1 ==========================
2 The Tahoe REST-ful Web API
3 ==========================
5 1. `Enabling the web-API port`_
6 2. `Basic Concepts: GET, PUT, DELETE, POST`_
11 4. `Slow Operations, Progress, and Cancelling`_
12 5. `Programmatic Operations`_
15 2. `Writing/Uploading a File`_
16 3. `Creating a New Directory`_
17 4. `Getting Information About A File Or Directory (as JSON)`_
18 5. `Attaching an Existing File or Directory by its read- or write-cap`_
19 6. `Adding Multiple Files or Directories to a Parent Directory at Once`_
20 7. `Unlinking a File or Directory`_
22 6. `Browser Operations: Human-Oriented Interfaces`_
24 1. `Viewing A Directory (as HTML)`_
25 2. `Viewing/Downloading a File`_
26 3. `Getting Information About A File Or Directory (as HTML)`_
27 4. `Creating a Directory`_
28 5. `Uploading a File`_
29 6. `Attaching An Existing File Or Directory (by URI)`_
30 7. `Unlinking A Child`_
31 8. `Renaming A Child`_
33 10. `Debugging and Testing Features`_
35 7. `Other Useful Pages`_
36 8. `Static Files in /public_html`_
37 9. `Safety and Security Issues -- Names vs. URIs`_
38 10. `Concurrency Issues`_
39 11. `Access Blacklist`_
42 Enabling the web-API port
43 =========================
45 Every Tahoe node is capable of running a built-in HTTP server. To enable
46 this, just write a port number into the "[node]web.port" line of your node's
47 tahoe.cfg file. For example, writing "web.port = 3456" into the "[node]"
48 section of $NODEDIR/tahoe.cfg will cause the node to run a webserver on port
51 This string is actually a Twisted "strports" specification, meaning you can
52 get more control over the interface to which the server binds by supplying
53 additional arguments. For more details, see the documentation on
54 `twisted.application.strports
55 <http://twistedmatrix.com/documents/current/api/twisted.application.strports.html>`_.
57 Writing "tcp:3456:interface=127.0.0.1" into the web.port line does the same
58 but binds to the loopback interface, ensuring that only the programs on the
59 local host can connect. Using "ssl:3456:privateKey=mykey.pem:certKey=cert.pem"
62 This webport can be set when the node is created by passing a --webport
63 option to the 'tahoe create-node' command. By default, the node listens on
64 port 3456, on the loopback (127.0.0.1) interface.
67 Basic Concepts: GET, PUT, DELETE, POST
68 ======================================
70 As described in `docs/architecture.rst <../architecture.rst>`_, each file
71 and directory in a Tahoe virtual filesystem is referenced by an identifier
72 that combines the designation of the object with the authority to do something
73 with it (such as read or modify the contents). This identifier is called a
74 "read-cap" or "write-cap", depending upon whether it enables read-only or
75 read-write access. These "caps" are also referred to as URIs (which may be
76 confusing because they are not currently `RFC3986
77 <http://tools.ietf.org/html/rfc3986>`_-compliant URIs).
79 The Tahoe web-based API is "REST-ful", meaning it implements the concepts of
80 "REpresentational State Transfer": the original scheme by which the World
81 Wide Web was intended to work. Each object (file or directory) is referenced
82 by a URL that includes the read- or write- cap. HTTP methods (GET, PUT, and
83 DELETE) are used to manipulate these objects. You can think of the URL as a
84 noun, and the method as a verb.
86 In REST, the GET method is used to retrieve information about an object, or
87 to retrieve some representation of the object itself. When the object is a
88 file, the basic GET method will simply return the contents of that file.
89 Other variations (generally implemented by adding query parameters to the
90 URL) will return information about the object, such as metadata. GET
91 operations are required to have no side-effects.
93 PUT is used to upload new objects into the filesystem, or to replace an
94 existing link or the contents of a mutable file. DELETE is used to unlink
95 objects from directories. Both PUT and DELETE are required to be idempotent:
96 performing the same operation multiple times must have the same side-effects
97 as only performing it once.
99 POST is used for more complicated actions that cannot be expressed as a GET,
100 PUT, or DELETE. POST operations can be thought of as a method call: sending
101 some message to the object referenced by the URL. In Tahoe, POST is also used
102 for operations that must be triggered by an HTML form (including upload and
103 unlinking), because otherwise a regular web browser has no way to accomplish
104 these tasks. In general, everything that can be done with a PUT or DELETE can
105 also be done with a POST.
107 Tahoe's web API is designed for two different kinds of consumer. The first is
108 a program that needs to manipulate the virtual file system. Such programs are
109 expected to use the RESTful interface described above. The second is a human
110 using a standard web browser to work with the filesystem. This user is given
111 a series of HTML pages with links to download files, and forms that use POST
112 actions to upload, rename, and unlink files.
114 When an error occurs, the HTTP response code will be set to an appropriate
115 400-series code (like 404 Not Found for an unknown childname, or 400 Bad Request
116 when the parameters to a web-API operation are invalid), and the HTTP response
117 body will usually contain a few lines of explanation as to the cause of the
118 error and possible responses. Unusual exceptions may result in a 500 Internal
119 Server Error as a catch-all, with a default response body containing
120 a Nevow-generated HTML-ized representation of the Python exception stack trace
121 that caused the problem. CLI programs which want to copy the response body to
122 stderr should provide an "Accept: text/plain" header to their requests to get
123 a plain text stack trace instead. If the Accept header contains ``*/*``, or
124 ``text/*``, or text/html (or if there is no Accept header), HTML tracebacks will
131 Tahoe uses a variety of read- and write- caps to identify files and
132 directories. The most common of these is the "immutable file read-cap", which
133 is used for most uploaded files. These read-caps look like the following::
135 URI:CHK:ime6pvkaxuetdfah2p2f35pe54:4btz54xk3tew6nd4y2ojpxj4m6wxjqqlwnztgre6gnjgtucd5r4a:3:10:202
137 The next most common is a "directory write-cap", which provides both read and
138 write access to a directory, and look like this::
140 URI:DIR2:djrdkfawoqihigoett4g6auz6a:jx5mplfpwexnoqff7y5e4zjus4lidm76dcuarpct7cckorh2dpgq
142 There are also "directory read-caps", which start with "URI:DIR2-RO:", and
143 give read-only access to a directory. Finally there are also mutable file
144 read- and write- caps, which start with "URI:SSK", and give access to mutable
147 (Later versions of Tahoe will make these strings shorter, and will remove the
148 unfortunate colons, which must be escaped when these caps are embedded in
151 To refer to any Tahoe object through the web API, you simply need to combine
152 a prefix (which indicates the HTTP server to use) with the cap (which
153 indicates which object inside that server to access). Since the default Tahoe
154 webport is 3456, the most common prefix is one that will use a local node
155 listening on this port::
157 http://127.0.0.1:3456/uri/ + $CAP
159 So, to access the directory named above (which happens to be the
160 publically-writeable sample directory on the Tahoe test grid, described at
161 http://allmydata.org/trac/tahoe/wiki/TestGrid), the URL would be::
163 http://127.0.0.1:3456/uri/URI%3ADIR2%3Adjrdkfawoqihigoett4g6auz6a%3Ajx5mplfpwexnoqff7y5e4zjus4lidm76dcuarpct7cckorh2dpgq/
165 (note that the colons in the directory-cap are url-encoded into "%3A"
168 Likewise, to access the file named above, use::
170 http://127.0.0.1:3456/uri/URI%3ACHK%3Aime6pvkaxuetdfah2p2f35pe54%3A4btz54xk3tew6nd4y2ojpxj4m6wxjqqlwnztgre6gnjgtucd5r4a%3A3%3A10%3A202
172 In the rest of this document, we'll use "$DIRCAP" as shorthand for a read-cap
173 or write-cap that refers to a directory, and "$FILECAP" to abbreviate a cap
174 that refers to a file (whether mutable or immutable). So those URLs above can
177 http://127.0.0.1:3456/uri/$DIRCAP/
178 http://127.0.0.1:3456/uri/$FILECAP
180 The operation summaries below will abbreviate these further, by eliding the
181 server prefix. They will be displayed like this::
190 Tahoe directories contain named child entries, just like directories in a regular
191 local filesystem. These child entries, called "dirnodes", consist of a name,
192 metadata, a write slot, and a read slot. The write and read slots normally contain
193 a write-cap and read-cap referring to the same object, which can be either a file
194 or a subdirectory. The write slot may be empty (actually, both may be empty,
195 but that is unusual).
197 If you have a Tahoe URL that refers to a directory, and want to reference a
198 named child inside it, just append the child name to the URL. For example, if
199 our sample directory contains a file named "welcome.txt", we can refer to
202 http://127.0.0.1:3456/uri/$DIRCAP/welcome.txt
204 (or http://127.0.0.1:3456/uri/URI%3ADIR2%3Adjrdkfawoqihigoett4g6auz6a%3Ajx5mplfpwexnoqff7y5e4zjus4lidm76dcuarpct7cckorh2dpgq/welcome.txt)
206 Multiple levels of subdirectories can be handled this way::
208 http://127.0.0.1:3456/uri/$DIRCAP/tahoe-source/docs/architecture.rst
210 In this document, when we need to refer to a URL that references a file using
211 this child-of-some-directory format, we'll use the following string::
213 /uri/$DIRCAP/[SUBDIRS../]FILENAME
215 The "[SUBDIRS../]" part means that there are zero or more (optional)
216 subdirectory names in the middle of the URL. The "FILENAME" at the end means
217 that this whole URL refers to a file of some sort, rather than to a
220 When we need to refer specifically to a directory in this way, we'll write::
222 /uri/$DIRCAP/[SUBDIRS../]SUBDIR
225 Note that all components of pathnames in URLs are required to be UTF-8
226 encoded, so "resume.doc" (with an acute accent on both E's) would be accessed
229 http://127.0.0.1:3456/uri/$DIRCAP/r%C3%A9sum%C3%A9.doc
231 Also note that the filenames inside upload POST forms are interpreted using
232 whatever character set was provided in the conventional '_charset' field, and
233 defaults to UTF-8 if not otherwise specified. The JSON representation of each
234 directory contains native Unicode strings. Tahoe directories are specified to
235 contain Unicode filenames, and cannot contain binary strings that are not
236 representable as such.
238 All Tahoe operations that refer to existing files or directories must include
239 a suitable read- or write- cap in the URL: the web-API server won't add one
240 for you. If you don't know the cap, you can't access the file. This allows
241 the security properties of Tahoe caps to be extended across the web-API
245 Slow Operations, Progress, and Cancelling
246 =========================================
248 Certain operations can be expected to take a long time. The "t=deep-check",
249 described below, will recursively visit every file and directory reachable
250 from a given starting point, which can take minutes or even hours for
251 extremely large directory structures. A single long-running HTTP request is a
252 fragile thing: proxies, NAT boxes, browsers, and users may all grow impatient
253 with waiting and give up on the connection.
255 For this reason, long-running operations have an "operation handle", which
256 can be used to poll for status/progress messages while the operation
257 proceeds. This handle can also be used to cancel the operation. These handles
258 are created by the client, and passed in as a an "ophandle=" query argument
259 to the POST or PUT request which starts the operation. The following
260 operations can then be used to retrieve status:
262 ``GET /operations/$HANDLE?output=HTML (with or without t=status)``
264 ``GET /operations/$HANDLE?output=JSON (same)``
266 These two retrieve the current status of the given operation. Each operation
267 presents a different sort of information, but in general the page retrieved
270 * whether the operation is complete, or if it is still running
271 * how much of the operation is complete, and how much is left, if possible
273 Note that the final status output can be quite large: a deep-manifest of a
274 directory structure with 300k directories and 200k unique files is about
275 275MB of JSON, and might take two minutes to generate. For this reason, the
276 full status is not provided until the operation has completed.
278 The HTML form will include a meta-refresh tag, which will cause a regular
279 web browser to reload the status page about 60 seconds later. This tag will
280 be removed once the operation has completed.
282 There may be more status information available under
283 /operations/$HANDLE/$ETC : i.e., the handle forms the root of a URL space.
285 ``POST /operations/$HANDLE?t=cancel``
287 This terminates the operation, and returns an HTML page explaining what was
288 cancelled. If the operation handle has already expired (see below), this
289 POST will return a 404, which indicates that the operation is no longer
290 running (either it was completed or terminated). The response body will be
291 the same as a GET /operations/$HANDLE on this operation handle, and the
292 handle will be expired immediately afterwards.
294 The operation handle will eventually expire, to avoid consuming an unbounded
295 amount of memory. The handle's time-to-live can be reset at any time, by
296 passing a retain-for= argument (with a count of seconds) to either the
297 initial POST that starts the operation, or the subsequent GET request which
298 asks about the operation. For example, if a 'GET
299 /operations/$HANDLE?output=JSON&retain-for=600' query is performed, the
300 handle will remain active for 600 seconds (10 minutes) after the GET was
303 In addition, if the GET includes a release-after-complete=True argument, and
304 the operation has completed, the operation handle will be released
307 If a retain-for= argument is not used, the default handle lifetimes are:
309 * handles will remain valid at least until their operation finishes
310 * uncollected handles for finished operations (i.e. handles for
311 operations that have finished but for which the GET page has not been
312 accessed since completion) will remain valid for four days, or for
313 the total time consumed by the operation, whichever is greater.
314 * collected handles (i.e. the GET page has been retrieved at least once
315 since the operation completed) will remain valid for one day.
317 Many "slow" operations can begin to use unacceptable amounts of memory when
318 operating on large directory structures. The memory usage increases when the
319 ophandle is polled, as the results must be copied into a JSON string, sent
320 over the wire, then parsed by a client. So, as an alternative, many "slow"
321 operations have streaming equivalents. These equivalents do not use operation
322 handles. Instead, they emit line-oriented status results immediately. Client
323 code can cancel the operation by simply closing the HTTP connection.
326 Programmatic Operations
327 =======================
329 Now that we know how to build URLs that refer to files and directories in a
330 Tahoe virtual filesystem, what sorts of operations can we do with those URLs?
331 This section contains a catalog of GET, PUT, DELETE, and POST operations that
332 can be performed on these URLs. This set of operations are aimed at programs
333 that use HTTP to communicate with a Tahoe node. A later section describes
334 operations that are intended for web browsers.
340 ``GET /uri/$FILECAP``
342 ``GET /uri/$DIRCAP/[SUBDIRS../]FILENAME``
344 This will retrieve the contents of the given file. The HTTP response body
345 will contain the sequence of bytes that make up the file.
347 To view files in a web browser, you may want more control over the
348 Content-Type and Content-Disposition headers. Please see the next section
349 "Browser Operations", for details on how to modify these URLs for that
353 Writing/Uploading A File
354 ------------------------
356 ``PUT /uri/$FILECAP``
358 ``PUT /uri/$DIRCAP/[SUBDIRS../]FILENAME``
360 Upload a file, using the data from the HTTP request body, and add whatever
361 child links and subdirectories are necessary to make the file available at
362 the given location. Once this operation succeeds, a GET on the same URL will
363 retrieve the same contents that were just uploaded. This will create any
364 necessary intermediate subdirectories.
366 To use the /uri/$FILECAP form, $FILECAP must be a write-cap for a mutable file.
368 In the /uri/$DIRCAP/[SUBDIRS../]FILENAME form, if the target file is a
369 writeable mutable file, that file's contents will be overwritten
370 in-place. If it is a read-cap for a mutable file, an error will occur.
371 If it is an immutable file, the old file will be discarded, and a new
372 one will be put in its place. If the target file is a writable mutable
373 file, you may also specify an "offset" parameter -- a byte offset that
374 determines where in the mutable file the data from the HTTP request
375 body is placed. This operation is relatively efficient for MDMF mutable
376 files, and is relatively inefficient (but still supported) for SDMF
377 mutable files. If no offset parameter is specified, then the entire
378 file is replaced with the data from the HTTP request body. For an
379 immutable file, the "offset" parameter is not valid.
381 When creating a new file, if "mutable=true" is in the query arguments, the
382 operation will create a mutable file instead of an immutable one.
384 This returns the file-cap of the resulting file. If a new file was created
385 by this method, the HTTP response code (as dictated by rfc2616) will be set
386 to 201 CREATED. If an existing file was replaced or modified, the response
389 Note that the 'curl -T localfile http://127.0.0.1:3456/uri/$DIRCAP/foo.txt'
390 command can be used to invoke this operation.
394 This uploads a file, and produces a file-cap for the contents, but does not
395 attach the file into the filesystem. No directories will be modified by
396 this operation. The file-cap is returned as the body of the HTTP response.
398 If "mutable=true" is in the query arguments, the operation will create a
399 mutable file, and return its write-cap in the HTTP respose. The default is
400 to create an immutable file, returning the read-cap as a response. If
401 you create a mutable file, you can also use the "mutable-type" query
402 parameter. If "mutable-type=sdmf", then the mutable file will be created
403 in the old SDMF mutable file format. This is desirable for files that
404 need to be read by old clients. If "mutable-type=mdmf", then the file
405 will be created in the new MDMF mutable file format. MDMF mutable files
406 can be downloaded more efficiently, and modified in-place efficiently,
407 but are not compatible with older versions of Tahoe-LAFS. If no
408 "mutable-type" argument is given, the file is created in whatever
409 format was configured in tahoe.cfg.
412 Creating A New Directory
413 ------------------------
415 ``POST /uri?t=mkdir``
419 Create a new empty directory and return its write-cap as the HTTP response
420 body. This does not make the newly created directory visible from the
421 filesystem. The "PUT" operation is provided for backwards compatibility:
422 new code should use POST.
424 ``POST /uri?t=mkdir-with-children``
426 Create a new directory, populated with a set of child nodes, and return its
427 write-cap as the HTTP response body. The new directory is not attached to
428 any other directory: the returned write-cap is the only reference to it.
430 Initial children are provided as the body of the POST form (this is more
431 efficient than doing separate mkdir and set_children operations). If the
432 body is empty, the new directory will be empty. If not empty, the body will
433 be interpreted as a UTF-8 JSON-encoded dictionary of children with which the
434 new directory should be populated, using the same format as would be
435 returned in the 'children' value of the t=json GET request, described below.
436 Each dictionary key should be a child name, and each value should be a list
437 of [TYPE, PROPDICT], where PROPDICT contains "rw_uri", "ro_uri", and
438 "metadata" keys (all others are ignored). For example, the PUT request body
442 "Fran\u00e7ais": [ "filenode", {
443 "ro_uri": "URI:CHK:...",
446 "ctime": 1202777696.7564139,
447 "mtime": 1202777696.7564139,
449 "linkcrtime": 1202777696.7564139,
450 "linkmotime": 1202777696.7564139
452 "subdir": [ "dirnode", {
453 "rw_uri": "URI:DIR2:...",
454 "ro_uri": "URI:DIR2-RO:...",
456 "ctime": 1202778102.7589991,
457 "mtime": 1202778111.2160511,
459 "linkcrtime": 1202777696.7564139,
460 "linkmotime": 1202777696.7564139
464 For forward-compatibility, a mutable directory can also contain caps in
465 a format that is unknown to the web-API server. When such caps are retrieved
466 from a mutable directory in a "ro_uri" field, they will be prefixed with
467 the string "ro.", indicating that they must not be decoded without
468 checking that they are read-only. The "ro." prefix must not be stripped
469 off without performing this check. (Future versions of the web-API server
470 will perform it where necessary.)
472 If both the "rw_uri" and "ro_uri" fields are present in a given PROPDICT,
473 and the web-API server recognizes the rw_uri as a write cap, then it will
474 reset the ro_uri to the corresponding read cap and discard the original
475 contents of ro_uri (in order to ensure that the two caps correspond to the
476 same object and that the ro_uri is in fact read-only). However this may not
477 happen for caps in a format unknown to the web-API server. Therefore, when
478 writing a directory the web-API client should ensure that the contents
479 of "rw_uri" and "ro_uri" for a given PROPDICT are a consistent
480 (write cap, read cap) pair if possible. If the web-API client only has
481 one cap and does not know whether it is a write cap or read cap, then
482 it is acceptable to set "rw_uri" to that cap and omit "ro_uri". The
483 client must not put a write cap into a "ro_uri" field.
485 The metadata may have a "no-write" field. If this is set to true in the
486 metadata of a link, it will not be possible to open that link for writing
487 via the SFTP frontend; see `<FTP-and-SFTP.rst>`_ for details.
488 Also, if the "no-write" field is set to true in the metadata of a link to
489 a mutable child, it will cause the link to be diminished to read-only.
491 Note that the web-API-using client application must not provide the
492 "Content-Type: multipart/form-data" header that usually accompanies HTML
493 form submissions, since the body is not formatted this way. Doing so will
494 cause a server error as the lower-level code misparses the request body.
496 Child file names should each be expressed as a Unicode string, then used as
497 keys of the dictionary. The dictionary should then be converted into JSON,
498 and the resulting string encoded into UTF-8. This UTF-8 bytestring should
499 then be used as the POST body.
501 ``POST /uri?t=mkdir-immutable``
503 Like t=mkdir-with-children above, but the new directory will be
504 deep-immutable. This means that the directory itself is immutable, and that
505 it can only contain objects that are treated as being deep-immutable, like
506 immutable files, literal files, and deep-immutable directories.
508 For forward-compatibility, a deep-immutable directory can also contain caps
509 in a format that is unknown to the web-API server. When such caps are retrieved
510 from a deep-immutable directory in a "ro_uri" field, they will be prefixed
511 with the string "imm.", indicating that they must not be decoded without
512 checking that they are immutable. The "imm." prefix must not be stripped
513 off without performing this check. (Future versions of the web-API server
514 will perform it where necessary.)
516 The cap for each child may be given either in the "rw_uri" or "ro_uri"
517 field of the PROPDICT (not both). If a cap is given in the "rw_uri" field,
518 then the web-API server will check that it is an immutable read-cap of a
519 *known* format, and give an error if it is not. If a cap is given in the
520 "ro_uri" field, then the web-API server will still check whether known
521 caps are immutable, but for unknown caps it will simply assume that the
522 cap can be stored, as described above. Note that an attacker would be
523 able to store any cap in an immutable directory, so this check when
524 creating the directory is only to help non-malicious clients to avoid
525 accidentally giving away more authority than intended.
527 A non-empty request body is mandatory, since after the directory is created,
528 it will not be possible to add more children to it.
530 ``POST /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=mkdir``
532 ``PUT /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=mkdir``
534 Create new directories as necessary to make sure that the named target
535 ($DIRCAP/SUBDIRS../SUBDIR) is a directory. This will create additional
536 intermediate mutable directories as necessary. If the named target directory
537 already exists, this will make no changes to it.
539 If the final directory is created, it will be empty.
541 This operation will return an error if a blocking file is present at any of
542 the parent names, preventing the server from creating the necessary parent
543 directory; or if it would require changing an immutable directory.
545 The write-cap of the new directory will be returned as the HTTP response
548 ``POST /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=mkdir-with-children``
550 Like /uri?t=mkdir-with-children, but the final directory is created as a
551 child of an existing mutable directory. This will create additional
552 intermediate mutable directories as necessary. If the final directory is
553 created, it will be populated with initial children from the POST request
554 body, as described above.
556 This operation will return an error if a blocking file is present at any of
557 the parent names, preventing the server from creating the necessary parent
558 directory; or if it would require changing an immutable directory; or if
559 the immediate parent directory already has a a child named SUBDIR.
561 ``POST /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=mkdir-immutable``
563 Like /uri?t=mkdir-immutable, but the final directory is created as a child
564 of an existing mutable directory. The final directory will be deep-immutable,
565 and will be populated with the children specified as a JSON dictionary in
566 the POST request body.
568 In Tahoe 1.6 this operation creates intermediate mutable directories if
569 necessary, but that behaviour should not be relied on; see ticket #920.
571 This operation will return an error if the parent directory is immutable,
572 or already has a child named SUBDIR.
574 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=mkdir&name=NAME``
576 Create a new empty mutable directory and attach it to the given existing
577 directory. This will create additional intermediate directories as necessary.
579 This operation will return an error if a blocking file is present at any of
580 the parent names, preventing the server from creating the necessary parent
581 directory, or if it would require changing any immutable directory.
583 The URL of this operation points to the parent of the bottommost new directory,
584 whereas the /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=mkdir operation above has a URL
585 that points directly to the bottommost new directory.
587 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=mkdir-with-children&name=NAME``
589 Like /uri/$DIRCAP/[SUBDIRS../]?t=mkdir&name=NAME, but the new directory will
590 be populated with initial children via the POST request body. This command
591 will create additional intermediate mutable directories as necessary.
593 This operation will return an error if a blocking file is present at any of
594 the parent names, preventing the server from creating the necessary parent
595 directory; or if it would require changing an immutable directory; or if
596 the immediate parent directory already has a a child named NAME.
598 Note that the name= argument must be passed as a queryarg, because the POST
599 request body is used for the initial children JSON.
601 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=mkdir-immutable&name=NAME``
603 Like /uri/$DIRCAP/[SUBDIRS../]?t=mkdir-with-children&name=NAME, but the
604 final directory will be deep-immutable. The children are specified as a
605 JSON dictionary in the POST request body. Again, the name= argument must be
606 passed as a queryarg.
608 In Tahoe 1.6 this operation creates intermediate mutable directories if
609 necessary, but that behaviour should not be relied on; see ticket #920.
611 This operation will return an error if the parent directory is immutable,
612 or already has a child named NAME.
615 Getting Information About A File Or Directory (as JSON)
616 -------------------------------------------------------
618 ``GET /uri/$FILECAP?t=json``
620 ``GET /uri/$DIRCAP?t=json``
622 ``GET /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=json``
624 ``GET /uri/$DIRCAP/[SUBDIRS../]FILENAME?t=json``
626 This returns a machine-parseable JSON-encoded description of the given
627 object. The JSON always contains a list, and the first element of the list is
628 always a flag that indicates whether the referenced object is a file or a
629 directory. If it is a capability to a file, then the information includes
630 file size and URI, like this::
632 GET /uri/$FILECAP?t=json :
636 "verify_uri": verify_uri,
641 If it is a capability to a directory followed by a path from that directory
642 to a file, then the information also includes metadata from the link to the
643 file in the parent directory, like this::
645 GET /uri/$DIRCAP/[SUBDIRS../]FILENAME?t=json
649 "verify_uri": verify_uri,
653 "ctime": 1202777696.7564139,
654 "mtime": 1202777696.7564139,
656 "linkcrtime": 1202777696.7564139,
657 "linkmotime": 1202777696.7564139
660 If it is a directory, then it includes information about the children of
661 this directory, as a mapping from child name to a set of data about the
662 child (the same data that would appear in a corresponding GET?t=json of the
663 child itself). The child entries also include metadata about each child,
664 including link-creation- and link-change- timestamps. The output looks like
667 GET /uri/$DIRCAP?t=json :
668 GET /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=json :
671 "rw_uri": read_write_uri,
672 "ro_uri": read_only_uri,
673 "verify_uri": verify_uri,
676 "foo.txt": [ "filenode", {
680 "ctime": 1202777696.7564139,
681 "mtime": 1202777696.7564139,
683 "linkcrtime": 1202777696.7564139,
684 "linkmotime": 1202777696.7564139
686 "subdir": [ "dirnode", {
690 "ctime": 1202778102.7589991,
691 "mtime": 1202778111.2160511,
693 "linkcrtime": 1202777696.7564139,
694 "linkmotime": 1202777696.7564139
698 In the above example, note how 'children' is a dictionary in which the keys
699 are child names and the values depend upon whether the child is a file or a
700 directory. The value is mostly the same as the JSON representation of the
701 child object (except that directories do not recurse -- the "children"
702 entry of the child is omitted, and the directory view includes the metadata
703 that is stored on the directory edge).
705 The rw_uri field will be present in the information about a directory
706 if and only if you have read-write access to that directory. The verify_uri
707 field will be present if and only if the object has a verify-cap
708 (non-distributed LIT files do not have verify-caps).
710 If the cap is of an unknown format, then the file size and verify_uri will
713 GET /uri/$UNKNOWNCAP?t=json :
716 "ro_uri": unknown_read_uri
719 GET /uri/$DIRCAP/[SUBDIRS../]UNKNOWNCHILDNAME?t=json :
722 "rw_uri": unknown_write_uri,
723 "ro_uri": unknown_read_uri,
726 "ctime": 1202777696.7564139,
727 "mtime": 1202777696.7564139,
729 "linkcrtime": 1202777696.7564139,
730 "linkmotime": 1202777696.7564139
733 As in the case of file nodes, the metadata will only be present when the
734 capability is to a directory followed by a path. The "mutable" field is also
735 not always present; when it is absent, the mutability of the object is not
741 The value of the 'tahoe':'linkmotime' key is updated whenever a link to a
742 child is set. The value of the 'tahoe':'linkcrtime' key is updated whenever
743 a link to a child is created -- i.e. when there was not previously a link
746 Note however, that if the edge in the Tahoe filesystem points to a mutable
747 file and the contents of that mutable file is changed, then the
748 'tahoe':'linkmotime' value on that edge will *not* be updated, since the
749 edge itself wasn't updated -- only the mutable file was.
751 The timestamps are represented as a number of seconds since the UNIX epoch
752 (1970-01-01 00:00:00 UTC), with leap seconds not being counted in the long
755 In Tahoe earlier than v1.4.0, 'mtime' and 'ctime' keys were populated
756 instead of the 'tahoe':'linkmotime' and 'tahoe':'linkcrtime' keys. Starting
757 in Tahoe v1.4.0, the 'linkmotime'/'linkcrtime' keys in the 'tahoe' sub-dict
758 are populated. However, prior to Tahoe v1.7beta, a bug caused the 'tahoe'
759 sub-dict to be deleted by web-API requests in which new metadata is
760 specified, and not to be added to existing child links that lack it.
762 From Tahoe v1.7.0 onward, the 'mtime' and 'ctime' fields are no longer
763 populated or updated (see ticket #924), except by "tahoe backup" as
764 explained below. For backward compatibility, when an existing link is
765 updated and 'tahoe':'linkcrtime' is not present in the previous metadata
766 but 'ctime' is, the old value of 'ctime' is used as the new value of
767 'tahoe':'linkcrtime'.
769 The reason we added the new fields in Tahoe v1.4.0 is that there is a
770 "set_children" API (described below) which you can use to overwrite the
771 values of the 'mtime'/'ctime' pair, and this API is used by the
772 "tahoe backup" command (in Tahoe v1.3.0 and later) to set the 'mtime' and
773 'ctime' values when backing up files from a local filesystem into the
774 Tahoe filesystem. As of Tahoe v1.4.0, the set_children API cannot be used
775 to set anything under the 'tahoe' key of the metadata dict -- if you
776 include 'tahoe' keys in your 'metadata' arguments then it will silently
779 Therefore, if the 'tahoe' sub-dict is present, you can rely on the
780 'linkcrtime' and 'linkmotime' values therein to have the semantics described
781 above. (This is assuming that only official Tahoe clients have been used to
782 write those links, and that their system clocks were set to what you expected
783 -- there is nothing preventing someone from editing their Tahoe client or
784 writing their own Tahoe client which would overwrite those values however
785 they like, and there is nothing to constrain their system clock from taking
788 When an edge is created or updated by "tahoe backup", the 'mtime' and
789 'ctime' keys on that edge are set as follows:
791 * 'mtime' is set to the timestamp read from the local filesystem for the
792 "mtime" of the local file in question, which means the last time the
793 contents of that file were changed.
795 * On Windows, 'ctime' is set to the creation timestamp for the file
796 read from the local filesystem. On other platforms, 'ctime' is set to
797 the UNIX "ctime" of the local file, which means the last time that
798 either the contents or the metadata of the local file was changed.
800 There are several ways that the 'ctime' field could be confusing:
802 1. You might be confused about whether it reflects the time of the creation
803 of a link in the Tahoe filesystem (by a version of Tahoe < v1.7.0) or a
804 timestamp copied in by "tahoe backup" from a local filesystem.
806 2. You might be confused about whether it is a copy of the file creation
807 time (if "tahoe backup" was run on a Windows system) or of the last
808 contents-or-metadata change (if "tahoe backup" was run on a different
811 3. You might be confused by the fact that changing the contents of a
812 mutable file in Tahoe doesn't have any effect on any links pointing at
813 that file in any directories, although "tahoe backup" sets the link
814 'ctime'/'mtime' to reflect timestamps about the local file corresponding
815 to the Tahoe file to which the link points.
817 4. Also, quite apart from Tahoe, you might be confused about the meaning
818 of the "ctime" in UNIX local filesystems, which people sometimes think
819 means file creation time, but which actually means, in UNIX local
820 filesystems, the most recent time that the file contents or the file
821 metadata (such as owner, permission bits, extended attributes, etc.)
822 has changed. Note that although "ctime" does not mean file creation time
823 in UNIX, links created by a version of Tahoe prior to v1.7.0, and never
824 written by "tahoe backup", will have 'ctime' set to the link creation
828 Attaching an Existing File or Directory by its read- or write-cap
829 -----------------------------------------------------------------
831 ``PUT /uri/$DIRCAP/[SUBDIRS../]CHILDNAME?t=uri``
833 This attaches a child object (either a file or directory) to a specified
834 location in the virtual filesystem. The child object is referenced by its
835 read- or write- cap, as provided in the HTTP request body. This will create
836 intermediate directories as necessary.
838 This is similar to a UNIX hardlink: by referencing a previously-uploaded file
839 (or previously-created directory) instead of uploading/creating a new one,
840 you can create two references to the same object.
842 The read- or write- cap of the child is provided in the body of the HTTP
843 request, and this same cap is returned in the response body.
845 The default behavior is to overwrite any existing object at the same
846 location. To prevent this (and make the operation return an error instead
847 of overwriting), add a "replace=false" argument, as "?t=uri&replace=false".
848 With replace=false, this operation will return an HTTP 409 "Conflict" error
849 if there is already an object at the given location, rather than
850 overwriting the existing object. To allow the operation to overwrite a
851 file, but return an error when trying to overwrite a directory, use
852 "replace=only-files" (this behavior is closer to the traditional UNIX "mv"
853 command). Note that "true", "t", and "1" are all synonyms for "True", and
854 "false", "f", and "0" are synonyms for "False", and the parameter is
857 Note that this operation does not take its child cap in the form of
858 separate "rw_uri" and "ro_uri" fields. Therefore, it cannot accept a
859 child cap in a format unknown to the web-API server, unless its URI
860 starts with "ro." or "imm.". This restriction is necessary because the
861 server is not able to attenuate an unknown write cap to a read cap.
862 Unknown URIs starting with "ro." or "imm.", on the other hand, are
863 assumed to represent read caps. The client should not prefix a write
864 cap with "ro." or "imm." and pass it to this operation, since that
865 would result in granting the cap's write authority to holders of the
869 Adding Multiple Files or Directories to a Parent Directory at Once
870 ------------------------------------------------------------------
872 ``POST /uri/$DIRCAP/[SUBDIRS..]?t=set_children``
874 ``POST /uri/$DIRCAP/[SUBDIRS..]?t=set-children`` (Tahoe >= v1.6)
876 This command adds multiple children to a directory in a single operation.
877 It reads the request body and interprets it as a JSON-encoded description
878 of the child names and read/write-caps that should be added.
880 The body should be a JSON-encoded dictionary, in the same format as the
881 "children" value returned by the "GET /uri/$DIRCAP?t=json" operation
882 described above. In this format, each key is a child names, and the
883 corresponding value is a tuple of (type, childinfo). "type" is ignored, and
884 "childinfo" is a dictionary that contains "rw_uri", "ro_uri", and
885 "metadata" keys. You can take the output of "GET /uri/$DIRCAP1?t=json" and
886 use it as the input to "POST /uri/$DIRCAP2?t=set_children" to make DIR2
887 look very much like DIR1 (except for any existing children of DIR2 that
888 were not overwritten, and any existing "tahoe" metadata keys as described
891 When the set_children request contains a child name that already exists in
892 the target directory, this command defaults to overwriting that child with
893 the new value (both child cap and metadata, but if the JSON data does not
894 contain a "metadata" key, the old child's metadata is preserved). The
895 command takes a boolean "overwrite=" query argument to control this
896 behavior. If you use "?t=set_children&overwrite=false", then an attempt to
897 replace an existing child will instead cause an error.
899 Any "tahoe" key in the new child's "metadata" value is ignored. Any
900 existing "tahoe" metadata is preserved. The metadata["tahoe"] value is
901 reserved for metadata generated by the tahoe node itself. The only two keys
902 currently placed here are "linkcrtime" and "linkmotime". For details, see
903 the section above entitled "Get Information About A File Or Directory (as
904 JSON)", in the "About the metadata" subsection.
906 Note that this command was introduced with the name "set_children", which
907 uses an underscore rather than a hyphen as other multi-word command names
908 do. The variant with a hyphen is now accepted, but clients that desire
909 backward compatibility should continue to use "set_children".
912 Unlinking a File or Directory
913 -----------------------------
915 ``DELETE /uri/$DIRCAP/[SUBDIRS../]CHILDNAME``
917 This removes the given name from its parent directory. CHILDNAME is the
918 name to be removed, and $DIRCAP/SUBDIRS.. indicates the directory that will
921 Note that this does not actually delete the file or directory that the name
922 points to from the tahoe grid -- it only unlinks the named reference from
923 this directory. If there are other names in this directory or in other
924 directories that point to the resource, then it will remain accessible
925 through those paths. Even if all names pointing to this object are removed
926 from their parent directories, then someone with possession of its read-cap
927 can continue to access the object through that cap.
929 The object will only become completely unreachable once 1: there are no
930 reachable directories that reference it, and 2: nobody is holding a read-
931 or write- cap to the object. (This behavior is very similar to the way
932 hardlinks and anonymous files work in traditional UNIX filesystems).
934 This operation will not modify more than a single directory. Intermediate
935 directories which were implicitly created by PUT or POST methods will *not*
936 be automatically removed by DELETE.
938 This method returns the file- or directory- cap of the object that was just
942 Browser Operations: Human-oriented interfaces
943 =============================================
945 This section describes the HTTP operations that provide support for humans
946 running a web browser. Most of these operations use HTML forms that use POST
947 to drive the Tahoe node. This section is intended for HTML authors who want
948 to write web pages that contain forms and buttons which manipulate the Tahoe
951 Note that for all POST operations, the arguments listed can be provided
952 either as URL query arguments or as form body fields. URL query arguments are
953 separated from the main URL by "?", and from each other by "&". For example,
954 "POST /uri/$DIRCAP?t=upload&mutable=true". Form body fields are usually
955 specified by using <input type="hidden"> elements. For clarity, the
956 descriptions below display the most significant arguments as URL query args.
959 Viewing A Directory (as HTML)
960 -----------------------------
962 ``GET /uri/$DIRCAP/[SUBDIRS../]``
964 This returns an HTML page, intended to be displayed to a human by a web
965 browser, which contains HREF links to all files and directories reachable
966 from this directory. These HREF links do not have a t= argument, meaning
967 that a human who follows them will get pages also meant for a human. It also
968 contains forms to upload new files, and to unlink files and directories
969 from their parent directory. Those forms use POST methods to do their job.
972 Viewing/Downloading a File
973 --------------------------
975 ``GET /uri/$FILECAP``
977 ``GET /uri/$DIRCAP/[SUBDIRS../]FILENAME``
979 This will retrieve the contents of the given file. The HTTP response body
980 will contain the sequence of bytes that make up the file.
982 If you want the HTTP response to include a useful Content-Type header,
983 either use the second form (which starts with a $DIRCAP), or add a
984 "filename=foo" query argument, like "GET /uri/$FILECAP?filename=foo.jpg".
985 The bare "GET /uri/$FILECAP" does not give the Tahoe node enough information
986 to determine a Content-Type (since Tahoe immutable files are merely
987 sequences of bytes, not typed+named file objects).
989 If the URL has both filename= and "save=true" in the query arguments, then
990 the server to add a "Content-Disposition: attachment" header, along with a
991 filename= parameter. When a user clicks on such a link, most browsers will
992 offer to let the user save the file instead of displaying it inline (indeed,
993 most browsers will refuse to display it inline). "true", "t", "1", and other
994 case-insensitive equivalents are all treated the same.
996 Character-set handling in URLs and HTTP headers is a dubious art [1]_. For
997 maximum compatibility, Tahoe simply copies the bytes from the filename=
998 argument into the Content-Disposition header's filename= parameter, without
999 trying to interpret them in any particular way.
1002 ``GET /named/$FILECAP/FILENAME``
1004 This is an alternate download form which makes it easier to get the correct
1005 filename. The Tahoe server will provide the contents of the given file, with
1006 a Content-Type header derived from the given filename. This form is used to
1007 get browsers to use the "Save Link As" feature correctly, and also helps
1008 command-line tools like "wget" and "curl" use the right filename. Note that
1009 this form can *only* be used with file caps; it is an error to use a
1010 directory cap after the /named/ prefix.
1013 Getting Information About A File Or Directory (as HTML)
1014 -------------------------------------------------------
1016 ``GET /uri/$FILECAP?t=info``
1018 ``GET /uri/$DIRCAP/?t=info``
1020 ``GET /uri/$DIRCAP/[SUBDIRS../]SUBDIR/?t=info``
1022 ``GET /uri/$DIRCAP/[SUBDIRS../]FILENAME?t=info``
1024 This returns a human-oriented HTML page with more detail about the selected
1025 file or directory object. This page contains the following items:
1029 * JSON representation
1030 * raw contents (text/plain)
1031 * access caps (URIs): verify-cap, read-cap, write-cap (for mutable objects)
1032 * check/verify/repair form
1033 * deep-check/deep-size/deep-stats/manifest (for directories)
1034 * replace-conents form (for mutable files)
1037 Creating a Directory
1038 --------------------
1040 ``POST /uri?t=mkdir``
1042 This creates a new empty directory, but does not attach it to the virtual
1045 If a "redirect_to_result=true" argument is provided, then the HTTP response
1046 will cause the web browser to be redirected to a /uri/$DIRCAP page that
1047 gives access to the newly-created directory. If you bookmark this page,
1048 you'll be able to get back to the directory again in the future. This is the
1049 recommended way to start working with a Tahoe server: create a new unlinked
1050 directory (using redirect_to_result=true), then bookmark the resulting
1051 /uri/$DIRCAP page. There is a "create directory" button on the Welcome page
1052 to invoke this action.
1054 If "redirect_to_result=true" is not provided (or is given a value of
1055 "false"), then the HTTP response body will simply be the write-cap of the
1058 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=mkdir&name=CHILDNAME``
1060 This creates a new empty directory as a child of the designated SUBDIR. This
1061 will create additional intermediate directories as necessary.
1063 If a "when_done=URL" argument is provided, the HTTP response will cause the
1064 web browser to redirect to the given URL. This provides a convenient way to
1065 return the browser to the directory that was just modified. Without a
1066 when_done= argument, the HTTP response will simply contain the write-cap of
1067 the directory that was just created.
1073 ``POST /uri?t=upload``
1075 This uploads a file, and produces a file-cap for the contents, but does not
1076 attach the file into the filesystem. No directories will be modified by
1079 The file must be provided as the "file" field of an HTML encoded form body,
1080 produced in response to an HTML form like this::
1082 <form action="/uri" method="POST" enctype="multipart/form-data">
1083 <input type="hidden" name="t" value="upload" />
1084 <input type="file" name="file" />
1085 <input type="submit" value="Upload Unlinked" />
1088 If a "when_done=URL" argument is provided, the response body will cause the
1089 browser to redirect to the given URL. If the when_done= URL has the string
1090 "%(uri)s" in it, that string will be replaced by a URL-escaped form of the
1091 newly created file-cap. (Note that without this substitution, there is no
1092 way to access the file that was just uploaded).
1094 The default (in the absence of when_done=) is to return an HTML page that
1095 describes the results of the upload. This page will contain information
1096 about which storage servers were used for the upload, how long each
1097 operation took, etc.
1099 If a "mutable=true" argument is provided, the operation will create a
1100 mutable file, and the response body will contain the write-cap instead of
1101 the upload results page. The default is to create an immutable file,
1102 returning the upload results page as a response. If you create a
1103 mutable file, you may choose to specify the format of that mutable file
1104 with the "mutable-type" parameter. If "mutable-type=mdmf", then the
1105 file will be created as an MDMF mutable file. If "mutable-type=sdmf",
1106 then the file will be created as an SDMF mutable file. If no value is
1107 specified, the file will be created in whatever format is specified in
1111 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=upload``
1113 This uploads a file, and attaches it as a new child of the given directory,
1114 which must be mutable. The file must be provided as the "file" field of an
1115 HTML-encoded form body, produced in response to an HTML form like this::
1117 <form action="." method="POST" enctype="multipart/form-data">
1118 <input type="hidden" name="t" value="upload" />
1119 <input type="file" name="file" />
1120 <input type="submit" value="Upload" />
1123 A "name=" argument can be provided to specify the new child's name,
1124 otherwise it will be taken from the "filename" field of the upload form
1125 (most web browsers will copy the last component of the original file's
1126 pathname into this field). To avoid confusion, name= is not allowed to
1129 If there is already a child with that name, and it is a mutable file, then
1130 its contents are replaced with the data being uploaded. If it is not a
1131 mutable file, the default behavior is to remove the existing child before
1132 creating a new one. To prevent this (and make the operation return an error
1133 instead of overwriting the old child), add a "replace=false" argument, as
1134 "?t=upload&replace=false". With replace=false, this operation will return an
1135 HTTP 409 "Conflict" error if there is already an object at the given
1136 location, rather than overwriting the existing object. Note that "true",
1137 "t", and "1" are all synonyms for "True", and "false", "f", and "0" are
1138 synonyms for "False". the parameter is case-insensitive.
1140 This will create additional intermediate directories as necessary, although
1141 since it is expected to be triggered by a form that was retrieved by "GET
1142 /uri/$DIRCAP/[SUBDIRS../]", it is likely that the parent directory will
1145 If a "mutable=true" argument is provided, any new file that is created will
1146 be a mutable file instead of an immutable one. <input type="checkbox"
1147 name="mutable" /> will give the user a way to set this option.
1149 If a "when_done=URL" argument is provided, the HTTP response will cause the
1150 web browser to redirect to the given URL. This provides a convenient way to
1151 return the browser to the directory that was just modified. Without a
1152 when_done= argument, the HTTP response will simply contain the file-cap of
1153 the file that was just uploaded (a write-cap for mutable files, or a
1154 read-cap for immutable files).
1156 ``POST /uri/$DIRCAP/[SUBDIRS../]FILENAME?t=upload``
1158 This also uploads a file and attaches it as a new child of the given
1159 directory, which must be mutable. It is a slight variant of the previous
1160 operation, as the URL refers to the target file rather than the parent
1161 directory. It is otherwise identical: this accepts mutable= and when_done=
1164 ``POST /uri/$FILECAP?t=upload``
1166 This modifies the contents of an existing mutable file in-place. An error is
1167 signalled if $FILECAP does not refer to a mutable file. It behaves just like
1168 the "PUT /uri/$FILECAP" form, but uses a POST for the benefit of HTML forms
1172 Attaching An Existing File Or Directory (by URI)
1173 ------------------------------------------------
1175 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=uri&name=CHILDNAME&uri=CHILDCAP``
1177 This attaches a given read- or write- cap "CHILDCAP" to the designated
1178 directory, with a specified child name. This behaves much like the PUT t=uri
1179 operation, and is a lot like a UNIX hardlink. It is subject to the same
1180 restrictions as that operation on the use of cap formats unknown to the
1183 This will create additional intermediate directories as necessary, although
1184 since it is expected to be triggered by a form that was retrieved by "GET
1185 /uri/$DIRCAP/[SUBDIRS../]", it is likely that the parent directory will
1188 This accepts the same replace= argument as POST t=upload.
1194 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=delete&name=CHILDNAME``
1196 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=unlink&name=CHILDNAME``
1198 This instructs the node to remove a child object (file or subdirectory) from
1199 the given directory, which must be mutable. Note that the entire subtree is
1200 unlinked from the parent. Unlike deleting a subdirectory in a UNIX local
1201 filesystem, the subtree need not be empty; if it isn't, then other references
1202 into the subtree will see that the child subdirectories are not modified by
1203 this operation. Only the link from the given directory to its child is severed.
1205 In Tahoe-LAFS v1.9.0 and later, t=unlink can be used as a synonym for t=delete.
1206 If interoperability with older web-API servers is required, t=delete should
1213 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=rename&from_name=OLD&to_name=NEW``
1215 This instructs the node to rename a child of the given directory, which must
1216 be mutable. This has a similar effect to removing the child, then adding the
1217 same child-cap under the new name, except that it preserves metadata. This
1218 operation cannot move the child to a different directory.
1220 This operation will replace any existing child of the new name, making it
1221 behave like the UNIX "``mv -f``" command.
1226 ``GET /uri?uri=$CAP``
1228 This causes a redirect to /uri/$CAP, and retains any additional query
1229 arguments (like filename= or save=). This is for the convenience of web
1230 forms which allow the user to paste in a read- or write- cap (obtained
1231 through some out-of-band channel, like IM or email).
1233 Note that this form merely redirects to the specific file or directory
1234 indicated by the $CAP: unlike the GET /uri/$DIRCAP form, you cannot
1235 traverse to children by appending additional path segments to the URL.
1237 ``GET /uri/$DIRCAP/[SUBDIRS../]?t=rename-form&name=$CHILDNAME``
1239 This provides a useful facility to browser-based user interfaces. It
1240 returns a page containing a form targetting the "POST $DIRCAP t=rename"
1241 functionality described above, with the provided $CHILDNAME present in the
1242 'from_name' field of that form. I.e. this presents a form offering to
1243 rename $CHILDNAME, requesting the new name, and submitting POST rename.
1245 ``GET /uri/$DIRCAP/[SUBDIRS../]CHILDNAME?t=uri``
1247 This returns the file- or directory- cap for the specified object.
1249 ``GET /uri/$DIRCAP/[SUBDIRS../]CHILDNAME?t=readonly-uri``
1251 This returns a read-only file- or directory- cap for the specified object.
1252 If the object is an immutable file, this will return the same value as
1256 Debugging and Testing Features
1257 ------------------------------
1259 These URLs are less-likely to be helpful to the casual Tahoe user, and are
1260 mainly intended for developers.
1262 ``POST $URL?t=check``
1264 This triggers the FileChecker to determine the current "health" of the
1265 given file or directory, by counting how many shares are available. The
1266 page that is returned will display the results. This can be used as a "show
1267 me detailed information about this file" page.
1269 If a verify=true argument is provided, the node will perform a more
1270 intensive check, downloading and verifying every single bit of every share.
1272 If an add-lease=true argument is provided, the node will also add (or
1273 renew) a lease to every share it encounters. Each lease will keep the share
1274 alive for a certain period of time (one month by default). Once the last
1275 lease expires or is explicitly cancelled, the storage server is allowed to
1278 If an output=JSON argument is provided, the response will be
1279 machine-readable JSON instead of human-oriented HTML. The data is a
1280 dictionary with the following keys::
1282 storage-index: a base32-encoded string with the objects's storage index,
1283 or an empty string for LIT files
1284 summary: a string, with a one-line summary of the stats of the file
1285 results: a dictionary that describes the state of the file. For LIT files,
1286 this dictionary has only the 'healthy' key, which will always be
1287 True. For distributed files, this dictionary has the following
1289 count-shares-good: the number of good shares that were found
1290 count-shares-needed: 'k', the number of shares required for recovery
1291 count-shares-expected: 'N', the number of total shares generated
1292 count-good-share-hosts: this was intended to be the number of distinct
1293 storage servers with good shares. It is currently
1294 (as of Tahoe-LAFS v1.8.0) computed incorrectly;
1296 count-wrong-shares: for mutable files, the number of shares for
1297 versions other than the 'best' one (highest
1298 sequence number, highest roothash). These are
1300 count-recoverable-versions: for mutable files, the number of
1301 recoverable versions of the file. For
1302 a healthy file, this will equal 1.
1303 count-unrecoverable-versions: for mutable files, the number of
1304 unrecoverable versions of the file.
1305 For a healthy file, this will be 0.
1306 count-corrupt-shares: the number of shares with integrity failures
1307 list-corrupt-shares: a list of "share locators", one for each share
1308 that was found to be corrupt. Each share locator
1309 is a list of (serverid, storage_index, sharenum).
1310 needs-rebalancing: (bool) True if there are multiple shares on a single
1311 storage server, indicating a reduction in reliability
1312 that could be resolved by moving shares to new
1314 servers-responding: list of base32-encoded storage server identifiers,
1315 one for each server which responded to the share
1317 healthy: (bool) True if the file is completely healthy, False otherwise.
1318 Healthy files have at least N good shares. Overlapping shares
1319 do not currently cause a file to be marked unhealthy. If there
1320 are at least N good shares, then corrupt shares do not cause the
1321 file to be marked unhealthy, although the corrupt shares will be
1322 listed in the results (list-corrupt-shares) and should be manually
1323 removed to wasting time in subsequent downloads (as the
1324 downloader rediscovers the corruption and uses alternate shares).
1325 Future compatibility: the meaning of this field may change to
1326 reflect whether the servers-of-happiness criterion is met
1328 sharemap: dict mapping share identifier to list of serverids
1329 (base32-encoded strings). This indicates which servers are
1330 holding which shares. For immutable files, the shareid is
1331 an integer (the share number, from 0 to N-1). For
1332 immutable files, it is a string of the form
1333 'seq%d-%s-sh%d', containing the sequence number, the
1334 roothash, and the share number.
1336 ``POST $URL?t=start-deep-check`` (must add &ophandle=XYZ)
1338 This initiates a recursive walk of all files and directories reachable from
1339 the target, performing a check on each one just like t=check. The result
1340 page will contain a summary of the results, including details on any
1341 file/directory that was not fully healthy.
1343 t=start-deep-check can only be invoked on a directory. An error (400
1344 BAD_REQUEST) will be signalled if it is invoked on a file. The recursive
1345 walker will deal with loops safely.
1347 This accepts the same verify= and add-lease= arguments as t=check.
1349 Since this operation can take a long time (perhaps a second per object),
1350 the ophandle= argument is required (see "Slow Operations, Progress, and
1351 Cancelling" above). The response to this POST will be a redirect to the
1352 corresponding /operations/$HANDLE page (with output=HTML or output=JSON to
1353 match the output= argument given to the POST). The deep-check operation
1354 will continue to run in the background, and the /operations page should be
1355 used to find out when the operation is done.
1357 Detailed check results for non-healthy files and directories will be
1358 available under /operations/$HANDLE/$STORAGEINDEX, and the HTML status will
1359 contain links to these detailed results.
1361 The HTML /operations/$HANDLE page for incomplete operations will contain a
1362 meta-refresh tag, set to 60 seconds, so that a browser which uses
1363 deep-check will automatically poll until the operation has completed.
1365 The JSON page (/options/$HANDLE?output=JSON) will contain a
1366 machine-readable JSON dictionary with the following keys::
1368 finished: a boolean, True if the operation is complete, else False. Some
1369 of the remaining keys may not be present until the operation
1371 root-storage-index: a base32-encoded string with the storage index of the
1372 starting point of the deep-check operation
1373 count-objects-checked: count of how many objects were checked. Note that
1374 non-distributed objects (i.e. small immutable LIT
1375 files) are not checked, since for these objects,
1376 the data is contained entirely in the URI.
1377 count-objects-healthy: how many of those objects were completely healthy
1378 count-objects-unhealthy: how many were damaged in some way
1379 count-corrupt-shares: how many shares were found to have corruption,
1380 summed over all objects examined
1381 list-corrupt-shares: a list of "share identifiers", one for each share
1382 that was found to be corrupt. Each share identifier
1383 is a list of (serverid, storage_index, sharenum).
1384 list-unhealthy-files: a list of (pathname, check-results) tuples, for
1385 each file that was not fully healthy. 'pathname' is
1386 a list of strings (which can be joined by "/"
1387 characters to turn it into a single string),
1388 relative to the directory on which deep-check was
1389 invoked. The 'check-results' field is the same as
1390 that returned by t=check&output=JSON, described
1392 stats: a dictionary with the same keys as the t=start-deep-stats command
1395 ``POST $URL?t=stream-deep-check``
1397 This initiates a recursive walk of all files and directories reachable from
1398 the target, performing a check on each one just like t=check. For each
1399 unique object (duplicates are skipped), a single line of JSON is emitted to
1400 the HTTP response channel (or an error indication, see below). When the walk
1401 is complete, a final line of JSON is emitted which contains the accumulated
1402 file-size/count "deep-stats" data.
1404 This command takes the same arguments as t=start-deep-check.
1406 A CLI tool can split the response stream on newlines into "response units",
1407 and parse each response unit as JSON. Each such parsed unit will be a
1408 dictionary, and will contain at least the "type" key: a string, one of
1409 "file", "directory", or "stats".
1411 For all units that have a type of "file" or "directory", the dictionary will
1412 contain the following keys::
1414 "path": a list of strings, with the path that is traversed to reach the
1416 "cap": a write-cap URI for the file or directory, if available, else a
1418 "verifycap": a verify-cap URI for the file or directory
1419 "repaircap": an URI for the weakest cap that can still be used to repair
1421 "storage-index": a base32 storage index for the object
1422 "check-results": a copy of the dictionary which would be returned by
1423 t=check&output=json, with three top-level keys:
1424 "storage-index", "summary", and "results", and a variety
1425 of counts and sharemaps in the "results" value.
1427 Note that non-distributed files (i.e. LIT files) will have values of None
1428 for verifycap, repaircap, and storage-index, since these files can neither
1429 be verified nor repaired, and are not stored on the storage servers.
1430 Likewise the check-results dictionary will be limited: an empty string for
1431 storage-index, and a results dictionary with only the "healthy" key.
1433 The last unit in the stream will have a type of "stats", and will contain
1434 the keys described in the "start-deep-stats" operation, below.
1436 If any errors occur during the traversal (specifically if a directory is
1437 unrecoverable, such that further traversal is not possible), an error
1438 indication is written to the response body, instead of the usual line of
1439 JSON. This error indication line will begin with the string "ERROR:" (in all
1440 caps), and contain a summary of the error on the rest of the line. The
1441 remaining lines of the response body will be a python exception. The client
1442 application should look for the ERROR: and stop processing JSON as soon as
1443 it is seen. Note that neither a file being unrecoverable nor a directory
1444 merely being unhealthy will cause traversal to stop. The line just before
1445 the ERROR: will describe the directory that was untraversable, since the
1446 unit is emitted to the HTTP response body before the child is traversed.
1449 ``POST $URL?t=check&repair=true``
1451 This performs a health check of the given file or directory, and if the
1452 checker determines that the object is not healthy (some shares are missing
1453 or corrupted), it will perform a "repair". During repair, any missing
1454 shares will be regenerated and uploaded to new servers.
1456 This accepts the same verify=true and add-lease= arguments as t=check. When
1457 an output=JSON argument is provided, the machine-readable JSON response
1458 will contain the following keys::
1460 storage-index: a base32-encoded string with the objects's storage index,
1461 or an empty string for LIT files
1462 repair-attempted: (bool) True if repair was attempted
1463 repair-successful: (bool) True if repair was attempted and the file was
1464 fully healthy afterwards. False if no repair was
1465 attempted, or if a repair attempt failed.
1466 pre-repair-results: a dictionary that describes the state of the file
1467 before any repair was performed. This contains exactly
1468 the same keys as the 'results' value of the t=check
1469 response, described above.
1470 post-repair-results: a dictionary that describes the state of the file
1471 after any repair was performed. If no repair was
1472 performed, post-repair-results and pre-repair-results
1473 will be the same. This contains exactly the same keys
1474 as the 'results' value of the t=check response,
1477 ``POST $URL?t=start-deep-check&repair=true`` (must add &ophandle=XYZ)
1479 This triggers a recursive walk of all files and directories, performing a
1480 t=check&repair=true on each one.
1482 Like t=start-deep-check without the repair= argument, this can only be
1483 invoked on a directory. An error (400 BAD_REQUEST) will be signalled if it
1484 is invoked on a file. The recursive walker will deal with loops safely.
1486 This accepts the same verify= and add-lease= arguments as
1487 t=start-deep-check. It uses the same ophandle= mechanism as
1488 start-deep-check. When an output=JSON argument is provided, the response
1489 will contain the following keys::
1491 finished: (bool) True if the operation has completed, else False
1492 root-storage-index: a base32-encoded string with the storage index of the
1493 starting point of the deep-check operation
1494 count-objects-checked: count of how many objects were checked
1496 count-objects-healthy-pre-repair: how many of those objects were completely
1497 healthy, before any repair
1498 count-objects-unhealthy-pre-repair: how many were damaged in some way
1499 count-objects-healthy-post-repair: how many of those objects were completely
1500 healthy, after any repair
1501 count-objects-unhealthy-post-repair: how many were damaged in some way
1503 count-repairs-attempted: repairs were attempted on this many objects.
1504 count-repairs-successful: how many repairs resulted in healthy objects
1505 count-repairs-unsuccessful: how many repairs resulted did not results in
1506 completely healthy objects
1507 count-corrupt-shares-pre-repair: how many shares were found to have
1508 corruption, summed over all objects
1509 examined, before any repair
1510 count-corrupt-shares-post-repair: how many shares were found to have
1511 corruption, summed over all objects
1512 examined, after any repair
1513 list-corrupt-shares: a list of "share identifiers", one for each share
1514 that was found to be corrupt (before any repair).
1515 Each share identifier is a list of (serverid,
1516 storage_index, sharenum).
1517 list-remaining-corrupt-shares: like list-corrupt-shares, but mutable shares
1518 that were successfully repaired are not
1519 included. These are shares that need
1520 manual processing. Since immutable shares
1521 cannot be modified by clients, all corruption
1522 in immutable shares will be listed here.
1523 list-unhealthy-files: a list of (pathname, check-results) tuples, for
1524 each file that was not fully healthy. 'pathname' is
1525 relative to the directory on which deep-check was
1526 invoked. The 'check-results' field is the same as
1527 that returned by t=check&repair=true&output=JSON,
1529 stats: a dictionary with the same keys as the t=start-deep-stats command
1532 ``POST $URL?t=stream-deep-check&repair=true``
1534 This triggers a recursive walk of all files and directories, performing a
1535 t=check&repair=true on each one. For each unique object (duplicates are
1536 skipped), a single line of JSON is emitted to the HTTP response channel (or
1537 an error indication). When the walk is complete, a final line of JSON is
1538 emitted which contains the accumulated file-size/count "deep-stats" data.
1540 This emits the same data as t=stream-deep-check (without the repair=true),
1541 except that the "check-results" field is replaced with a
1542 "check-and-repair-results" field, which contains the keys returned by
1543 t=check&repair=true&output=json (i.e. repair-attempted, repair-successful,
1544 pre-repair-results, and post-repair-results). The output does not contain
1545 the summary dictionary that is provied by t=start-deep-check&repair=true
1546 (the one with count-objects-checked and list-unhealthy-files), since the
1547 receiving client is expected to calculate those values itself from the
1548 stream of per-object check-and-repair-results.
1550 Note that the "ERROR:" indication will only be emitted if traversal stops,
1551 which will only occur if an unrecoverable directory is encountered. If a
1552 file or directory repair fails, the traversal will continue, and the repair
1553 failure will be indicated in the JSON data (in the "repair-successful" key).
1555 ``POST $DIRURL?t=start-manifest`` (must add &ophandle=XYZ)
1557 This operation generates a "manfest" of the given directory tree, mostly
1558 for debugging. This is a table of (path, filecap/dircap), for every object
1559 reachable from the starting directory. The path will be slash-joined, and
1560 the filecap/dircap will contain a link to the object in question. This page
1561 gives immediate access to every object in the virtual filesystem subtree.
1563 This operation uses the same ophandle= mechanism as deep-check. The
1564 corresponding /operations/$HANDLE page has three different forms. The
1565 default is output=HTML.
1567 If output=text is added to the query args, the results will be a text/plain
1568 list. The first line is special: it is either "finished: yes" or "finished:
1569 no"; if the operation is not finished, you must periodically reload the
1570 page until it completes. The rest of the results are a plaintext list, with
1571 one file/dir per line, slash-separated, with the filecap/dircap separated
1574 If output=JSON is added to the queryargs, then the results will be a
1575 JSON-formatted dictionary with six keys. Note that because large directory
1576 structures can result in very large JSON results, the full results will not
1577 be available until the operation is complete (i.e. until output["finished"]
1580 finished (bool): if False then you must reload the page until True
1581 origin_si (base32 str): the storage index of the starting point
1582 manifest: list of (path, cap) tuples, where path is a list of strings.
1583 verifycaps: list of (printable) verify cap strings
1584 storage-index: list of (base32) storage index strings
1585 stats: a dictionary with the same keys as the t=start-deep-stats command
1588 ``POST $DIRURL?t=start-deep-size`` (must add &ophandle=XYZ)
1590 This operation generates a number (in bytes) containing the sum of the
1591 filesize of all directories and immutable files reachable from the given
1592 directory. This is a rough lower bound of the total space consumed by this
1593 subtree. It does not include space consumed by mutable files, nor does it
1594 take expansion or encoding overhead into account. Later versions of the
1595 code may improve this estimate upwards.
1597 The /operations/$HANDLE status output consists of two lines of text::
1602 ``POST $DIRURL?t=start-deep-stats`` (must add &ophandle=XYZ)
1604 This operation performs a recursive walk of all files and directories
1605 reachable from the given directory, and generates a collection of
1606 statistics about those objects.
1608 The result (obtained from the /operations/$OPHANDLE page) is a
1609 JSON-serialized dictionary with the following keys (note that some of these
1610 keys may be missing until 'finished' is True)::
1612 finished: (bool) True if the operation has finished, else False
1613 count-immutable-files: count of how many CHK files are in the set
1614 count-mutable-files: same, for mutable files (does not include directories)
1615 count-literal-files: same, for LIT files (data contained inside the URI)
1616 count-files: sum of the above three
1617 count-directories: count of directories
1618 count-unknown: count of unrecognized objects (perhaps from the future)
1619 size-immutable-files: total bytes for all CHK files in the set, =deep-size
1620 size-mutable-files (TODO): same, for current version of all mutable files
1621 size-literal-files: same, for LIT files
1622 size-directories: size of directories (includes size-literal-files)
1623 size-files-histogram: list of (minsize, maxsize, count) buckets,
1624 with a histogram of filesizes, 5dB/bucket,
1625 for both literal and immutable files
1626 largest-directory: number of children in the largest directory
1627 largest-immutable-file: number of bytes in the largest CHK file
1629 size-mutable-files is not implemented, because it would require extra
1630 queries to each mutable file to get their size. This may be implemented in
1633 Assuming no sharing, the basic space consumed by a single root directory is
1634 the sum of size-immutable-files, size-mutable-files, and size-directories.
1635 The actual disk space used by the shares is larger, because of the
1636 following sources of overhead::
1639 expansion due to erasure coding
1640 share management data (leases)
1641 backend (ext3) minimum block size
1643 ``POST $URL?t=stream-manifest``
1645 This operation performs a recursive walk of all files and directories
1646 reachable from the given starting point. For each such unique object
1647 (duplicates are skipped), a single line of JSON is emitted to the HTTP
1648 response channel (or an error indication, see below). When the walk is
1649 complete, a final line of JSON is emitted which contains the accumulated
1650 file-size/count "deep-stats" data.
1652 A CLI tool can split the response stream on newlines into "response units",
1653 and parse each response unit as JSON. Each such parsed unit will be a
1654 dictionary, and will contain at least the "type" key: a string, one of
1655 "file", "directory", or "stats".
1657 For all units that have a type of "file" or "directory", the dictionary will
1658 contain the following keys::
1660 "path": a list of strings, with the path that is traversed to reach the
1662 "cap": a write-cap URI for the file or directory, if available, else a
1664 "verifycap": a verify-cap URI for the file or directory
1665 "repaircap": an URI for the weakest cap that can still be used to repair
1667 "storage-index": a base32 storage index for the object
1669 Note that non-distributed files (i.e. LIT files) will have values of None
1670 for verifycap, repaircap, and storage-index, since these files can neither
1671 be verified nor repaired, and are not stored on the storage servers.
1673 The last unit in the stream will have a type of "stats", and will contain
1674 the keys described in the "start-deep-stats" operation, below.
1676 If any errors occur during the traversal (specifically if a directory is
1677 unrecoverable, such that further traversal is not possible), an error
1678 indication is written to the response body, instead of the usual line of
1679 JSON. This error indication line will begin with the string "ERROR:" (in all
1680 caps), and contain a summary of the error on the rest of the line. The
1681 remaining lines of the response body will be a python exception. The client
1682 application should look for the ERROR: and stop processing JSON as soon as
1683 it is seen. The line just before the ERROR: will describe the directory that
1684 was untraversable, since the manifest entry is emitted to the HTTP response
1685 body before the child is traversed.
1691 The portion of the web namespace that begins with "/uri" (and "/named") is
1692 dedicated to giving users (both humans and programs) access to the Tahoe
1693 virtual filesystem. The rest of the namespace provides status information
1694 about the state of the Tahoe node.
1696 ``GET /`` (the root page)
1698 This is the "Welcome Page", and contains a few distinct sections::
1700 Node information: library versions, local nodeid, services being provided.
1702 Filesystem Access Forms: create a new directory, view a file/directory by
1703 URI, upload a file (unlinked), download a file by
1706 Grid Status: introducer information, helper information, connected storage
1711 This page lists all active uploads and downloads, and contains a short list
1712 of recent upload/download operations. Each operation has a link to a page
1713 that describes file sizes, servers that were involved, and the time consumed
1714 in each phase of the operation.
1716 A GET of /status/?t=json will contain a machine-readable subset of the same
1717 data. It returns a JSON-encoded dictionary. The only key defined at this
1718 time is "active", with a value that is a list of operation dictionaries, one
1719 for each active operation. Once an operation is completed, it will no longer
1720 appear in data["active"] .
1722 Each op-dict contains a "type" key, one of "upload", "download",
1723 "mapupdate", "publish", or "retrieve" (the first two are for immutable
1724 files, while the latter three are for mutable files and directories).
1726 The "upload" op-dict will contain the following keys::
1728 type (string): "upload"
1729 storage-index-string (string): a base32-encoded storage index
1730 total-size (int): total size of the file
1731 status (string): current status of the operation
1732 progress-hash (float): 1.0 when the file has been hashed
1733 progress-ciphertext (float): 1.0 when the file has been encrypted.
1734 progress-encode-push (float): 1.0 when the file has been encoded and
1735 pushed to the storage servers. For helper
1736 uploads, the ciphertext value climbs to 1.0
1737 first, then encoding starts. For unassisted
1738 uploads, ciphertext and encode-push progress
1739 will climb at the same pace.
1741 The "download" op-dict will contain the following keys::
1743 type (string): "download"
1744 storage-index-string (string): a base32-encoded storage index
1745 total-size (int): total size of the file
1746 status (string): current status of the operation
1747 progress (float): 1.0 when the file has been fully downloaded
1749 Front-ends which want to report progress information are advised to simply
1750 average together all the progress-* indicators. A slightly more accurate
1751 value can be found by ignoring the progress-hash value (since the current
1752 implementation hashes synchronously, so clients will probably never see
1753 progress-hash!=1.0).
1755 ``GET /provisioning/``
1757 This page provides a basic tool to predict the likely storage and bandwidth
1758 requirements of a large Tahoe grid. It provides forms to input things like
1759 total number of users, number of files per user, average file size, number
1760 of servers, expansion ratio, hard drive failure rate, etc. It then provides
1761 numbers like how many disks per server will be needed, how many read
1762 operations per second should be expected, and the likely MTBF for files in
1763 the grid. This information is very preliminary, and the model upon which it
1764 is based still needs a lot of work.
1766 ``GET /helper_status/``
1768 If the node is running a helper (i.e. if [helper]enabled is set to True in
1769 tahoe.cfg), then this page will provide a list of all the helper operations
1770 currently in progress. If "?t=json" is added to the URL, it will return a
1771 JSON-formatted list of helper statistics, which can then be used to produce
1772 graphs to indicate how busy the helper is.
1774 ``GET /statistics/``
1776 This page provides "node statistics", which are collected from a variety of
1779 load_monitor: every second, the node schedules a timer for one second in
1780 the future, then measures how late the subsequent callback
1781 is. The "load_average" is this tardiness, measured in
1782 seconds, averaged over the last minute. It is an indication
1783 of a busy node, one which is doing more work than can be
1784 completed in a timely fashion. The "max_load" value is the
1785 highest value that has been seen in the last 60 seconds.
1787 cpu_monitor: every minute, the node uses time.clock() to measure how much
1788 CPU time it has used, and it uses this value to produce
1789 1min/5min/15min moving averages. These values range from 0%
1790 (0.0) to 100% (1.0), and indicate what fraction of the CPU
1791 has been used by the Tahoe node. Not all operating systems
1792 provide meaningful data to time.clock(): they may report 100%
1793 CPU usage at all times.
1795 uploader: this counts how many immutable files (and bytes) have been
1796 uploaded since the node was started
1798 downloader: this counts how many immutable files have been downloaded
1799 since the node was started
1801 publishes: this counts how many mutable files (including directories) have
1802 been modified since the node was started
1804 retrieves: this counts how many mutable files (including directories) have
1805 been read since the node was started
1807 There are other statistics that are tracked by the node. The "raw stats"
1808 section shows a formatted dump of all of them.
1810 By adding "?t=json" to the URL, the node will return a JSON-formatted
1811 dictionary of stats values, which can be used by other tools to produce
1812 graphs of node behavior. The misc/munin/ directory in the source
1813 distribution provides some tools to produce these graphs.
1815 ``GET /`` (introducer status)
1817 For Introducer nodes, the welcome page displays information about both
1818 clients and servers which are connected to the introducer. Servers make
1819 "service announcements", and these are listed in a table. Clients will
1820 subscribe to hear about service announcements, and these subscriptions are
1821 listed in a separate table. Both tables contain information about what
1822 version of Tahoe is being run by the remote node, their advertised and
1823 outbound IP addresses, their nodeid and nickname, and how long they have
1826 By adding "?t=json" to the URL, the node will return a JSON-formatted
1827 dictionary of stats values, which can be used to produce graphs of connected
1828 clients over time. This dictionary has the following keys::
1830 ["subscription_summary"] : a dictionary mapping service name (like
1831 "storage") to an integer with the number of
1832 clients that have subscribed to hear about that
1834 ["announcement_summary"] : a dictionary mapping service name to an integer
1835 with the number of servers which are announcing
1837 ["announcement_distinct_hosts"] : a dictionary mapping service name to an
1838 integer which represents the number of
1839 distinct hosts that are providing that
1840 service. If two servers have announced
1841 FURLs which use the same hostnames (but
1842 different ports and tubids), they are
1843 considered to be on the same host.
1846 Static Files in /public_html
1847 ============================
1849 The web-API server will take any request for a URL that starts with /static
1850 and serve it from a configurable directory which defaults to
1851 $BASEDIR/public_html . This is configured by setting the "[node]web.static"
1852 value in $BASEDIR/tahoe.cfg . If this is left at the default value of
1853 "public_html", then http://localhost:3456/static/subdir/foo.html will be
1854 served with the contents of the file $BASEDIR/public_html/subdir/foo.html .
1856 This can be useful to serve a javascript application which provides a
1857 prettier front-end to the rest of the Tahoe web-API.
1860 Safety and Security Issues -- Names vs. URIs
1861 ============================================
1863 Summary: use explicit file- and dir- caps whenever possible, to reduce the
1864 potential for surprises when the filesystem structure is changed.
1866 Tahoe provides a mutable filesystem, but the ways that the filesystem can
1867 change are limited. The only thing that can change is that the mapping from
1868 child names to child objects that each directory contains can be changed by
1869 adding a new child name pointing to an object, removing an existing child name,
1870 or changing an existing child name to point to a different object.
1872 Obviously if you query Tahoe for information about the filesystem and then act
1873 to change the filesystem (such as by getting a listing of the contents of a
1874 directory and then adding a file to the directory), then the filesystem might
1875 have been changed after you queried it and before you acted upon it. However,
1876 if you use the URI instead of the pathname of an object when you act upon the
1877 object, then the only change that can happen is if the object is a directory
1878 then the set of child names it has might be different. If, on the other hand,
1879 you act upon the object using its pathname, then a different object might be in
1880 that place, which can result in more kinds of surprises.
1882 For example, suppose you are writing code which recursively downloads the
1883 contents of a directory. The first thing your code does is fetch the listing
1884 of the contents of the directory. For each child that it fetched, if that
1885 child is a file then it downloads the file, and if that child is a directory
1886 then it recurses into that directory. Now, if the download and the recurse
1887 actions are performed using the child's name, then the results might be
1888 wrong, because for example a child name that pointed to a sub-directory when
1889 you listed the directory might have been changed to point to a file (in which
1890 case your attempt to recurse into it would result in an error and the file
1891 would be skipped), or a child name that pointed to a file when you listed the
1892 directory might now point to a sub-directory (in which case your attempt to
1893 download the child would result in a file containing HTML text describing the
1896 If your recursive algorithm uses the uri of the child instead of the name of
1897 the child, then those kinds of mistakes just can't happen. Note that both the
1898 child's name and the child's URI are included in the results of listing the
1899 parent directory, so it isn't any harder to use the URI for this purpose.
1901 The read and write caps in a given directory node are separate URIs, and
1902 can't be assumed to point to the same object even if they were retrieved in
1903 the same operation (although the web-API server attempts to ensure this
1904 in most cases). If you need to rely on that property, you should explicitly
1905 verify it. More generally, you should not make assumptions about the
1906 internal consistency of the contents of mutable directories. As a result
1907 of the signatures on mutable object versions, it is guaranteed that a given
1908 version was written in a single update, but -- as in the case of a file --
1909 the contents may have been chosen by a malicious writer in a way that is
1910 designed to confuse applications that rely on their consistency.
1912 In general, use names if you want "whatever object (whether file or
1913 directory) is found by following this name (or sequence of names) when my
1914 request reaches the server". Use URIs if you want "this particular object".
1920 Tahoe uses both mutable and immutable files. Mutable files can be created
1921 explicitly by doing an upload with ?mutable=true added, or implicitly by
1922 creating a new directory (since a directory is just a special way to
1923 interpret a given mutable file).
1925 Mutable files suffer from the same consistency-vs-availability tradeoff that
1926 all distributed data storage systems face. It is not possible to
1927 simultaneously achieve perfect consistency and perfect availability in the
1928 face of network partitions (servers being unreachable or faulty).
1930 Tahoe tries to achieve a reasonable compromise, but there is a basic rule in
1931 place, known as the Prime Coordination Directive: "Don't Do That". What this
1932 means is that if write-access to a mutable file is available to several
1933 parties, then those parties are responsible for coordinating their activities
1934 to avoid multiple simultaneous updates. This could be achieved by having
1935 these parties talk to each other and using some sort of locking mechanism, or
1936 by serializing all changes through a single writer.
1938 The consequences of performing uncoordinated writes can vary. Some of the
1939 writers may lose their changes, as somebody else wins the race condition. In
1940 many cases the file will be left in an "unhealthy" state, meaning that there
1941 are not as many redundant shares as we would like (reducing the reliability
1942 of the file against server failures). In the worst case, the file can be left
1943 in such an unhealthy state that no version is recoverable, even the old ones.
1944 It is this small possibility of data loss that prompts us to issue the Prime
1945 Coordination Directive.
1947 Tahoe nodes implement internal serialization to make sure that a single Tahoe
1948 node cannot conflict with itself. For example, it is safe to issue two
1949 directory modification requests to a single tahoe node's web-API server at the
1950 same time, because the Tahoe node will internally delay one of them until
1951 after the other has finished being applied. (This feature was introduced in
1952 Tahoe-1.1; back with Tahoe-1.0 the web client was responsible for serializing
1953 web requests themselves).
1955 For more details, please see the "Consistency vs Availability" and "The Prime
1956 Coordination Directive" sections of `mutable.rst <../specifications/mutable.rst>`_.
1962 Gateway nodes may find it necessary to prohibit access to certain files. The
1963 web-API has a facility to block access to filecaps by their storage index,
1964 returning a 403 "Forbidden" error instead of the original file.
1966 This blacklist is recorded in $NODEDIR/access.blacklist, and contains one
1967 blocked file per line. Comment lines (starting with ``#``) are ignored. Each
1968 line consists of the storage-index (in the usual base32 format as displayed
1969 by the "More Info" page, or by the "tahoe debug dump-cap" command), followed
1970 by whitespace, followed by a reason string, which will be included in the 403
1971 error message. This could hold a URL to a page that explains why the file is
1972 blocked, for example.
1974 So for example, if you found a need to block access to a file with filecap
1975 ``URI:CHK:n7r3m6wmomelk4sep3kw5cvduq:os7ijw5c3maek7pg65e5254k2fzjflavtpejjyhshpsxuqzhcwwq:3:20:14861``,
1976 you could do the following::
1978 tahoe debug dump-cap URI:CHK:n7r3m6wmomelk4sep3kw5cvduq:os7ijw5c3maek7pg65e5254k2fzjflavtpejjyhshpsxuqzhcwwq:3:20:14861
1979 -> storage index: whpepioyrnff7orecjolvbudeu
1980 echo "whpepioyrnff7orecjolvbudeu my puppy told me to" >>$NODEDIR/access.blacklist
1981 tahoe restart $NODEDIR
1982 tahoe get URI:CHK:n7r3m6wmomelk4sep3kw5cvduq:os7ijw5c3maek7pg65e5254k2fzjflavtpejjyhshpsxuqzhcwwq:3:20:14861
1983 -> error, 403 Access Prohibited: my puppy told me to
1985 The ``access.blacklist`` file will be checked each time a file or directory
1986 is accessed: the file's ``mtime`` is used to decide whether it need to be
1987 reloaded. Therefore no node restart is necessary when creating the initial
1988 blacklist, nor when adding second, third, or additional entries to the list.
1989 When modifying the file, be careful to update it atomically, otherwise a
1990 request may arrive while the file is only halfway written, and the partial
1991 file may be incorrectly parsed.
1993 The blacklist is applied to all access paths (including FTP, SFTP, and CLI
1994 operations), not just the web-API. The blacklist also applies to directories.
1995 If a directory is blacklisted, the gateway will refuse access to both that
1996 directory and any child files/directories underneath it, when accessed via
1997 "DIRCAP/SUBDIR/FILENAME" -style URLs. Users who go directly to the child
1998 file/dir will bypass the blacklist.
2000 The node will log the SI of the file being blocked, and the reason code, into
2001 the ``logs/twistd.log`` file.
2004 .. [1] URLs and HTTP and UTF-8, Oh My
2006 HTTP does not provide a mechanism to specify the character set used to
2007 encode non-ASCII names in URLs
2008 (`RFC3986#2.1 <http://tools.ietf.org/html/rfc3986#section-2.1>`_).
2009 We prefer the convention that the ``filename=`` argument shall be a
2010 URL-escaped UTF-8 encoded Unicode string.
2011 For example, suppose we want to provoke the server into using a filename of
2012 "f i a n c e-acute e" (i.e. f i a n c U+00E9 e). The UTF-8 encoding of this
2013 is 0x66 0x69 0x61 0x6e 0x63 0xc3 0xa9 0x65 (or "fianc\\xC3\\xA9e", as python's
2014 ``repr()`` function would show). To encode this into a URL, the non-printable
2015 characters must be escaped with the urlencode ``%XX`` mechanism, giving
2016 us "fianc%C3%A9e". Thus, the first line of the HTTP request will be
2017 "``GET /uri/CAP...?save=true&filename=fianc%C3%A9e HTTP/1.1``". Not all
2018 browsers provide this: IE7 by default uses the Latin-1 encoding, which is
2019 "fianc%E9e" (although it has a configuration option to send URLs as UTF-8).
2021 The response header will need to indicate a non-ASCII filename. The actual
2022 mechanism to do this is not clear. For ASCII filenames, the response header
2025 Content-Disposition: attachment; filename="english.txt"
2027 If Tahoe were to enforce the UTF-8 convention, it would need to decode the
2028 URL argument into a Unicode string, and then encode it back into a sequence
2029 of bytes when creating the response header. One possibility would be to use
2030 unencoded UTF-8. Developers suggest that IE7 might accept this::
2032 #1: Content-Disposition: attachment; filename="fianc\xC3\xA9e"
2033 (note, the last four bytes of that line, not including the newline, are
2034 0xC3 0xA9 0x65 0x22)
2036 `RFC2231#4 <http://tools.ietf.org/html/rfc2231#section-4>`_
2037 (dated 1997): suggests that the following might work, and
2038 `some developers have reported <http://markmail.org/message/dsjyokgl7hv64ig3>`_
2039 that it is supported by Firefox (but not IE7)::
2041 #2: Content-Disposition: attachment; filename*=utf-8''fianc%C3%A9e
2043 My reading of `RFC2616#19.5.1 <http://tools.ietf.org/html/rfc2616#section-19.5.1>`_
2044 (which defines Content-Disposition) says that the filename= parameter is
2045 defined to be wrapped in quotes (presumably to allow spaces without breaking
2046 the parsing of subsequent parameters), which would give us::
2048 #3: Content-Disposition: attachment; filename*=utf-8''"fianc%C3%A9e"
2050 However this is contrary to the examples in the email thread listed above.
2052 Developers report that IE7 (when it is configured for UTF-8 URL encoding,
2053 which is not the default in Asian countries), will accept::
2055 #4: Content-Disposition: attachment; filename=fianc%C3%A9e
2057 However, for maximum compatibility, Tahoe simply copies bytes from the URL
2058 into the response header, rather than enforcing the UTF-8 convention. This
2059 means it does not try to decode the filename from the URL argument, nor does
2060 it encode the filename into the response header.