1386 This triggers the FileChecker to determine the current "health" of the
1387 given file or directory, by counting how many shares are available. The
1388 page that is returned will display the results. This can be used as a "show
1389 me detailed information about this file" page.
1391 If a verify=true argument is provided, the node will perform a more
1392 intensive check, downloading and verifying every single bit of every share.
1394 If an add-lease=true argument is provided, the node will also add (or
1395 renew) a lease to every share it encounters. Each lease will keep the share
1396 alive for a certain period of time (one month by default). Once the last
1397 lease expires or is explicitly cancelled, the storage server is allowed to
1400 If an output=JSON argument is provided, the response will be
1401 machine-readable JSON instead of human-oriented HTML. The data is a
1402 dictionary with the following keys::
1404 storage-index: a base32-encoded string with the objects's storage index,
1405 or an empty string for LIT files
1406 summary: a string, with a one-line summary of the stats of the file
1407 results: a dictionary that describes the state of the file. For LIT files,
1408 this dictionary has only the 'healthy' key, which will always be
1409 True. For distributed files, this dictionary has the following
1411 count-shares-good: the number of good shares that were found
1412 count-shares-needed: 'k', the number of shares required for recovery
1413 count-shares-expected: 'N', the number of total shares generated
1414 count-good-share-hosts: the number of distinct storage servers with
1415 good shares. Note that a high value does not
1416 necessarily imply good share distribution,
1417 because some of these servers may only hold
1419 count-wrong-shares: for mutable files, the number of shares for
1420 versions other than the 'best' one (highest
1421 sequence number, highest roothash). These are
1422 either old, or created by an uncoordinated or
1423 not fully successful write.
1424 count-recoverable-versions: for mutable files, the number of
1425 recoverable versions of the file. For
1426 a healthy file, this will equal 1.
1427 count-unrecoverable-versions: for mutable files, the number of
1428 unrecoverable versions of the file.
1429 For a healthy file, this will be 0.
1430 count-corrupt-shares: the number of shares with integrity failures
1431 list-corrupt-shares: a list of "share locators", one for each share
1432 that was found to be corrupt. Each share locator
1433 is a list of (serverid, storage_index, sharenum).
1434 needs-rebalancing: (bool) True if there are multiple shares on a single
1435 storage server, indicating a reduction in reliability
1436 that could be resolved by moving shares to new
1438 servers-responding: list of base32-encoded storage server identifiers,
1439 one for each server which responded to the share
1441 healthy: (bool) True if the file is completely healthy, False otherwise.
1442 Healthy files have at least N good shares. Overlapping shares
1443 do not currently cause a file to be marked unhealthy. If there
1444 are at least N good shares, then corrupt shares do not cause the
1445 file to be marked unhealthy, although the corrupt shares will be
1446 listed in the results (list-corrupt-shares) and should be manually
1447 removed to wasting time in subsequent downloads (as the
1448 downloader rediscovers the corruption and uses alternate shares).
1449 Future compatibility: the meaning of this field may change to
1450 reflect whether the servers-of-happiness criterion is met
1452 sharemap: dict mapping share identifier to list of serverids
1453 (base32-encoded strings). This indicates which servers are
1454 holding which shares. For immutable files, the shareid is
1455 an integer (the share number, from 0 to N-1). For
1456 immutable files, it is a string of the form
1457 'seq%d-%s-sh%d', containing the sequence number, the
1458 roothash, and the share number.
1460 ``POST $URL?t=start-deep-check`` (must add &ophandle=XYZ)
1462 This initiates a recursive walk of all files and directories reachable from
1463 the target, performing a check on each one just like t=check. The result
1464 page will contain a summary of the results, including details on any
1465 file/directory that was not fully healthy.
1467 t=start-deep-check can only be invoked on a directory. An error (400
1468 BAD_REQUEST) will be signalled if it is invoked on a file. The recursive
1469 walker will deal with loops safely.
1471 This accepts the same verify= and add-lease= arguments as t=check.
1473 Since this operation can take a long time (perhaps a second per object),
1474 the ophandle= argument is required (see "Slow Operations, Progress, and
1475 Cancelling" above). The response to this POST will be a redirect to the
1476 corresponding /operations/$HANDLE page (with output=HTML or output=JSON to
1477 match the output= argument given to the POST). The deep-check operation
1478 will continue to run in the background, and the /operations page should be
1479 used to find out when the operation is done.
1481 Detailed check results for non-healthy files and directories will be
1482 available under /operations/$HANDLE/$STORAGEINDEX, and the HTML status will
1483 contain links to these detailed results.
1485 The HTML /operations/$HANDLE page for incomplete operations will contain a
1486 meta-refresh tag, set to 60 seconds, so that a browser which uses
1487 deep-check will automatically poll until the operation has completed.
1489 The JSON page (/options/$HANDLE?output=JSON) will contain a
1490 machine-readable JSON dictionary with the following keys::
1492 finished: a boolean, True if the operation is complete, else False. Some
1493 of the remaining keys may not be present until the operation
1495 root-storage-index: a base32-encoded string with the storage index of the
1496 starting point of the deep-check operation
1497 count-objects-checked: count of how many objects were checked. Note that
1498 non-distributed objects (i.e. small immutable LIT
1499 files) are not checked, since for these objects,
1500 the data is contained entirely in the URI.
1501 count-objects-healthy: how many of those objects were completely healthy
1502 count-objects-unhealthy: how many were damaged in some way
1503 count-corrupt-shares: how many shares were found to have corruption,
1504 summed over all objects examined
1505 list-corrupt-shares: a list of "share identifiers", one for each share
1506 that was found to be corrupt. Each share identifier
1507 is a list of (serverid, storage_index, sharenum).
1508 list-unhealthy-files: a list of (pathname, check-results) tuples, for
1509 each file that was not fully healthy. 'pathname' is
1510 a list of strings (which can be joined by "/"
1511 characters to turn it into a single string),
1512 relative to the directory on which deep-check was
1513 invoked. The 'check-results' field is the same as
1514 that returned by t=check&output=JSON, described
1516 stats: a dictionary with the same keys as the t=start-deep-stats command
1519 ``POST $URL?t=stream-deep-check``
1521 This initiates a recursive walk of all files and directories reachable from
1522 the target, performing a check on each one just like t=check. For each
1523 unique object (duplicates are skipped), a single line of JSON is emitted to
1524 the HTTP response channel (or an error indication, see below). When the walk
1525 is complete, a final line of JSON is emitted which contains the accumulated
1526 file-size/count "deep-stats" data.
1528 This command takes the same arguments as t=start-deep-check.
1530 A CLI tool can split the response stream on newlines into "response units",
1531 and parse each response unit as JSON. Each such parsed unit will be a
1532 dictionary, and will contain at least the "type" key: a string, one of
1533 "file", "directory", or "stats".
1535 For all units that have a type of "file" or "directory", the dictionary will
1536 contain the following keys::
1538 "path": a list of strings, with the path that is traversed to reach the
1540 "cap": a write-cap URI for the file or directory, if available, else a
1542 "verifycap": a verify-cap URI for the file or directory
1543 "repaircap": an URI for the weakest cap that can still be used to repair
1545 "storage-index": a base32 storage index for the object
1546 "check-results": a copy of the dictionary which would be returned by
1547 t=check&output=json, with three top-level keys:
1548 "storage-index", "summary", and "results", and a variety
1549 of counts and sharemaps in the "results" value.
1551 Note that non-distributed files (i.e. LIT files) will have values of None
1552 for verifycap, repaircap, and storage-index, since these files can neither
1553 be verified nor repaired, and are not stored on the storage servers.
1554 Likewise the check-results dictionary will be limited: an empty string for
1555 storage-index, and a results dictionary with only the "healthy" key.
1557 The last unit in the stream will have a type of "stats", and will contain
1558 the keys described in the "start-deep-stats" operation, below.
1560 If any errors occur during the traversal (specifically if a directory is
1561 unrecoverable, such that further traversal is not possible), an error
1562 indication is written to the response body, instead of the usual line of
1563 JSON. This error indication line will begin with the string "ERROR:" (in all
1564 caps), and contain a summary of the error on the rest of the line. The
1565 remaining lines of the response body will be a python exception. The client
1566 application should look for the ERROR: and stop processing JSON as soon as
1567 it is seen. Note that neither a file being unrecoverable nor a directory
1568 merely being unhealthy will cause traversal to stop. The line just before
1569 the ERROR: will describe the directory that was untraversable, since the
1570 unit is emitted to the HTTP response body before the child is traversed.
1573 ``POST $URL?t=check&repair=true``
1575 This performs a health check of the given file or directory, and if the
1576 checker determines that the object is not healthy (some shares are missing
1577 or corrupted), it will perform a "repair". During repair, any missing
1578 shares will be regenerated and uploaded to new servers.
1580 This accepts the same verify=true and add-lease= arguments as t=check. When
1581 an output=JSON argument is provided, the machine-readable JSON response
1582 will contain the following keys::
1584 storage-index: a base32-encoded string with the objects's storage index,
1585 or an empty string for LIT files
1586 repair-attempted: (bool) True if repair was attempted
1587 repair-successful: (bool) True if repair was attempted and the file was
1588 fully healthy afterwards. False if no repair was
1589 attempted, or if a repair attempt failed.
1590 pre-repair-results: a dictionary that describes the state of the file
1591 before any repair was performed. This contains exactly
1592 the same keys as the 'results' value of the t=check
1593 response, described above.
1594 post-repair-results: a dictionary that describes the state of the file
1595 after any repair was performed. If no repair was
1596 performed, post-repair-results and pre-repair-results
1597 will be the same. This contains exactly the same keys
1598 as the 'results' value of the t=check response,
1601 ``POST $URL?t=start-deep-check&repair=true`` (must add &ophandle=XYZ)
1603 This triggers a recursive walk of all files and directories, performing a
1604 t=check&repair=true on each one.
1606 Like t=start-deep-check without the repair= argument, this can only be
1607 invoked on a directory. An error (400 BAD_REQUEST) will be signalled if it
1608 is invoked on a file. The recursive walker will deal with loops safely.
1610 This accepts the same verify= and add-lease= arguments as
1611 t=start-deep-check. It uses the same ophandle= mechanism as
1612 start-deep-check. When an output=JSON argument is provided, the response
1613 will contain the following keys::
1615 finished: (bool) True if the operation has completed, else False
1616 root-storage-index: a base32-encoded string with the storage index of the
1617 starting point of the deep-check operation
1618 count-objects-checked: count of how many objects were checked
1620 count-objects-healthy-pre-repair: how many of those objects were completely
1621 healthy, before any repair
1622 count-objects-unhealthy-pre-repair: how many were damaged in some way
1623 count-objects-healthy-post-repair: how many of those objects were completely
1624 healthy, after any repair
1625 count-objects-unhealthy-post-repair: how many were damaged in some way
1627 count-repairs-attempted: repairs were attempted on this many objects.
1628 count-repairs-successful: how many repairs resulted in healthy objects
1629 count-repairs-unsuccessful: how many repairs resulted did not results in
1630 completely healthy objects
1631 count-corrupt-shares-pre-repair: how many shares were found to have
1632 corruption, summed over all objects
1633 examined, before any repair
1634 count-corrupt-shares-post-repair: how many shares were found to have
1635 corruption, summed over all objects
1636 examined, after any repair
1637 list-corrupt-shares: a list of "share identifiers", one for each share
1638 that was found to be corrupt (before any repair).
1639 Each share identifier is a list of (serverid,
1640 storage_index, sharenum).
1641 list-remaining-corrupt-shares: like list-corrupt-shares, but mutable shares
1642 that were successfully repaired are not
1643 included. These are shares that need
1644 manual processing. Since immutable shares
1645 cannot be modified by clients, all corruption
1646 in immutable shares will be listed here.
1647 list-unhealthy-files: a list of (pathname, check-results) tuples, for
1648 each file that was not fully healthy. 'pathname' is
1649 relative to the directory on which deep-check was
1650 invoked. The 'check-results' field is the same as
1651 that returned by t=check&repair=true&output=JSON,
1653 stats: a dictionary with the same keys as the t=start-deep-stats command
1656 ``POST $URL?t=stream-deep-check&repair=true``
1658 This triggers a recursive walk of all files and directories, performing a
1659 t=check&repair=true on each one. For each unique object (duplicates are
1660 skipped), a single line of JSON is emitted to the HTTP response channel (or
1661 an error indication). When the walk is complete, a final line of JSON is
1662 emitted which contains the accumulated file-size/count "deep-stats" data.
1664 This emits the same data as t=stream-deep-check (without the repair=true),
1665 except that the "check-results" field is replaced with a
1666 "check-and-repair-results" field, which contains the keys returned by
1667 t=check&repair=true&output=json (i.e. repair-attempted, repair-successful,
1668 pre-repair-results, and post-repair-results). The output does not contain
1669 the summary dictionary that is provied by t=start-deep-check&repair=true
1670 (the one with count-objects-checked and list-unhealthy-files), since the
1671 receiving client is expected to calculate those values itself from the
1672 stream of per-object check-and-repair-results.
1674 Note that the "ERROR:" indication will only be emitted if traversal stops,
1675 which will only occur if an unrecoverable directory is encountered. If a
1676 file or directory repair fails, the traversal will continue, and the repair
1677 failure will be indicated in the JSON data (in the "repair-successful" key).
1679 ``POST $DIRURL?t=start-manifest`` (must add &ophandle=XYZ)
1681 This operation generates a "manfest" of the given directory tree, mostly
1682 for debugging. This is a table of (path, filecap/dircap), for every object
1683 reachable from the starting directory. The path will be slash-joined, and
1684 the filecap/dircap will contain a link to the object in question. This page
1685 gives immediate access to every object in the virtual filesystem subtree.
1687 This operation uses the same ophandle= mechanism as deep-check. The
1688 corresponding /operations/$HANDLE page has three different forms. The
1689 default is output=HTML.
1691 If output=text is added to the query args, the results will be a text/plain
1692 list. The first line is special: it is either "finished: yes" or "finished:
1693 no"; if the operation is not finished, you must periodically reload the
1694 page until it completes. The rest of the results are a plaintext list, with
1695 one file/dir per line, slash-separated, with the filecap/dircap separated
1698 If output=JSON is added to the queryargs, then the results will be a
1699 JSON-formatted dictionary with six keys. Note that because large directory
1700 structures can result in very large JSON results, the full results will not
1701 be available until the operation is complete (i.e. until output["finished"]
1704 finished (bool): if False then you must reload the page until True
1705 origin_si (base32 str): the storage index of the starting point
1706 manifest: list of (path, cap) tuples, where path is a list of strings.
1707 verifycaps: list of (printable) verify cap strings
1708 storage-index: list of (base32) storage index strings
1709 stats: a dictionary with the same keys as the t=start-deep-stats command
1712 ``POST $DIRURL?t=start-deep-size`` (must add &ophandle=XYZ)
1714 This operation generates a number (in bytes) containing the sum of the
1715 filesize of all directories and immutable files reachable from the given
1716 directory. This is a rough lower bound of the total space consumed by this
1717 subtree. It does not include space consumed by mutable files, nor does it
1718 take expansion or encoding overhead into account. Later versions of the
1719 code may improve this estimate upwards.
1721 The /operations/$HANDLE status output consists of two lines of text::
1726 ``POST $DIRURL?t=start-deep-stats`` (must add &ophandle=XYZ)
1728 This operation performs a recursive walk of all files and directories
1729 reachable from the given directory, and generates a collection of
1730 statistics about those objects.
1732 The result (obtained from the /operations/$OPHANDLE page) is a
1733 JSON-serialized dictionary with the following keys (note that some of these
1734 keys may be missing until 'finished' is True)::
1736 finished: (bool) True if the operation has finished, else False
1737 count-immutable-files: count of how many CHK files are in the set
1738 count-mutable-files: same, for mutable files (does not include directories)
1739 count-literal-files: same, for LIT files (data contained inside the URI)
1740 count-files: sum of the above three
1741 count-directories: count of directories
1742 count-unknown: count of unrecognized objects (perhaps from the future)
1743 size-immutable-files: total bytes for all CHK files in the set, =deep-size
1744 size-mutable-files (TODO): same, for current version of all mutable files
1745 size-literal-files: same, for LIT files
1746 size-directories: size of directories (includes size-literal-files)
1747 size-files-histogram: list of (minsize, maxsize, count) buckets,
1748 with a histogram of filesizes, 5dB/bucket,
1749 for both literal and immutable files
1750 largest-directory: number of children in the largest directory
1751 largest-immutable-file: number of bytes in the largest CHK file
1753 size-mutable-files is not implemented, because it would require extra
1754 queries to each mutable file to get their size. This may be implemented in
1757 Assuming no sharing, the basic space consumed by a single root directory is
1758 the sum of size-immutable-files, size-mutable-files, and size-directories.
1759 The actual disk space used by the shares is larger, because of the
1760 following sources of overhead::
1763 expansion due to erasure coding
1764 share management data (leases)
1765 backend (ext3) minimum block size
1767 ``POST $URL?t=stream-manifest``
1769 This operation performs a recursive walk of all files and directories
1770 reachable from the given starting point. For each such unique object
1771 (duplicates are skipped), a single line of JSON is emitted to the HTTP
1772 response channel (or an error indication, see below). When the walk is
1773 complete, a final line of JSON is emitted which contains the accumulated
1774 file-size/count "deep-stats" data.
1776 A CLI tool can split the response stream on newlines into "response units",
1777 and parse each response unit as JSON. Each such parsed unit will be a
1778 dictionary, and will contain at least the "type" key: a string, one of
1779 "file", "directory", or "stats".
1781 For all units that have a type of "file" or "directory", the dictionary will
1782 contain the following keys::
1784 "path": a list of strings, with the path that is traversed to reach the
1786 "cap": a write-cap URI for the file or directory, if available, else a
1788 "verifycap": a verify-cap URI for the file or directory
1789 "repaircap": an URI for the weakest cap that can still be used to repair
1791 "storage-index": a base32 storage index for the object
1793 Note that non-distributed files (i.e. LIT files) will have values of None
1794 for verifycap, repaircap, and storage-index, since these files can neither
1795 be verified nor repaired, and are not stored on the storage servers.
1797 The last unit in the stream will have a type of "stats", and will contain
1798 the keys described in the "start-deep-stats" operation, below.
1800 If any errors occur during the traversal (specifically if a directory is
1801 unrecoverable, such that further traversal is not possible), an error
1802 indication is written to the response body, instead of the usual line of
1803 JSON. This error indication line will begin with the string "ERROR:" (in all
1804 caps), and contain a summary of the error on the rest of the line. The
1805 remaining lines of the response body will be a python exception. The client
1806 application should look for the ERROR: and stop processing JSON as soon as
1807 it is seen. The line just before the ERROR: will describe the directory that
1808 was untraversable, since the manifest entry is emitted to the HTTP response
1809 body before the child is traversed.
1815 The portion of the web namespace that begins with "/uri" (and "/named") is
1816 dedicated to giving users (both humans and programs) access to the Tahoe
1817 virtual filesystem. The rest of the namespace provides status information
1818 about the state of the Tahoe node.
1820 ``GET /`` (the root page)
1822 This is the "Welcome Page", and contains a few distinct sections::
1824 Node information: library versions, local nodeid, services being provided.
1826 Filesystem Access Forms: create a new directory, view a file/directory by
1827 URI, upload a file (unlinked), download a file by
1830 Grid Status: introducer information, helper information, connected storage
1835 This page lists all active uploads and downloads, and contains a short list
1836 of recent upload/download operations. Each operation has a link to a page
1837 that describes file sizes, servers that were involved, and the time consumed
1838 in each phase of the operation.
1840 A GET of /status/?t=json will contain a machine-readable subset of the same
1841 data. It returns a JSON-encoded dictionary. The only key defined at this
1842 time is "active", with a value that is a list of operation dictionaries, one
1843 for each active operation. Once an operation is completed, it will no longer
1844 appear in data["active"] .
1846 Each op-dict contains a "type" key, one of "upload", "download",
1847 "mapupdate", "publish", or "retrieve" (the first two are for immutable
1848 files, while the latter three are for mutable files and directories).
1850 The "upload" op-dict will contain the following keys::
1852 type (string): "upload"
1853 storage-index-string (string): a base32-encoded storage index
1854 total-size (int): total size of the file
1855 status (string): current status of the operation
1856 progress-hash (float): 1.0 when the file has been hashed
1857 progress-ciphertext (float): 1.0 when the file has been encrypted.
1858 progress-encode-push (float): 1.0 when the file has been encoded and
1859 pushed to the storage servers. For helper
1860 uploads, the ciphertext value climbs to 1.0
1861 first, then encoding starts. For unassisted
1862 uploads, ciphertext and encode-push progress
1863 will climb at the same pace.
1865 The "download" op-dict will contain the following keys::
1867 type (string): "download"
1868 storage-index-string (string): a base32-encoded storage index
1869 total-size (int): total size of the file
1870 status (string): current status of the operation
1871 progress (float): 1.0 when the file has been fully downloaded
1873 Front-ends which want to report progress information are advised to simply
1874 average together all the progress-* indicators. A slightly more accurate
1875 value can be found by ignoring the progress-hash value (since the current
1876 implementation hashes synchronously, so clients will probably never see
1877 progress-hash!=1.0).
1879 ``GET /helper_status/``
1881 If the node is running a helper (i.e. if [helper]enabled is set to True in
1882 tahoe.cfg), then this page will provide a list of all the helper operations
1883 currently in progress. If "?t=json" is added to the URL, it will return a
1884 JSON-formatted list of helper statistics, which can then be used to produce
1885 graphs to indicate how busy the helper is.
1887 ``GET /statistics/``
1889 This page provides "node statistics", which are collected from a variety of
1892 load_monitor: every second, the node schedules a timer for one second in
1893 the future, then measures how late the subsequent callback
1894 is. The "load_average" is this tardiness, measured in
1895 seconds, averaged over the last minute. It is an indication
1896 of a busy node, one which is doing more work than can be
1897 completed in a timely fashion. The "max_load" value is the
1898 highest value that has been seen in the last 60 seconds.
1900 cpu_monitor: every minute, the node uses time.clock() to measure how much
1901 CPU time it has used, and it uses this value to produce
1902 1min/5min/15min moving averages. These values range from 0%
1903 (0.0) to 100% (1.0), and indicate what fraction of the CPU
1904 has been used by the Tahoe node. Not all operating systems
1905 provide meaningful data to time.clock(): they may report 100%
1906 CPU usage at all times.
1908 uploader: this counts how many immutable files (and bytes) have been
1909 uploaded since the node was started
1911 downloader: this counts how many immutable files have been downloaded
1912 since the node was started
1914 publishes: this counts how many mutable files (including directories) have
1915 been modified since the node was started
1917 retrieves: this counts how many mutable files (including directories) have
1918 been read since the node was started
1920 There are other statistics that are tracked by the node. The "raw stats"
1921 section shows a formatted dump of all of them.
1923 By adding "?t=json" to the URL, the node will return a JSON-formatted
1924 dictionary of stats values, which can be used by other tools to produce
1925 graphs of node behavior. The misc/munin/ directory in the source
1926 distribution provides some tools to produce these graphs.
1928 ``GET /`` (introducer status)
1930 For Introducer nodes, the welcome page displays information about both
1931 clients and servers which are connected to the introducer. Servers make
1932 "service announcements", and these are listed in a table. Clients will
1933 subscribe to hear about service announcements, and these subscriptions are
1934 listed in a separate table. Both tables contain information about what
1935 version of Tahoe is being run by the remote node, their advertised and
1936 outbound IP addresses, their nodeid and nickname, and how long they have
1939 By adding "?t=json" to the URL, the node will return a JSON-formatted
1940 dictionary of stats values, which can be used to produce graphs of connected
1941 clients over time. This dictionary has the following keys::
1943 ["subscription_summary"] : a dictionary mapping service name (like
1944 "storage") to an integer with the number of
1945 clients that have subscribed to hear about that
1947 ["announcement_summary"] : a dictionary mapping service name to an integer
1948 with the number of servers which are announcing
1950 ["announcement_distinct_hosts"] : a dictionary mapping service name to an
1951 integer which represents the number of
1952 distinct hosts that are providing that
1953 service. If two servers have announced
1954 FURLs which use the same hostnames (but
1955 different ports and tubids), they are
1956 considered to be on the same host.
1959 Static Files in /public_html
1960 ============================
1962 The web-API server will take any request for a URL that starts with /static
1963 and serve it from a configurable directory which defaults to
1964 $BASEDIR/public_html . This is configured by setting the "[node]web.static"
1965 value in $BASEDIR/tahoe.cfg . If this is left at the default value of
1966 "public_html", then http://127.0.0.1:3456/static/subdir/foo.html will be
1967 served with the contents of the file $BASEDIR/public_html/subdir/foo.html .
1969 This can be useful to serve a javascript application which provides a
1970 prettier front-end to the rest of the Tahoe web-API.
1973 Safety and Security Issues -- Names vs. URIs
1974 ============================================
1976 Summary: use explicit file- and dir- caps whenever possible, to reduce the
1977 potential for surprises when the filesystem structure is changed.
1979 Tahoe provides a mutable filesystem, but the ways that the filesystem can
1980 change are limited. The only thing that can change is that the mapping from
1981 child names to child objects that each directory contains can be changed by
1982 adding a new child name pointing to an object, removing an existing child name,
1983 or changing an existing child name to point to a different object.
1985 Obviously if you query Tahoe for information about the filesystem and then act
1986 to change the filesystem (such as by getting a listing of the contents of a
1987 directory and then adding a file to the directory), then the filesystem might
1988 have been changed after you queried it and before you acted upon it. However,
1989 if you use the URI instead of the pathname of an object when you act upon the
1990 object, then the only change that can happen is if the object is a directory
1991 then the set of child names it has might be different. If, on the other hand,
1992 you act upon the object using its pathname, then a different object might be in
1993 that place, which can result in more kinds of surprises.
1995 For example, suppose you are writing code which recursively downloads the
1996 contents of a directory. The first thing your code does is fetch the listing
1997 of the contents of the directory. For each child that it fetched, if that
1998 child is a file then it downloads the file, and if that child is a directory
1999 then it recurses into that directory. Now, if the download and the recurse
2000 actions are performed using the child's name, then the results might be
2001 wrong, because for example a child name that pointed to a sub-directory when
2002 you listed the directory might have been changed to point to a file (in which
2003 case your attempt to recurse into it would result in an error and the file
2004 would be skipped), or a child name that pointed to a file when you listed the
2005 directory might now point to a sub-directory (in which case your attempt to
2006 download the child would result in a file containing HTML text describing the
2009 If your recursive algorithm uses the uri of the child instead of the name of
2010 the child, then those kinds of mistakes just can't happen. Note that both the
2011 child's name and the child's URI are included in the results of listing the
2012 parent directory, so it isn't any harder to use the URI for this purpose.
2014 The read and write caps in a given directory node are separate URIs, and
2015 can't be assumed to point to the same object even if they were retrieved in
2016 the same operation (although the web-API server attempts to ensure this
2017 in most cases). If you need to rely on that property, you should explicitly
2018 verify it. More generally, you should not make assumptions about the
2019 internal consistency of the contents of mutable directories. As a result
2020 of the signatures on mutable object versions, it is guaranteed that a given
2021 version was written in a single update, but -- as in the case of a file --
2022 the contents may have been chosen by a malicious writer in a way that is
2023 designed to confuse applications that rely on their consistency.
2025 In general, use names if you want "whatever object (whether file or
2026 directory) is found by following this name (or sequence of names) when my
2027 request reaches the server". Use URIs if you want "this particular object".
2033 Tahoe uses both mutable and immutable files. Mutable files can be created
2034 explicitly by doing an upload with ?mutable=true added, or implicitly by
2035 creating a new directory (since a directory is just a special way to
2036 interpret a given mutable file).
2038 Mutable files suffer from the same consistency-vs-availability tradeoff that
2039 all distributed data storage systems face. It is not possible to
2040 simultaneously achieve perfect consistency and perfect availability in the
2041 face of network partitions (servers being unreachable or faulty).
2043 Tahoe tries to achieve a reasonable compromise, but there is a basic rule in
2044 place, known as the Prime Coordination Directive: "Don't Do That". What this
2045 means is that if write-access to a mutable file is available to several
2046 parties, then those parties are responsible for coordinating their activities
2047 to avoid multiple simultaneous updates. This could be achieved by having
2048 these parties talk to each other and using some sort of locking mechanism, or
2049 by serializing all changes through a single writer.
2051 The consequences of performing uncoordinated writes can vary. Some of the
2052 writers may lose their changes, as somebody else wins the race condition. In
2053 many cases the file will be left in an "unhealthy" state, meaning that there
2054 are not as many redundant shares as we would like (reducing the reliability
2055 of the file against server failures). In the worst case, the file can be left
2056 in such an unhealthy state that no version is recoverable, even the old ones.
2057 It is this small possibility of data loss that prompts us to issue the Prime
2058 Coordination Directive.
2060 Tahoe nodes implement internal serialization to make sure that a single Tahoe
2061 node cannot conflict with itself. For example, it is safe to issue two
2062 directory modification requests to a single tahoe node's web-API server at the
2063 same time, because the Tahoe node will internally delay one of them until
2064 after the other has finished being applied. (This feature was introduced in
2065 Tahoe-1.1; back with Tahoe-1.0 the web client was responsible for serializing
2066 web requests themselves).
2068 For more details, please see the "Consistency vs Availability" and "The Prime
2069 Coordination Directive" sections of `mutable.rst <../specifications/mutable.rst>`_.
2075 Gateway nodes may find it necessary to prohibit access to certain files. The
2076 web-API has a facility to block access to filecaps by their storage index,
2077 returning a 403 "Forbidden" error instead of the original file.
2079 This blacklist is recorded in $NODEDIR/access.blacklist, and contains one
2080 blocked file per line. Comment lines (starting with ``#``) are ignored. Each
2081 line consists of the storage-index (in the usual base32 format as displayed
2082 by the "More Info" page, or by the "tahoe debug dump-cap" command), followed
2083 by whitespace, followed by a reason string, which will be included in the 403
2084 error message. This could hold a URL to a page that explains why the file is
2085 blocked, for example.
2087 So for example, if you found a need to block access to a file with filecap
2088 ``URI:CHK:n7r3m6wmomelk4sep3kw5cvduq:os7ijw5c3maek7pg65e5254k2fzjflavtpejjyhshpsxuqzhcwwq:3:20:14861``,
2089 you could do the following::
2091 tahoe debug dump-cap URI:CHK:n7r3m6wmomelk4sep3kw5cvduq:os7ijw5c3maek7pg65e5254k2fzjflavtpejjyhshpsxuqzhcwwq:3:20:14861
2092 -> storage index: whpepioyrnff7orecjolvbudeu
2093 echo "whpepioyrnff7orecjolvbudeu my puppy told me to" >>$NODEDIR/access.blacklist
2094 tahoe restart $NODEDIR
2095 tahoe get URI:CHK:n7r3m6wmomelk4sep3kw5cvduq:os7ijw5c3maek7pg65e5254k2fzjflavtpejjyhshpsxuqzhcwwq:3:20:14861
2096 -> error, 403 Access Prohibited: my puppy told me to
2098 The ``access.blacklist`` file will be checked each time a file or directory
2099 is accessed: the file's ``mtime`` is used to decide whether it need to be
2100 reloaded. Therefore no node restart is necessary when creating the initial
2101 blacklist, nor when adding second, third, or additional entries to the list.
2102 When modifying the file, be careful to update it atomically, otherwise a
2103 request may arrive while the file is only halfway written, and the partial
2104 file may be incorrectly parsed.
2106 The blacklist is applied to all access paths (including SFTP, FTP, and CLI
2107 operations), not just the web-API. The blacklist also applies to directories.
2108 If a directory is blacklisted, the gateway will refuse access to both that
2109 directory and any child files/directories underneath it, when accessed via
2110 "DIRCAP/SUBDIR/FILENAME" -style URLs. Users who go directly to the child
2111 file/dir will bypass the blacklist.
2113 The node will log the SI of the file being blocked, and the reason code, into
2114 the ``logs/twistd.log`` file.
2117 .. [1] URLs and HTTP and UTF-8, Oh My
2119 HTTP does not provide a mechanism to specify the character set used to
2120 encode non-ASCII names in URLs
2121 (`RFC3986#2.1 <https://tools.ietf.org/html/rfc3986#section-2.1>`_).
2122 We prefer the convention that the ``filename=`` argument shall be a
2123 URL-escaped UTF-8 encoded Unicode string.
2124 For example, suppose we want to provoke the server into using a filename of
2125 "f i a n c e-acute e" (i.e. f i a n c U+00E9 e). The UTF-8 encoding of this
2126 is 0x66 0x69 0x61 0x6e 0x63 0xc3 0xa9 0x65 (or "fianc\\xC3\\xA9e", as python's
2127 ``repr()`` function would show). To encode this into a URL, the non-printable
2128 characters must be escaped with the urlencode ``%XX`` mechanism, giving
2129 us "fianc%C3%A9e". Thus, the first line of the HTTP request will be
2130 "``GET /uri/CAP...?save=true&filename=fianc%C3%A9e HTTP/1.1``". Not all
2131 browsers provide this: IE7 by default uses the Latin-1 encoding, which is
2132 "fianc%E9e" (although it has a configuration option to send URLs as UTF-8).
2134 The response header will need to indicate a non-ASCII filename. The actual
2135 mechanism to do this is not clear. For ASCII filenames, the response header
2138 Content-Disposition: attachment; filename="english.txt"
2140 If Tahoe were to enforce the UTF-8 convention, it would need to decode the
2141 URL argument into a Unicode string, and then encode it back into a sequence
2142 of bytes when creating the response header. One possibility would be to use
2143 unencoded UTF-8. Developers suggest that IE7 might accept this::
2145 #1: Content-Disposition: attachment; filename="fianc\xC3\xA9e"
2146 (note, the last four bytes of that line, not including the newline, are
2147 0xC3 0xA9 0x65 0x22)
2149 `RFC2231#4 <https://tools.ietf.org/html/rfc2231#section-4>`_
2150 (dated 1997): suggests that the following might work, and
2151 `some developers have reported <http://markmail.org/message/dsjyokgl7hv64ig3>`_
2152 that it is supported by Firefox (but not IE7)::
2154 #2: Content-Disposition: attachment; filename*=utf-8''fianc%C3%A9e
2156 My reading of `RFC2616#19.5.1 <https://tools.ietf.org/html/rfc2616#section-19.5.1>`_
2157 (which defines Content-Disposition) says that the filename= parameter is
2158 defined to be wrapped in quotes (presumably to allow spaces without breaking
2159 the parsing of subsequent parameters), which would give us::
2161 #3: Content-Disposition: attachment; filename*=utf-8''"fianc%C3%A9e"
2163 However this is contrary to the examples in the email thread listed above.
2165 Developers report that IE7 (when it is configured for UTF-8 URL encoding,
2166 which is not the default in Asian countries), will accept::
2168 #4: Content-Disposition: attachment; filename=fianc%C3%A9e
2170 However, for maximum compatibility, Tahoe simply copies bytes from the URL
2171 into the response header, rather than enforcing the UTF-8 convention. This
2172 means it does not try to decode the filename from the URL argument, nor does
2173 it encode the filename into the response header.