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. `Get 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. `Deleting 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. `Get 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. `Deleting 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`_
40 Enabling the web-API port
41 =========================
43 Every Tahoe node is capable of running a built-in HTTP server. To enable
44 this, just write a port number into the "[node]web.port" line of your node's
45 tahoe.cfg file. For example, writing "web.port = 3456" into the "[node]"
46 section of $NODEDIR/tahoe.cfg will cause the node to run a webserver on port
49 This string is actually a Twisted "strports" specification, meaning you can
50 get more control over the interface to which the server binds by supplying
51 additional arguments. For more details, see the documentation on
52 `twisted.application.strports
53 <http://twistedmatrix.com/documents/current/api/twisted.application.strports.html>`_.
55 Writing "tcp:3456:interface=127.0.0.1" into the web.port line does the same
56 but binds to the loopback interface, ensuring that only the programs on the
57 local host can connect. Using "ssl:3456:privateKey=mykey.pem:certKey=cert.pem"
60 This webport can be set when the node is created by passing a --webport
61 option to the 'tahoe create-node' command. By default, the node listens on
62 port 3456, on the loopback (127.0.0.1) interface.
64 Basic Concepts: GET, PUT, DELETE, POST
65 ======================================
67 As described in `docs/architecture.rst <../architecture.rst>`_, each file
68 and directory in a Tahoe virtual filesystem is referenced by an identifier
69 that combines the designation of the object with the authority to do something
70 with it (such as read or modify the contents). This identifier is called a
71 "read-cap" or "write-cap", depending upon whether it enables read-only or
72 read-write access. These "caps" are also referred to as URIs (which may be
73 confusing because they are not currently `RFC3986
74 <http://tools.ietf.org/html/rfc3986>`_-compliant URIs).
76 The Tahoe web-based API is "REST-ful", meaning it implements the concepts of
77 "REpresentational State Transfer": the original scheme by which the World
78 Wide Web was intended to work. Each object (file or directory) is referenced
79 by a URL that includes the read- or write- cap. HTTP methods (GET, PUT, and
80 DELETE) are used to manipulate these objects. You can think of the URL as a
81 noun, and the method as a verb.
83 In REST, the GET method is used to retrieve information about an object, or
84 to retrieve some representation of the object itself. When the object is a
85 file, the basic GET method will simply return the contents of that file.
86 Other variations (generally implemented by adding query parameters to the
87 URL) will return information about the object, such as metadata. GET
88 operations are required to have no side-effects.
90 PUT is used to upload new objects into the filesystem, or to replace an
91 existing link or the contents of a mutable file. DELETE is used to unlink
92 objects from directories. Both PUT and DELETE are required to be idempotent:
93 performing the same operation multiple times must have the same side-effects
94 as only performing it once.
96 POST is used for more complicated actions that cannot be expressed as a GET,
97 PUT, or DELETE. POST operations can be thought of as a method call: sending
98 some message to the object referenced by the URL. In Tahoe, POST is also used
99 for operations that must be triggered by an HTML form (including upload and
100 delete), because otherwise a regular web browser has no way to accomplish
101 these tasks. In general, everything that can be done with a PUT or DELETE can
102 also be done with a POST.
104 Tahoe's web API is designed for two different kinds of consumer. The first is
105 a program that needs to manipulate the virtual file system. Such programs are
106 expected to use the RESTful interface described above. The second is a human
107 using a standard web browser to work with the filesystem. This user is given
108 a series of HTML pages with links to download files, and forms that use POST
109 actions to upload, rename, and delete files.
111 When an error occurs, the HTTP response code will be set to an appropriate
112 400-series code (like 404 Not Found for an unknown childname, or 400 Bad Request
113 when the parameters to a web-API operation are invalid), and the HTTP response
114 body will usually contain a few lines of explanation as to the cause of the
115 error and possible responses. Unusual exceptions may result in a 500 Internal
116 Server Error as a catch-all, with a default response body containing
117 a Nevow-generated HTML-ized representation of the Python exception stack trace
118 that caused the problem. CLI programs which want to copy the response body to
119 stderr should provide an "Accept: text/plain" header to their requests to get
120 a plain text stack trace instead. If the Accept header contains ``*/*``, or
121 ``text/*``, or text/html (or if there is no Accept header), HTML tracebacks will
127 Tahoe uses a variety of read- and write- caps to identify files and
128 directories. The most common of these is the "immutable file read-cap", which
129 is used for most uploaded files. These read-caps look like the following::
131 URI:CHK:ime6pvkaxuetdfah2p2f35pe54:4btz54xk3tew6nd4y2ojpxj4m6wxjqqlwnztgre6gnjgtucd5r4a:3:10:202
133 The next most common is a "directory write-cap", which provides both read and
134 write access to a directory, and look like this::
136 URI:DIR2:djrdkfawoqihigoett4g6auz6a:jx5mplfpwexnoqff7y5e4zjus4lidm76dcuarpct7cckorh2dpgq
138 There are also "directory read-caps", which start with "URI:DIR2-RO:", and
139 give read-only access to a directory. Finally there are also mutable file
140 read- and write- caps, which start with "URI:SSK", and give access to mutable
143 (Later versions of Tahoe will make these strings shorter, and will remove the
144 unfortunate colons, which must be escaped when these caps are embedded in
147 To refer to any Tahoe object through the web API, you simply need to combine
148 a prefix (which indicates the HTTP server to use) with the cap (which
149 indicates which object inside that server to access). Since the default Tahoe
150 webport is 3456, the most common prefix is one that will use a local node
151 listening on this port::
153 http://127.0.0.1:3456/uri/ + $CAP
155 So, to access the directory named above (which happens to be the
156 publically-writeable sample directory on the Tahoe test grid, described at
157 http://allmydata.org/trac/tahoe/wiki/TestGrid), the URL would be::
159 http://127.0.0.1:3456/uri/URI%3ADIR2%3Adjrdkfawoqihigoett4g6auz6a%3Ajx5mplfpwexnoqff7y5e4zjus4lidm76dcuarpct7cckorh2dpgq/
161 (note that the colons in the directory-cap are url-encoded into "%3A"
164 Likewise, to access the file named above, use::
166 http://127.0.0.1:3456/uri/URI%3ACHK%3Aime6pvkaxuetdfah2p2f35pe54%3A4btz54xk3tew6nd4y2ojpxj4m6wxjqqlwnztgre6gnjgtucd5r4a%3A3%3A10%3A202
168 In the rest of this document, we'll use "$DIRCAP" as shorthand for a read-cap
169 or write-cap that refers to a directory, and "$FILECAP" to abbreviate a cap
170 that refers to a file (whether mutable or immutable). So those URLs above can
173 http://127.0.0.1:3456/uri/$DIRCAP/
174 http://127.0.0.1:3456/uri/$FILECAP
176 The operation summaries below will abbreviate these further, by eliding the
177 server prefix. They will be displayed like this::
186 Tahoe directories contain named child entries, just like directories in a regular
187 local filesystem. These child entries, called "dirnodes", consist of a name,
188 metadata, a write slot, and a read slot. The write and read slots normally contain
189 a write-cap and read-cap referring to the same object, which can be either a file
190 or a subdirectory. The write slot may be empty (actually, both may be empty,
191 but that is unusual).
193 If you have a Tahoe URL that refers to a directory, and want to reference a
194 named child inside it, just append the child name to the URL. For example, if
195 our sample directory contains a file named "welcome.txt", we can refer to
198 http://127.0.0.1:3456/uri/$DIRCAP/welcome.txt
200 (or http://127.0.0.1:3456/uri/URI%3ADIR2%3Adjrdkfawoqihigoett4g6auz6a%3Ajx5mplfpwexnoqff7y5e4zjus4lidm76dcuarpct7cckorh2dpgq/welcome.txt)
202 Multiple levels of subdirectories can be handled this way::
204 http://127.0.0.1:3456/uri/$DIRCAP/tahoe-source/docs/architecture.rst
206 In this document, when we need to refer to a URL that references a file using
207 this child-of-some-directory format, we'll use the following string::
209 /uri/$DIRCAP/[SUBDIRS../]FILENAME
211 The "[SUBDIRS../]" part means that there are zero or more (optional)
212 subdirectory names in the middle of the URL. The "FILENAME" at the end means
213 that this whole URL refers to a file of some sort, rather than to a
216 When we need to refer specifically to a directory in this way, we'll write::
218 /uri/$DIRCAP/[SUBDIRS../]SUBDIR
221 Note that all components of pathnames in URLs are required to be UTF-8
222 encoded, so "resume.doc" (with an acute accent on both E's) would be accessed
225 http://127.0.0.1:3456/uri/$DIRCAP/r%C3%A9sum%C3%A9.doc
227 Also note that the filenames inside upload POST forms are interpreted using
228 whatever character set was provided in the conventional '_charset' field, and
229 defaults to UTF-8 if not otherwise specified. The JSON representation of each
230 directory contains native Unicode strings. Tahoe directories are specified to
231 contain Unicode filenames, and cannot contain binary strings that are not
232 representable as such.
234 All Tahoe operations that refer to existing files or directories must include
235 a suitable read- or write- cap in the URL: the web-API server won't add one
236 for you. If you don't know the cap, you can't access the file. This allows
237 the security properties of Tahoe caps to be extended across the web-API
240 Slow Operations, Progress, and Cancelling
241 =========================================
243 Certain operations can be expected to take a long time. The "t=deep-check",
244 described below, will recursively visit every file and directory reachable
245 from a given starting point, which can take minutes or even hours for
246 extremely large directory structures. A single long-running HTTP request is a
247 fragile thing: proxies, NAT boxes, browsers, and users may all grow impatient
248 with waiting and give up on the connection.
250 For this reason, long-running operations have an "operation handle", which
251 can be used to poll for status/progress messages while the operation
252 proceeds. This handle can also be used to cancel the operation. These handles
253 are created by the client, and passed in as a an "ophandle=" query argument
254 to the POST or PUT request which starts the operation. The following
255 operations can then be used to retrieve status:
257 ``GET /operations/$HANDLE?output=HTML (with or without t=status)``
259 ``GET /operations/$HANDLE?output=JSON (same)``
261 These two retrieve the current status of the given operation. Each operation
262 presents a different sort of information, but in general the page retrieved
265 * whether the operation is complete, or if it is still running
266 * how much of the operation is complete, and how much is left, if possible
268 Note that the final status output can be quite large: a deep-manifest of a
269 directory structure with 300k directories and 200k unique files is about
270 275MB of JSON, and might take two minutes to generate. For this reason, the
271 full status is not provided until the operation has completed.
273 The HTML form will include a meta-refresh tag, which will cause a regular
274 web browser to reload the status page about 60 seconds later. This tag will
275 be removed once the operation has completed.
277 There may be more status information available under
278 /operations/$HANDLE/$ETC : i.e., the handle forms the root of a URL space.
280 ``POST /operations/$HANDLE?t=cancel``
282 This terminates the operation, and returns an HTML page explaining what was
283 cancelled. If the operation handle has already expired (see below), this
284 POST will return a 404, which indicates that the operation is no longer
285 running (either it was completed or terminated). The response body will be
286 the same as a GET /operations/$HANDLE on this operation handle, and the
287 handle will be expired immediately afterwards.
289 The operation handle will eventually expire, to avoid consuming an unbounded
290 amount of memory. The handle's time-to-live can be reset at any time, by
291 passing a retain-for= argument (with a count of seconds) to either the
292 initial POST that starts the operation, or the subsequent GET request which
293 asks about the operation. For example, if a 'GET
294 /operations/$HANDLE?output=JSON&retain-for=600' query is performed, the
295 handle will remain active for 600 seconds (10 minutes) after the GET was
298 In addition, if the GET includes a release-after-complete=True argument, and
299 the operation has completed, the operation handle will be released
302 If a retain-for= argument is not used, the default handle lifetimes are:
304 * handles will remain valid at least until their operation finishes
305 * uncollected handles for finished operations (i.e. handles for
306 operations that have finished but for which the GET page has not been
307 accessed since completion) will remain valid for four days, or for
308 the total time consumed by the operation, whichever is greater.
309 * collected handles (i.e. the GET page has been retrieved at least once
310 since the operation completed) will remain valid for one day.
312 Many "slow" operations can begin to use unacceptable amounts of memory when
313 operating on large directory structures. The memory usage increases when the
314 ophandle is polled, as the results must be copied into a JSON string, sent
315 over the wire, then parsed by a client. So, as an alternative, many "slow"
316 operations have streaming equivalents. These equivalents do not use operation
317 handles. Instead, they emit line-oriented status results immediately. Client
318 code can cancel the operation by simply closing the HTTP connection.
320 Programmatic Operations
321 =======================
323 Now that we know how to build URLs that refer to files and directories in a
324 Tahoe virtual filesystem, what sorts of operations can we do with those URLs?
325 This section contains a catalog of GET, PUT, DELETE, and POST operations that
326 can be performed on these URLs. This set of operations are aimed at programs
327 that use HTTP to communicate with a Tahoe node. A later section describes
328 operations that are intended for web browsers.
333 ``GET /uri/$FILECAP``
335 ``GET /uri/$DIRCAP/[SUBDIRS../]FILENAME``
337 This will retrieve the contents of the given file. The HTTP response body
338 will contain the sequence of bytes that make up the file.
340 To view files in a web browser, you may want more control over the
341 Content-Type and Content-Disposition headers. Please see the next section
342 "Browser Operations", for details on how to modify these URLs for that
345 Writing/Uploading A File
346 ------------------------
348 ``PUT /uri/$FILECAP``
350 ``PUT /uri/$DIRCAP/[SUBDIRS../]FILENAME``
352 Upload a file, using the data from the HTTP request body, and add whatever
353 child links and subdirectories are necessary to make the file available at
354 the given location. Once this operation succeeds, a GET on the same URL will
355 retrieve the same contents that were just uploaded. This will create any
356 necessary intermediate subdirectories.
358 To use the /uri/$FILECAP form, $FILECAP must be a write-cap for a mutable file.
360 In the /uri/$DIRCAP/[SUBDIRS../]FILENAME form, if the target file is a
361 writeable mutable file, that file's contents will be overwritten in-place. If
362 it is a read-cap for a mutable file, an error will occur. If it is an
363 immutable file, the old file will be discarded, and a new one will be put in
366 When creating a new file, if "mutable=true" is in the query arguments, the
367 operation will create a mutable file instead of an immutable one.
369 This returns the file-cap of the resulting file. If a new file was created
370 by this method, the HTTP response code (as dictated by rfc2616) will be set
371 to 201 CREATED. If an existing file was replaced or modified, the response
374 Note that the 'curl -T localfile http://127.0.0.1:3456/uri/$DIRCAP/foo.txt'
375 command can be used to invoke this operation.
379 This uploads a file, and produces a file-cap for the contents, but does not
380 attach the file into the filesystem. No directories will be modified by
381 this operation. The file-cap is returned as the body of the HTTP response.
383 If "mutable=true" is in the query arguments, the operation will create a
384 mutable file, and return its write-cap in the HTTP respose. The default is
385 to create an immutable file, returning the read-cap as a response.
387 Creating A New Directory
388 ------------------------
390 ``POST /uri?t=mkdir``
394 Create a new empty directory and return its write-cap as the HTTP response
395 body. This does not make the newly created directory visible from the
396 filesystem. The "PUT" operation is provided for backwards compatibility:
397 new code should use POST.
399 ``POST /uri?t=mkdir-with-children``
401 Create a new directory, populated with a set of child nodes, and return its
402 write-cap as the HTTP response body. The new directory is not attached to
403 any other directory: the returned write-cap is the only reference to it.
405 Initial children are provided as the body of the POST form (this is more
406 efficient than doing separate mkdir and set_children operations). If the
407 body is empty, the new directory will be empty. If not empty, the body will
408 be interpreted as a UTF-8 JSON-encoded dictionary of children with which the
409 new directory should be populated, using the same format as would be
410 returned in the 'children' value of the t=json GET request, described below.
411 Each dictionary key should be a child name, and each value should be a list
412 of [TYPE, PROPDICT], where PROPDICT contains "rw_uri", "ro_uri", and
413 "metadata" keys (all others are ignored). For example, the PUT request body
417 "Fran\u00e7ais": [ "filenode", {
418 "ro_uri": "URI:CHK:...",
421 "ctime": 1202777696.7564139,
422 "mtime": 1202777696.7564139,
424 "linkcrtime": 1202777696.7564139,
425 "linkmotime": 1202777696.7564139
427 "subdir": [ "dirnode", {
428 "rw_uri": "URI:DIR2:...",
429 "ro_uri": "URI:DIR2-RO:...",
431 "ctime": 1202778102.7589991,
432 "mtime": 1202778111.2160511,
434 "linkcrtime": 1202777696.7564139,
435 "linkmotime": 1202777696.7564139
439 For forward-compatibility, a mutable directory can also contain caps in
440 a format that is unknown to the web-API server. When such caps are retrieved
441 from a mutable directory in a "ro_uri" field, they will be prefixed with
442 the string "ro.", indicating that they must not be decoded without
443 checking that they are read-only. The "ro." prefix must not be stripped
444 off without performing this check. (Future versions of the web-API server
445 will perform it where necessary.)
447 If both the "rw_uri" and "ro_uri" fields are present in a given PROPDICT,
448 and the web-API server recognizes the rw_uri as a write cap, then it will
449 reset the ro_uri to the corresponding read cap and discard the original
450 contents of ro_uri (in order to ensure that the two caps correspond to the
451 same object and that the ro_uri is in fact read-only). However this may not
452 happen for caps in a format unknown to the web-API server. Therefore, when
453 writing a directory the web-API client should ensure that the contents
454 of "rw_uri" and "ro_uri" for a given PROPDICT are a consistent
455 (write cap, read cap) pair if possible. If the web-API client only has
456 one cap and does not know whether it is a write cap or read cap, then
457 it is acceptable to set "rw_uri" to that cap and omit "ro_uri". The
458 client must not put a write cap into a "ro_uri" field.
460 The metadata may have a "no-write" field. If this is set to true in the
461 metadata of a link, it will not be possible to open that link for writing
462 via the SFTP frontend; see `<FTP-and-SFTP.rst>`_ for details.
463 Also, if the "no-write" field is set to true in the metadata of a link to
464 a mutable child, it will cause the link to be diminished to read-only.
466 Note that the web-API-using client application must not provide the
467 "Content-Type: multipart/form-data" header that usually accompanies HTML
468 form submissions, since the body is not formatted this way. Doing so will
469 cause a server error as the lower-level code misparses the request body.
471 Child file names should each be expressed as a Unicode string, then used as
472 keys of the dictionary. The dictionary should then be converted into JSON,
473 and the resulting string encoded into UTF-8. This UTF-8 bytestring should
474 then be used as the POST body.
476 ``POST /uri?t=mkdir-immutable``
478 Like t=mkdir-with-children above, but the new directory will be
479 deep-immutable. This means that the directory itself is immutable, and that
480 it can only contain objects that are treated as being deep-immutable, like
481 immutable files, literal files, and deep-immutable directories.
483 For forward-compatibility, a deep-immutable directory can also contain caps
484 in a format that is unknown to the web-API server. When such caps are retrieved
485 from a deep-immutable directory in a "ro_uri" field, they will be prefixed
486 with the string "imm.", indicating that they must not be decoded without
487 checking that they are immutable. The "imm." prefix must not be stripped
488 off without performing this check. (Future versions of the web-API server
489 will perform it where necessary.)
491 The cap for each child may be given either in the "rw_uri" or "ro_uri"
492 field of the PROPDICT (not both). If a cap is given in the "rw_uri" field,
493 then the web-API server will check that it is an immutable read-cap of a
494 *known* format, and give an error if it is not. If a cap is given in the
495 "ro_uri" field, then the web-API server will still check whether known
496 caps are immutable, but for unknown caps it will simply assume that the
497 cap can be stored, as described above. Note that an attacker would be
498 able to store any cap in an immutable directory, so this check when
499 creating the directory is only to help non-malicious clients to avoid
500 accidentally giving away more authority than intended.
502 A non-empty request body is mandatory, since after the directory is created,
503 it will not be possible to add more children to it.
505 ``POST /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=mkdir``
507 ``PUT /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=mkdir``
509 Create new directories as necessary to make sure that the named target
510 ($DIRCAP/SUBDIRS../SUBDIR) is a directory. This will create additional
511 intermediate mutable directories as necessary. If the named target directory
512 already exists, this will make no changes to it.
514 If the final directory is created, it will be empty.
516 This operation will return an error if a blocking file is present at any of
517 the parent names, preventing the server from creating the necessary parent
518 directory; or if it would require changing an immutable directory.
520 The write-cap of the new directory will be returned as the HTTP response
523 ``POST /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=mkdir-with-children``
525 Like /uri?t=mkdir-with-children, but the final directory is created as a
526 child of an existing mutable directory. This will create additional
527 intermediate mutable directories as necessary. If the final directory is
528 created, it will be populated with initial children from the POST request
529 body, as described above.
531 This operation will return an error if a blocking file is present at any of
532 the parent names, preventing the server from creating the necessary parent
533 directory; or if it would require changing an immutable directory; or if
534 the immediate parent directory already has a a child named SUBDIR.
536 ``POST /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=mkdir-immutable``
538 Like /uri?t=mkdir-immutable, but the final directory is created as a child
539 of an existing mutable directory. The final directory will be deep-immutable,
540 and will be populated with the children specified as a JSON dictionary in
541 the POST request body.
543 In Tahoe 1.6 this operation creates intermediate mutable directories if
544 necessary, but that behaviour should not be relied on; see ticket #920.
546 This operation will return an error if the parent directory is immutable,
547 or already has a child named SUBDIR.
549 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=mkdir&name=NAME``
551 Create a new empty mutable directory and attach it to the given existing
552 directory. This will create additional intermediate directories as necessary.
554 This operation will return an error if a blocking file is present at any of
555 the parent names, preventing the server from creating the necessary parent
556 directory, or if it would require changing any immutable directory.
558 The URL of this operation points to the parent of the bottommost new directory,
559 whereas the /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=mkdir operation above has a URL
560 that points directly to the bottommost new directory.
562 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=mkdir-with-children&name=NAME``
564 Like /uri/$DIRCAP/[SUBDIRS../]?t=mkdir&name=NAME, but the new directory will
565 be populated with initial children via the POST request body. This command
566 will create additional intermediate mutable directories as necessary.
568 This operation will return an error if a blocking file is present at any of
569 the parent names, preventing the server from creating the necessary parent
570 directory; or if it would require changing an immutable directory; or if
571 the immediate parent directory already has a a child named NAME.
573 Note that the name= argument must be passed as a queryarg, because the POST
574 request body is used for the initial children JSON.
576 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=mkdir-immutable&name=NAME``
578 Like /uri/$DIRCAP/[SUBDIRS../]?t=mkdir-with-children&name=NAME, but the
579 final directory will be deep-immutable. The children are specified as a
580 JSON dictionary in the POST request body. Again, the name= argument must be
581 passed as a queryarg.
583 In Tahoe 1.6 this operation creates intermediate mutable directories if
584 necessary, but that behaviour should not be relied on; see ticket #920.
586 This operation will return an error if the parent directory is immutable,
587 or already has a child named NAME.
589 Get Information About A File Or Directory (as JSON)
590 ---------------------------------------------------
592 ``GET /uri/$FILECAP?t=json``
594 ``GET /uri/$DIRCAP?t=json``
596 ``GET /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=json``
598 ``GET /uri/$DIRCAP/[SUBDIRS../]FILENAME?t=json``
600 This returns a machine-parseable JSON-encoded description of the given
601 object. The JSON always contains a list, and the first element of the list is
602 always a flag that indicates whether the referenced object is a file or a
603 directory. If it is a capability to a file, then the information includes
604 file size and URI, like this::
606 GET /uri/$FILECAP?t=json :
610 "verify_uri": verify_uri,
615 If it is a capability to a directory followed by a path from that directory
616 to a file, then the information also includes metadata from the link to the
617 file in the parent directory, like this::
619 GET /uri/$DIRCAP/[SUBDIRS../]FILENAME?t=json
623 "verify_uri": verify_uri,
627 "ctime": 1202777696.7564139,
628 "mtime": 1202777696.7564139,
630 "linkcrtime": 1202777696.7564139,
631 "linkmotime": 1202777696.7564139
634 If it is a directory, then it includes information about the children of
635 this directory, as a mapping from child name to a set of data about the
636 child (the same data that would appear in a corresponding GET?t=json of the
637 child itself). The child entries also include metadata about each child,
638 including link-creation- and link-change- timestamps. The output looks like
641 GET /uri/$DIRCAP?t=json :
642 GET /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=json :
645 "rw_uri": read_write_uri,
646 "ro_uri": read_only_uri,
647 "verify_uri": verify_uri,
650 "foo.txt": [ "filenode", {
654 "ctime": 1202777696.7564139,
655 "mtime": 1202777696.7564139,
657 "linkcrtime": 1202777696.7564139,
658 "linkmotime": 1202777696.7564139
660 "subdir": [ "dirnode", {
664 "ctime": 1202778102.7589991,
665 "mtime": 1202778111.2160511,
667 "linkcrtime": 1202777696.7564139,
668 "linkmotime": 1202777696.7564139
672 In the above example, note how 'children' is a dictionary in which the keys
673 are child names and the values depend upon whether the child is a file or a
674 directory. The value is mostly the same as the JSON representation of the
675 child object (except that directories do not recurse -- the "children"
676 entry of the child is omitted, and the directory view includes the metadata
677 that is stored on the directory edge).
679 The rw_uri field will be present in the information about a directory
680 if and only if you have read-write access to that directory. The verify_uri
681 field will be present if and only if the object has a verify-cap
682 (non-distributed LIT files do not have verify-caps).
684 If the cap is of an unknown format, then the file size and verify_uri will
687 GET /uri/$UNKNOWNCAP?t=json :
690 "ro_uri": unknown_read_uri
693 GET /uri/$DIRCAP/[SUBDIRS../]UNKNOWNCHILDNAME?t=json :
696 "rw_uri": unknown_write_uri,
697 "ro_uri": unknown_read_uri,
700 "ctime": 1202777696.7564139,
701 "mtime": 1202777696.7564139,
703 "linkcrtime": 1202777696.7564139,
704 "linkmotime": 1202777696.7564139
707 As in the case of file nodes, the metadata will only be present when the
708 capability is to a directory followed by a path. The "mutable" field is also
709 not always present; when it is absent, the mutability of the object is not
715 The value of the 'tahoe':'linkmotime' key is updated whenever a link to a
716 child is set. The value of the 'tahoe':'linkcrtime' key is updated whenever
717 a link to a child is created -- i.e. when there was not previously a link
720 Note however, that if the edge in the Tahoe filesystem points to a mutable
721 file and the contents of that mutable file is changed, then the
722 'tahoe':'linkmotime' value on that edge will *not* be updated, since the
723 edge itself wasn't updated -- only the mutable file was.
725 The timestamps are represented as a number of seconds since the UNIX epoch
726 (1970-01-01 00:00:00 UTC), with leap seconds not being counted in the long
729 In Tahoe earlier than v1.4.0, 'mtime' and 'ctime' keys were populated
730 instead of the 'tahoe':'linkmotime' and 'tahoe':'linkcrtime' keys. Starting
731 in Tahoe v1.4.0, the 'linkmotime'/'linkcrtime' keys in the 'tahoe' sub-dict
732 are populated. However, prior to Tahoe v1.7beta, a bug caused the 'tahoe'
733 sub-dict to be deleted by web-API requests in which new metadata is
734 specified, and not to be added to existing child links that lack it.
736 From Tahoe v1.7.0 onward, the 'mtime' and 'ctime' fields are no longer
737 populated or updated (see ticket #924), except by "tahoe backup" as
738 explained below. For backward compatibility, when an existing link is
739 updated and 'tahoe':'linkcrtime' is not present in the previous metadata
740 but 'ctime' is, the old value of 'ctime' is used as the new value of
741 'tahoe':'linkcrtime'.
743 The reason we added the new fields in Tahoe v1.4.0 is that there is a
744 "set_children" API (described below) which you can use to overwrite the
745 values of the 'mtime'/'ctime' pair, and this API is used by the
746 "tahoe backup" command (in Tahoe v1.3.0 and later) to set the 'mtime' and
747 'ctime' values when backing up files from a local filesystem into the
748 Tahoe filesystem. As of Tahoe v1.4.0, the set_children API cannot be used
749 to set anything under the 'tahoe' key of the metadata dict -- if you
750 include 'tahoe' keys in your 'metadata' arguments then it will silently
753 Therefore, if the 'tahoe' sub-dict is present, you can rely on the
754 'linkcrtime' and 'linkmotime' values therein to have the semantics described
755 above. (This is assuming that only official Tahoe clients have been used to
756 write those links, and that their system clocks were set to what you expected
757 -- there is nothing preventing someone from editing their Tahoe client or
758 writing their own Tahoe client which would overwrite those values however
759 they like, and there is nothing to constrain their system clock from taking
762 When an edge is created or updated by "tahoe backup", the 'mtime' and
763 'ctime' keys on that edge are set as follows:
765 * 'mtime' is set to the timestamp read from the local filesystem for the
766 "mtime" of the local file in question, which means the last time the
767 contents of that file were changed.
769 * On Windows, 'ctime' is set to the creation timestamp for the file
770 read from the local filesystem. On other platforms, 'ctime' is set to
771 the UNIX "ctime" of the local file, which means the last time that
772 either the contents or the metadata of the local file was changed.
774 There are several ways that the 'ctime' field could be confusing:
776 1. You might be confused about whether it reflects the time of the creation
777 of a link in the Tahoe filesystem (by a version of Tahoe < v1.7.0) or a
778 timestamp copied in by "tahoe backup" from a local filesystem.
780 2. You might be confused about whether it is a copy of the file creation
781 time (if "tahoe backup" was run on a Windows system) or of the last
782 contents-or-metadata change (if "tahoe backup" was run on a different
785 3. You might be confused by the fact that changing the contents of a
786 mutable file in Tahoe doesn't have any effect on any links pointing at
787 that file in any directories, although "tahoe backup" sets the link
788 'ctime'/'mtime' to reflect timestamps about the local file corresponding
789 to the Tahoe file to which the link points.
791 4. Also, quite apart from Tahoe, you might be confused about the meaning
792 of the "ctime" in UNIX local filesystems, which people sometimes think
793 means file creation time, but which actually means, in UNIX local
794 filesystems, the most recent time that the file contents or the file
795 metadata (such as owner, permission bits, extended attributes, etc.)
796 has changed. Note that although "ctime" does not mean file creation time
797 in UNIX, links created by a version of Tahoe prior to v1.7.0, and never
798 written by "tahoe backup", will have 'ctime' set to the link creation
802 Attaching an existing File or Directory by its read- or write-cap
803 -----------------------------------------------------------------
805 ``PUT /uri/$DIRCAP/[SUBDIRS../]CHILDNAME?t=uri``
807 This attaches a child object (either a file or directory) to a specified
808 location in the virtual filesystem. The child object is referenced by its
809 read- or write- cap, as provided in the HTTP request body. This will create
810 intermediate directories as necessary.
812 This is similar to a UNIX hardlink: by referencing a previously-uploaded file
813 (or previously-created directory) instead of uploading/creating a new one,
814 you can create two references to the same object.
816 The read- or write- cap of the child is provided in the body of the HTTP
817 request, and this same cap is returned in the response body.
819 The default behavior is to overwrite any existing object at the same
820 location. To prevent this (and make the operation return an error instead
821 of overwriting), add a "replace=false" argument, as "?t=uri&replace=false".
822 With replace=false, this operation will return an HTTP 409 "Conflict" error
823 if there is already an object at the given location, rather than
824 overwriting the existing object. To allow the operation to overwrite a
825 file, but return an error when trying to overwrite a directory, use
826 "replace=only-files" (this behavior is closer to the traditional UNIX "mv"
827 command). Note that "true", "t", and "1" are all synonyms for "True", and
828 "false", "f", and "0" are synonyms for "False", and the parameter is
831 Note that this operation does not take its child cap in the form of
832 separate "rw_uri" and "ro_uri" fields. Therefore, it cannot accept a
833 child cap in a format unknown to the web-API server, unless its URI
834 starts with "ro." or "imm.". This restriction is necessary because the
835 server is not able to attenuate an unknown write cap to a read cap.
836 Unknown URIs starting with "ro." or "imm.", on the other hand, are
837 assumed to represent read caps. The client should not prefix a write
838 cap with "ro." or "imm." and pass it to this operation, since that
839 would result in granting the cap's write authority to holders of the
842 Adding multiple files or directories to a parent directory at once
843 ------------------------------------------------------------------
845 ``POST /uri/$DIRCAP/[SUBDIRS..]?t=set_children``
847 ``POST /uri/$DIRCAP/[SUBDIRS..]?t=set-children`` (Tahoe >= v1.6)
849 This command adds multiple children to a directory in a single operation.
850 It reads the request body and interprets it as a JSON-encoded description
851 of the child names and read/write-caps that should be added.
853 The body should be a JSON-encoded dictionary, in the same format as the
854 "children" value returned by the "GET /uri/$DIRCAP?t=json" operation
855 described above. In this format, each key is a child names, and the
856 corresponding value is a tuple of (type, childinfo). "type" is ignored, and
857 "childinfo" is a dictionary that contains "rw_uri", "ro_uri", and
858 "metadata" keys. You can take the output of "GET /uri/$DIRCAP1?t=json" and
859 use it as the input to "POST /uri/$DIRCAP2?t=set_children" to make DIR2
860 look very much like DIR1 (except for any existing children of DIR2 that
861 were not overwritten, and any existing "tahoe" metadata keys as described
864 When the set_children request contains a child name that already exists in
865 the target directory, this command defaults to overwriting that child with
866 the new value (both child cap and metadata, but if the JSON data does not
867 contain a "metadata" key, the old child's metadata is preserved). The
868 command takes a boolean "overwrite=" query argument to control this
869 behavior. If you use "?t=set_children&overwrite=false", then an attempt to
870 replace an existing child will instead cause an error.
872 Any "tahoe" key in the new child's "metadata" value is ignored. Any
873 existing "tahoe" metadata is preserved. The metadata["tahoe"] value is
874 reserved for metadata generated by the tahoe node itself. The only two keys
875 currently placed here are "linkcrtime" and "linkmotime". For details, see
876 the section above entitled "Get Information About A File Or Directory (as
877 JSON)", in the "About the metadata" subsection.
879 Note that this command was introduced with the name "set_children", which
880 uses an underscore rather than a hyphen as other multi-word command names
881 do. The variant with a hyphen is now accepted, but clients that desire
882 backward compatibility should continue to use "set_children".
885 Deleting a File or Directory
886 ----------------------------
888 ``DELETE /uri/$DIRCAP/[SUBDIRS../]CHILDNAME``
890 This removes the given name from its parent directory. CHILDNAME is the
891 name to be removed, and $DIRCAP/SUBDIRS.. indicates the directory that will
894 Note that this does not actually delete the file or directory that the name
895 points to from the tahoe grid -- it only removes the named reference from
896 this directory. If there are other names in this directory or in other
897 directories that point to the resource, then it will remain accessible
898 through those paths. Even if all names pointing to this object are removed
899 from their parent directories, then someone with possession of its read-cap
900 can continue to access the object through that cap.
902 The object will only become completely unreachable once 1: there are no
903 reachable directories that reference it, and 2: nobody is holding a read-
904 or write- cap to the object. (This behavior is very similar to the way
905 hardlinks and anonymous files work in traditional UNIX filesystems).
907 This operation will not modify more than a single directory. Intermediate
908 directories which were implicitly created by PUT or POST methods will *not*
909 be automatically removed by DELETE.
911 This method returns the file- or directory- cap of the object that was just
914 Browser Operations: Human-oriented interfaces
915 =============================================
917 This section describes the HTTP operations that provide support for humans
918 running a web browser. Most of these operations use HTML forms that use POST
919 to drive the Tahoe node. This section is intended for HTML authors who want
920 to write web pages that contain forms and buttons which manipulate the Tahoe
923 Note that for all POST operations, the arguments listed can be provided
924 either as URL query arguments or as form body fields. URL query arguments are
925 separated from the main URL by "?", and from each other by "&". For example,
926 "POST /uri/$DIRCAP?t=upload&mutable=true". Form body fields are usually
927 specified by using <input type="hidden"> elements. For clarity, the
928 descriptions below display the most significant arguments as URL query args.
930 Viewing A Directory (as HTML)
931 -----------------------------
933 ``GET /uri/$DIRCAP/[SUBDIRS../]``
935 This returns an HTML page, intended to be displayed to a human by a web
936 browser, which contains HREF links to all files and directories reachable
937 from this directory. These HREF links do not have a t= argument, meaning
938 that a human who follows them will get pages also meant for a human. It also
939 contains forms to upload new files, and to unlink files and directories
940 from their parent directory. Those forms use POST methods to do their job.
942 Viewing/Downloading a File
943 --------------------------
945 ``GET /uri/$FILECAP``
947 ``GET /uri/$DIRCAP/[SUBDIRS../]FILENAME``
949 This will retrieve the contents of the given file. The HTTP response body
950 will contain the sequence of bytes that make up the file.
952 If you want the HTTP response to include a useful Content-Type header,
953 either use the second form (which starts with a $DIRCAP), or add a
954 "filename=foo" query argument, like "GET /uri/$FILECAP?filename=foo.jpg".
955 The bare "GET /uri/$FILECAP" does not give the Tahoe node enough information
956 to determine a Content-Type (since Tahoe immutable files are merely
957 sequences of bytes, not typed+named file objects).
959 If the URL has both filename= and "save=true" in the query arguments, then
960 the server to add a "Content-Disposition: attachment" header, along with a
961 filename= parameter. When a user clicks on such a link, most browsers will
962 offer to let the user save the file instead of displaying it inline (indeed,
963 most browsers will refuse to display it inline). "true", "t", "1", and other
964 case-insensitive equivalents are all treated the same.
966 Character-set handling in URLs and HTTP headers is a dubious art [1]_. For
967 maximum compatibility, Tahoe simply copies the bytes from the filename=
968 argument into the Content-Disposition header's filename= parameter, without
969 trying to interpret them in any particular way.
972 ``GET /named/$FILECAP/FILENAME``
974 This is an alternate download form which makes it easier to get the correct
975 filename. The Tahoe server will provide the contents of the given file, with
976 a Content-Type header derived from the given filename. This form is used to
977 get browsers to use the "Save Link As" feature correctly, and also helps
978 command-line tools like "wget" and "curl" use the right filename. Note that
979 this form can *only* be used with file caps; it is an error to use a
980 directory cap after the /named/ prefix.
982 Get Information About A File Or Directory (as HTML)
983 ---------------------------------------------------
985 ``GET /uri/$FILECAP?t=info``
987 ``GET /uri/$DIRCAP/?t=info``
989 ``GET /uri/$DIRCAP/[SUBDIRS../]SUBDIR/?t=info``
991 ``GET /uri/$DIRCAP/[SUBDIRS../]FILENAME?t=info``
993 This returns a human-oriented HTML page with more detail about the selected
994 file or directory object. This page contains the following items:
998 * JSON representation
999 * raw contents (text/plain)
1000 * access caps (URIs): verify-cap, read-cap, write-cap (for mutable objects)
1001 * check/verify/repair form
1002 * deep-check/deep-size/deep-stats/manifest (for directories)
1003 * replace-conents form (for mutable files)
1005 Creating a Directory
1006 --------------------
1008 ``POST /uri?t=mkdir``
1010 This creates a new empty directory, but does not attach it to the virtual
1013 If a "redirect_to_result=true" argument is provided, then the HTTP response
1014 will cause the web browser to be redirected to a /uri/$DIRCAP page that
1015 gives access to the newly-created directory. If you bookmark this page,
1016 you'll be able to get back to the directory again in the future. This is the
1017 recommended way to start working with a Tahoe server: create a new unlinked
1018 directory (using redirect_to_result=true), then bookmark the resulting
1019 /uri/$DIRCAP page. There is a "create directory" button on the Welcome page
1020 to invoke this action.
1022 If "redirect_to_result=true" is not provided (or is given a value of
1023 "false"), then the HTTP response body will simply be the write-cap of the
1026 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=mkdir&name=CHILDNAME``
1028 This creates a new empty directory as a child of the designated SUBDIR. This
1029 will create additional intermediate directories as necessary.
1031 If a "when_done=URL" argument is provided, the HTTP response will cause the
1032 web browser to redirect to the given URL. This provides a convenient way to
1033 return the browser to the directory that was just modified. Without a
1034 when_done= argument, the HTTP response will simply contain the write-cap of
1035 the directory that was just created.
1041 ``POST /uri?t=upload``
1043 This uploads a file, and produces a file-cap for the contents, but does not
1044 attach the file into the filesystem. No directories will be modified by
1047 The file must be provided as the "file" field of an HTML encoded form body,
1048 produced in response to an HTML form like this::
1050 <form action="/uri" method="POST" enctype="multipart/form-data">
1051 <input type="hidden" name="t" value="upload" />
1052 <input type="file" name="file" />
1053 <input type="submit" value="Upload Unlinked" />
1056 If a "when_done=URL" argument is provided, the response body will cause the
1057 browser to redirect to the given URL. If the when_done= URL has the string
1058 "%(uri)s" in it, that string will be replaced by a URL-escaped form of the
1059 newly created file-cap. (Note that without this substitution, there is no
1060 way to access the file that was just uploaded).
1062 The default (in the absence of when_done=) is to return an HTML page that
1063 describes the results of the upload. This page will contain information
1064 about which storage servers were used for the upload, how long each
1065 operation took, etc.
1067 If a "mutable=true" argument is provided, the operation will create a
1068 mutable file, and the response body will contain the write-cap instead of
1069 the upload results page. The default is to create an immutable file,
1070 returning the upload results page as a response.
1073 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=upload``
1075 This uploads a file, and attaches it as a new child of the given directory,
1076 which must be mutable. The file must be provided as the "file" field of an
1077 HTML-encoded form body, produced in response to an HTML form like this::
1079 <form action="." method="POST" enctype="multipart/form-data">
1080 <input type="hidden" name="t" value="upload" />
1081 <input type="file" name="file" />
1082 <input type="submit" value="Upload" />
1085 A "name=" argument can be provided to specify the new child's name,
1086 otherwise it will be taken from the "filename" field of the upload form
1087 (most web browsers will copy the last component of the original file's
1088 pathname into this field). To avoid confusion, name= is not allowed to
1091 If there is already a child with that name, and it is a mutable file, then
1092 its contents are replaced with the data being uploaded. If it is not a
1093 mutable file, the default behavior is to remove the existing child before
1094 creating a new one. To prevent this (and make the operation return an error
1095 instead of overwriting the old child), add a "replace=false" argument, as
1096 "?t=upload&replace=false". With replace=false, this operation will return an
1097 HTTP 409 "Conflict" error if there is already an object at the given
1098 location, rather than overwriting the existing object. Note that "true",
1099 "t", and "1" are all synonyms for "True", and "false", "f", and "0" are
1100 synonyms for "False". the parameter is case-insensitive.
1102 This will create additional intermediate directories as necessary, although
1103 since it is expected to be triggered by a form that was retrieved by "GET
1104 /uri/$DIRCAP/[SUBDIRS../]", it is likely that the parent directory will
1107 If a "mutable=true" argument is provided, any new file that is created will
1108 be a mutable file instead of an immutable one. <input type="checkbox"
1109 name="mutable" /> will give the user a way to set this option.
1111 If a "when_done=URL" argument is provided, the HTTP response will cause the
1112 web browser to redirect to the given URL. This provides a convenient way to
1113 return the browser to the directory that was just modified. Without a
1114 when_done= argument, the HTTP response will simply contain the file-cap of
1115 the file that was just uploaded (a write-cap for mutable files, or a
1116 read-cap for immutable files).
1118 ``POST /uri/$DIRCAP/[SUBDIRS../]FILENAME?t=upload``
1120 This also uploads a file and attaches it as a new child of the given
1121 directory, which must be mutable. It is a slight variant of the previous
1122 operation, as the URL refers to the target file rather than the parent
1123 directory. It is otherwise identical: this accepts mutable= and when_done=
1126 ``POST /uri/$FILECAP?t=upload``
1128 This modifies the contents of an existing mutable file in-place. An error is
1129 signalled if $FILECAP does not refer to a mutable file. It behaves just like
1130 the "PUT /uri/$FILECAP" form, but uses a POST for the benefit of HTML forms
1133 Attaching An Existing File Or Directory (by URI)
1134 ------------------------------------------------
1136 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=uri&name=CHILDNAME&uri=CHILDCAP``
1138 This attaches a given read- or write- cap "CHILDCAP" to the designated
1139 directory, with a specified child name. This behaves much like the PUT t=uri
1140 operation, and is a lot like a UNIX hardlink. It is subject to the same
1141 restrictions as that operation on the use of cap formats unknown to the
1144 This will create additional intermediate directories as necessary, although
1145 since it is expected to be triggered by a form that was retrieved by "GET
1146 /uri/$DIRCAP/[SUBDIRS../]", it is likely that the parent directory will
1149 This accepts the same replace= argument as POST t=upload.
1154 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=delete&name=CHILDNAME``
1156 This instructs the node to remove a child object (file or subdirectory) from
1157 the given directory, which must be mutable. Note that the entire subtree is
1158 unlinked from the parent. Unlike deleting a subdirectory in a UNIX local
1159 filesystem, the subtree need not be empty; if it isn't, then other references
1160 into the subtree will see that the child subdirectories are not modified by
1161 this operation. Only the link from the given directory to its child is severed.
1163 In Tahoe-LAFS v1.9.0 and later, t=unlink can be used as a synonym for t=delete.
1164 If interoperability with older web-API servers is required, t=delete should
1171 ``POST /uri/$DIRCAP/[SUBDIRS../]?t=rename&from_name=OLD&to_name=NEW``
1173 This instructs the node to rename a child of the given directory, which must
1174 be mutable. This has a similar effect to removing the child, then adding the
1175 same child-cap under the new name, except that it preserves metadata. This
1176 operation cannot move the child to a different directory.
1178 This operation will replace any existing child of the new name, making it
1179 behave like the UNIX "``mv -f``" command.
1184 ``GET /uri?uri=$CAP``
1186 This causes a redirect to /uri/$CAP, and retains any additional query
1187 arguments (like filename= or save=). This is for the convenience of web
1188 forms which allow the user to paste in a read- or write- cap (obtained
1189 through some out-of-band channel, like IM or email).
1191 Note that this form merely redirects to the specific file or directory
1192 indicated by the $CAP: unlike the GET /uri/$DIRCAP form, you cannot
1193 traverse to children by appending additional path segments to the URL.
1195 ``GET /uri/$DIRCAP/[SUBDIRS../]?t=rename-form&name=$CHILDNAME``
1197 This provides a useful facility to browser-based user interfaces. It
1198 returns a page containing a form targetting the "POST $DIRCAP t=rename"
1199 functionality described above, with the provided $CHILDNAME present in the
1200 'from_name' field of that form. I.e. this presents a form offering to
1201 rename $CHILDNAME, requesting the new name, and submitting POST rename.
1203 ``GET /uri/$DIRCAP/[SUBDIRS../]CHILDNAME?t=uri``
1205 This returns the file- or directory- cap for the specified object.
1207 ``GET /uri/$DIRCAP/[SUBDIRS../]CHILDNAME?t=readonly-uri``
1209 This returns a read-only file- or directory- cap for the specified object.
1210 If the object is an immutable file, this will return the same value as
1213 Debugging and Testing Features
1214 ------------------------------
1216 These URLs are less-likely to be helpful to the casual Tahoe user, and are
1217 mainly intended for developers.
1219 ``POST $URL?t=check``
1221 This triggers the FileChecker to determine the current "health" of the
1222 given file or directory, by counting how many shares are available. The
1223 page that is returned will display the results. This can be used as a "show
1224 me detailed information about this file" page.
1226 If a verify=true argument is provided, the node will perform a more
1227 intensive check, downloading and verifying every single bit of every share.
1229 If an add-lease=true argument is provided, the node will also add (or
1230 renew) a lease to every share it encounters. Each lease will keep the share
1231 alive for a certain period of time (one month by default). Once the last
1232 lease expires or is explicitly cancelled, the storage server is allowed to
1235 If an output=JSON argument is provided, the response will be
1236 machine-readable JSON instead of human-oriented HTML. The data is a
1237 dictionary with the following keys::
1239 storage-index: a base32-encoded string with the objects's storage index,
1240 or an empty string for LIT files
1241 summary: a string, with a one-line summary of the stats of the file
1242 results: a dictionary that describes the state of the file. For LIT files,
1243 this dictionary has only the 'healthy' key, which will always be
1244 True. For distributed files, this dictionary has the following
1246 count-shares-good: the number of good shares that were found
1247 count-shares-needed: 'k', the number of shares required for recovery
1248 count-shares-expected: 'N', the number of total shares generated
1249 count-good-share-hosts: this was intended to be the number of distinct
1250 storage servers with good shares. It is currently
1251 (as of Tahoe-LAFS v1.8.0) computed incorrectly;
1253 count-wrong-shares: for mutable files, the number of shares for
1254 versions other than the 'best' one (highest
1255 sequence number, highest roothash). These are
1257 count-recoverable-versions: for mutable files, the number of
1258 recoverable versions of the file. For
1259 a healthy file, this will equal 1.
1260 count-unrecoverable-versions: for mutable files, the number of
1261 unrecoverable versions of the file.
1262 For a healthy file, this will be 0.
1263 count-corrupt-shares: the number of shares with integrity failures
1264 list-corrupt-shares: a list of "share locators", one for each share
1265 that was found to be corrupt. Each share locator
1266 is a list of (serverid, storage_index, sharenum).
1267 needs-rebalancing: (bool) True if there are multiple shares on a single
1268 storage server, indicating a reduction in reliability
1269 that could be resolved by moving shares to new
1271 servers-responding: list of base32-encoded storage server identifiers,
1272 one for each server which responded to the share
1274 healthy: (bool) True if the file is completely healthy, False otherwise.
1275 Healthy files have at least N good shares. Overlapping shares
1276 do not currently cause a file to be marked unhealthy. If there
1277 are at least N good shares, then corrupt shares do not cause the
1278 file to be marked unhealthy, although the corrupt shares will be
1279 listed in the results (list-corrupt-shares) and should be manually
1280 removed to wasting time in subsequent downloads (as the
1281 downloader rediscovers the corruption and uses alternate shares).
1282 Future compatibility: the meaning of this field may change to
1283 reflect whether the servers-of-happiness criterion is met
1285 sharemap: dict mapping share identifier to list of serverids
1286 (base32-encoded strings). This indicates which servers are
1287 holding which shares. For immutable files, the shareid is
1288 an integer (the share number, from 0 to N-1). For
1289 immutable files, it is a string of the form
1290 'seq%d-%s-sh%d', containing the sequence number, the
1291 roothash, and the share number.
1293 ``POST $URL?t=start-deep-check`` (must add &ophandle=XYZ)
1295 This initiates a recursive walk of all files and directories reachable from
1296 the target, performing a check on each one just like t=check. The result
1297 page will contain a summary of the results, including details on any
1298 file/directory that was not fully healthy.
1300 t=start-deep-check can only be invoked on a directory. An error (400
1301 BAD_REQUEST) will be signalled if it is invoked on a file. The recursive
1302 walker will deal with loops safely.
1304 This accepts the same verify= and add-lease= arguments as t=check.
1306 Since this operation can take a long time (perhaps a second per object),
1307 the ophandle= argument is required (see "Slow Operations, Progress, and
1308 Cancelling" above). The response to this POST will be a redirect to the
1309 corresponding /operations/$HANDLE page (with output=HTML or output=JSON to
1310 match the output= argument given to the POST). The deep-check operation
1311 will continue to run in the background, and the /operations page should be
1312 used to find out when the operation is done.
1314 Detailed check results for non-healthy files and directories will be
1315 available under /operations/$HANDLE/$STORAGEINDEX, and the HTML status will
1316 contain links to these detailed results.
1318 The HTML /operations/$HANDLE page for incomplete operations will contain a
1319 meta-refresh tag, set to 60 seconds, so that a browser which uses
1320 deep-check will automatically poll until the operation has completed.
1322 The JSON page (/options/$HANDLE?output=JSON) will contain a
1323 machine-readable JSON dictionary with the following keys::
1325 finished: a boolean, True if the operation is complete, else False. Some
1326 of the remaining keys may not be present until the operation
1328 root-storage-index: a base32-encoded string with the storage index of the
1329 starting point of the deep-check operation
1330 count-objects-checked: count of how many objects were checked. Note that
1331 non-distributed objects (i.e. small immutable LIT
1332 files) are not checked, since for these objects,
1333 the data is contained entirely in the URI.
1334 count-objects-healthy: how many of those objects were completely healthy
1335 count-objects-unhealthy: how many were damaged in some way
1336 count-corrupt-shares: how many shares were found to have corruption,
1337 summed over all objects examined
1338 list-corrupt-shares: a list of "share identifiers", one for each share
1339 that was found to be corrupt. Each share identifier
1340 is a list of (serverid, storage_index, sharenum).
1341 list-unhealthy-files: a list of (pathname, check-results) tuples, for
1342 each file that was not fully healthy. 'pathname' is
1343 a list of strings (which can be joined by "/"
1344 characters to turn it into a single string),
1345 relative to the directory on which deep-check was
1346 invoked. The 'check-results' field is the same as
1347 that returned by t=check&output=JSON, described
1349 stats: a dictionary with the same keys as the t=start-deep-stats command
1352 ``POST $URL?t=stream-deep-check``
1354 This initiates a recursive walk of all files and directories reachable from
1355 the target, performing a check on each one just like t=check. For each
1356 unique object (duplicates are skipped), a single line of JSON is emitted to
1357 the HTTP response channel (or an error indication, see below). When the walk
1358 is complete, a final line of JSON is emitted which contains the accumulated
1359 file-size/count "deep-stats" data.
1361 This command takes the same arguments as t=start-deep-check.
1363 A CLI tool can split the response stream on newlines into "response units",
1364 and parse each response unit as JSON. Each such parsed unit will be a
1365 dictionary, and will contain at least the "type" key: a string, one of
1366 "file", "directory", or "stats".
1368 For all units that have a type of "file" or "directory", the dictionary will
1369 contain the following keys::
1371 "path": a list of strings, with the path that is traversed to reach the
1373 "cap": a write-cap URI for the file or directory, if available, else a
1375 "verifycap": a verify-cap URI for the file or directory
1376 "repaircap": an URI for the weakest cap that can still be used to repair
1378 "storage-index": a base32 storage index for the object
1379 "check-results": a copy of the dictionary which would be returned by
1380 t=check&output=json, with three top-level keys:
1381 "storage-index", "summary", and "results", and a variety
1382 of counts and sharemaps in the "results" value.
1384 Note that non-distributed files (i.e. LIT files) will have values of None
1385 for verifycap, repaircap, and storage-index, since these files can neither
1386 be verified nor repaired, and are not stored on the storage servers.
1387 Likewise the check-results dictionary will be limited: an empty string for
1388 storage-index, and a results dictionary with only the "healthy" key.
1390 The last unit in the stream will have a type of "stats", and will contain
1391 the keys described in the "start-deep-stats" operation, below.
1393 If any errors occur during the traversal (specifically if a directory is
1394 unrecoverable, such that further traversal is not possible), an error
1395 indication is written to the response body, instead of the usual line of
1396 JSON. This error indication line will begin with the string "ERROR:" (in all
1397 caps), and contain a summary of the error on the rest of the line. The
1398 remaining lines of the response body will be a python exception. The client
1399 application should look for the ERROR: and stop processing JSON as soon as
1400 it is seen. Note that neither a file being unrecoverable nor a directory
1401 merely being unhealthy will cause traversal to stop. The line just before
1402 the ERROR: will describe the directory that was untraversable, since the
1403 unit is emitted to the HTTP response body before the child is traversed.
1406 ``POST $URL?t=check&repair=true``
1408 This performs a health check of the given file or directory, and if the
1409 checker determines that the object is not healthy (some shares are missing
1410 or corrupted), it will perform a "repair". During repair, any missing
1411 shares will be regenerated and uploaded to new servers.
1413 This accepts the same verify=true and add-lease= arguments as t=check. When
1414 an output=JSON argument is provided, the machine-readable JSON response
1415 will contain the following keys::
1417 storage-index: a base32-encoded string with the objects's storage index,
1418 or an empty string for LIT files
1419 repair-attempted: (bool) True if repair was attempted
1420 repair-successful: (bool) True if repair was attempted and the file was
1421 fully healthy afterwards. False if no repair was
1422 attempted, or if a repair attempt failed.
1423 pre-repair-results: a dictionary that describes the state of the file
1424 before any repair was performed. This contains exactly
1425 the same keys as the 'results' value of the t=check
1426 response, described above.
1427 post-repair-results: a dictionary that describes the state of the file
1428 after any repair was performed. If no repair was
1429 performed, post-repair-results and pre-repair-results
1430 will be the same. This contains exactly the same keys
1431 as the 'results' value of the t=check response,
1434 ``POST $URL?t=start-deep-check&repair=true`` (must add &ophandle=XYZ)
1436 This triggers a recursive walk of all files and directories, performing a
1437 t=check&repair=true on each one.
1439 Like t=start-deep-check without the repair= argument, this can only be
1440 invoked on a directory. An error (400 BAD_REQUEST) will be signalled if it
1441 is invoked on a file. The recursive walker will deal with loops safely.
1443 This accepts the same verify= and add-lease= arguments as
1444 t=start-deep-check. It uses the same ophandle= mechanism as
1445 start-deep-check. When an output=JSON argument is provided, the response
1446 will contain the following keys::
1448 finished: (bool) True if the operation has completed, else False
1449 root-storage-index: a base32-encoded string with the storage index of the
1450 starting point of the deep-check operation
1451 count-objects-checked: count of how many objects were checked
1453 count-objects-healthy-pre-repair: how many of those objects were completely
1454 healthy, before any repair
1455 count-objects-unhealthy-pre-repair: how many were damaged in some way
1456 count-objects-healthy-post-repair: how many of those objects were completely
1457 healthy, after any repair
1458 count-objects-unhealthy-post-repair: how many were damaged in some way
1460 count-repairs-attempted: repairs were attempted on this many objects.
1461 count-repairs-successful: how many repairs resulted in healthy objects
1462 count-repairs-unsuccessful: how many repairs resulted did not results in
1463 completely healthy objects
1464 count-corrupt-shares-pre-repair: how many shares were found to have
1465 corruption, summed over all objects
1466 examined, before any repair
1467 count-corrupt-shares-post-repair: how many shares were found to have
1468 corruption, summed over all objects
1469 examined, after any repair
1470 list-corrupt-shares: a list of "share identifiers", one for each share
1471 that was found to be corrupt (before any repair).
1472 Each share identifier is a list of (serverid,
1473 storage_index, sharenum).
1474 list-remaining-corrupt-shares: like list-corrupt-shares, but mutable shares
1475 that were successfully repaired are not
1476 included. These are shares that need
1477 manual processing. Since immutable shares
1478 cannot be modified by clients, all corruption
1479 in immutable shares will be listed here.
1480 list-unhealthy-files: a list of (pathname, check-results) tuples, for
1481 each file that was not fully healthy. 'pathname' is
1482 relative to the directory on which deep-check was
1483 invoked. The 'check-results' field is the same as
1484 that returned by t=check&repair=true&output=JSON,
1486 stats: a dictionary with the same keys as the t=start-deep-stats command
1489 ``POST $URL?t=stream-deep-check&repair=true``
1491 This triggers a recursive walk of all files and directories, performing a
1492 t=check&repair=true on each one. For each unique object (duplicates are
1493 skipped), a single line of JSON is emitted to the HTTP response channel (or
1494 an error indication). When the walk is complete, a final line of JSON is
1495 emitted which contains the accumulated file-size/count "deep-stats" data.
1497 This emits the same data as t=stream-deep-check (without the repair=true),
1498 except that the "check-results" field is replaced with a
1499 "check-and-repair-results" field, which contains the keys returned by
1500 t=check&repair=true&output=json (i.e. repair-attempted, repair-successful,
1501 pre-repair-results, and post-repair-results). The output does not contain
1502 the summary dictionary that is provied by t=start-deep-check&repair=true
1503 (the one with count-objects-checked and list-unhealthy-files), since the
1504 receiving client is expected to calculate those values itself from the
1505 stream of per-object check-and-repair-results.
1507 Note that the "ERROR:" indication will only be emitted if traversal stops,
1508 which will only occur if an unrecoverable directory is encountered. If a
1509 file or directory repair fails, the traversal will continue, and the repair
1510 failure will be indicated in the JSON data (in the "repair-successful" key).
1512 ``POST $DIRURL?t=start-manifest`` (must add &ophandle=XYZ)
1514 This operation generates a "manfest" of the given directory tree, mostly
1515 for debugging. This is a table of (path, filecap/dircap), for every object
1516 reachable from the starting directory. The path will be slash-joined, and
1517 the filecap/dircap will contain a link to the object in question. This page
1518 gives immediate access to every object in the virtual filesystem subtree.
1520 This operation uses the same ophandle= mechanism as deep-check. The
1521 corresponding /operations/$HANDLE page has three different forms. The
1522 default is output=HTML.
1524 If output=text is added to the query args, the results will be a text/plain
1525 list. The first line is special: it is either "finished: yes" or "finished:
1526 no"; if the operation is not finished, you must periodically reload the
1527 page until it completes. The rest of the results are a plaintext list, with
1528 one file/dir per line, slash-separated, with the filecap/dircap separated
1531 If output=JSON is added to the queryargs, then the results will be a
1532 JSON-formatted dictionary with six keys. Note that because large directory
1533 structures can result in very large JSON results, the full results will not
1534 be available until the operation is complete (i.e. until output["finished"]
1537 finished (bool): if False then you must reload the page until True
1538 origin_si (base32 str): the storage index of the starting point
1539 manifest: list of (path, cap) tuples, where path is a list of strings.
1540 verifycaps: list of (printable) verify cap strings
1541 storage-index: list of (base32) storage index strings
1542 stats: a dictionary with the same keys as the t=start-deep-stats command
1545 ``POST $DIRURL?t=start-deep-size`` (must add &ophandle=XYZ)
1547 This operation generates a number (in bytes) containing the sum of the
1548 filesize of all directories and immutable files reachable from the given
1549 directory. This is a rough lower bound of the total space consumed by this
1550 subtree. It does not include space consumed by mutable files, nor does it
1551 take expansion or encoding overhead into account. Later versions of the
1552 code may improve this estimate upwards.
1554 The /operations/$HANDLE status output consists of two lines of text::
1559 ``POST $DIRURL?t=start-deep-stats`` (must add &ophandle=XYZ)
1561 This operation performs a recursive walk of all files and directories
1562 reachable from the given directory, and generates a collection of
1563 statistics about those objects.
1565 The result (obtained from the /operations/$OPHANDLE page) is a
1566 JSON-serialized dictionary with the following keys (note that some of these
1567 keys may be missing until 'finished' is True)::
1569 finished: (bool) True if the operation has finished, else False
1570 count-immutable-files: count of how many CHK files are in the set
1571 count-mutable-files: same, for mutable files (does not include directories)
1572 count-literal-files: same, for LIT files (data contained inside the URI)
1573 count-files: sum of the above three
1574 count-directories: count of directories
1575 count-unknown: count of unrecognized objects (perhaps from the future)
1576 size-immutable-files: total bytes for all CHK files in the set, =deep-size
1577 size-mutable-files (TODO): same, for current version of all mutable files
1578 size-literal-files: same, for LIT files
1579 size-directories: size of directories (includes size-literal-files)
1580 size-files-histogram: list of (minsize, maxsize, count) buckets,
1581 with a histogram of filesizes, 5dB/bucket,
1582 for both literal and immutable files
1583 largest-directory: number of children in the largest directory
1584 largest-immutable-file: number of bytes in the largest CHK file
1586 size-mutable-files is not implemented, because it would require extra
1587 queries to each mutable file to get their size. This may be implemented in
1590 Assuming no sharing, the basic space consumed by a single root directory is
1591 the sum of size-immutable-files, size-mutable-files, and size-directories.
1592 The actual disk space used by the shares is larger, because of the
1593 following sources of overhead::
1596 expansion due to erasure coding
1597 share management data (leases)
1598 backend (ext3) minimum block size
1600 ``POST $URL?t=stream-manifest``
1602 This operation performs a recursive walk of all files and directories
1603 reachable from the given starting point. For each such unique object
1604 (duplicates are skipped), a single line of JSON is emitted to the HTTP
1605 response channel (or an error indication, see below). When the walk is
1606 complete, a final line of JSON is emitted which contains the accumulated
1607 file-size/count "deep-stats" data.
1609 A CLI tool can split the response stream on newlines into "response units",
1610 and parse each response unit as JSON. Each such parsed unit will be a
1611 dictionary, and will contain at least the "type" key: a string, one of
1612 "file", "directory", or "stats".
1614 For all units that have a type of "file" or "directory", the dictionary will
1615 contain the following keys::
1617 "path": a list of strings, with the path that is traversed to reach the
1619 "cap": a write-cap URI for the file or directory, if available, else a
1621 "verifycap": a verify-cap URI for the file or directory
1622 "repaircap": an URI for the weakest cap that can still be used to repair
1624 "storage-index": a base32 storage index for the object
1626 Note that non-distributed files (i.e. LIT files) will have values of None
1627 for verifycap, repaircap, and storage-index, since these files can neither
1628 be verified nor repaired, and are not stored on the storage servers.
1630 The last unit in the stream will have a type of "stats", and will contain
1631 the keys described in the "start-deep-stats" operation, below.
1633 If any errors occur during the traversal (specifically if a directory is
1634 unrecoverable, such that further traversal is not possible), an error
1635 indication is written to the response body, instead of the usual line of
1636 JSON. This error indication line will begin with the string "ERROR:" (in all
1637 caps), and contain a summary of the error on the rest of the line. The
1638 remaining lines of the response body will be a python exception. The client
1639 application should look for the ERROR: and stop processing JSON as soon as
1640 it is seen. The line just before the ERROR: will describe the directory that
1641 was untraversable, since the manifest entry is emitted to the HTTP response
1642 body before the child is traversed.
1647 The portion of the web namespace that begins with "/uri" (and "/named") is
1648 dedicated to giving users (both humans and programs) access to the Tahoe
1649 virtual filesystem. The rest of the namespace provides status information
1650 about the state of the Tahoe node.
1652 ``GET /`` (the root page)
1654 This is the "Welcome Page", and contains a few distinct sections::
1656 Node information: library versions, local nodeid, services being provided.
1658 Filesystem Access Forms: create a new directory, view a file/directory by
1659 URI, upload a file (unlinked), download a file by
1662 Grid Status: introducer information, helper information, connected storage
1667 This page lists all active uploads and downloads, and contains a short list
1668 of recent upload/download operations. Each operation has a link to a page
1669 that describes file sizes, servers that were involved, and the time consumed
1670 in each phase of the operation.
1672 A GET of /status/?t=json will contain a machine-readable subset of the same
1673 data. It returns a JSON-encoded dictionary. The only key defined at this
1674 time is "active", with a value that is a list of operation dictionaries, one
1675 for each active operation. Once an operation is completed, it will no longer
1676 appear in data["active"] .
1678 Each op-dict contains a "type" key, one of "upload", "download",
1679 "mapupdate", "publish", or "retrieve" (the first two are for immutable
1680 files, while the latter three are for mutable files and directories).
1682 The "upload" op-dict will contain the following keys::
1684 type (string): "upload"
1685 storage-index-string (string): a base32-encoded storage index
1686 total-size (int): total size of the file
1687 status (string): current status of the operation
1688 progress-hash (float): 1.0 when the file has been hashed
1689 progress-ciphertext (float): 1.0 when the file has been encrypted.
1690 progress-encode-push (float): 1.0 when the file has been encoded and
1691 pushed to the storage servers. For helper
1692 uploads, the ciphertext value climbs to 1.0
1693 first, then encoding starts. For unassisted
1694 uploads, ciphertext and encode-push progress
1695 will climb at the same pace.
1697 The "download" op-dict will contain the following keys::
1699 type (string): "download"
1700 storage-index-string (string): a base32-encoded storage index
1701 total-size (int): total size of the file
1702 status (string): current status of the operation
1703 progress (float): 1.0 when the file has been fully downloaded
1705 Front-ends which want to report progress information are advised to simply
1706 average together all the progress-* indicators. A slightly more accurate
1707 value can be found by ignoring the progress-hash value (since the current
1708 implementation hashes synchronously, so clients will probably never see
1709 progress-hash!=1.0).
1711 ``GET /provisioning/``
1713 This page provides a basic tool to predict the likely storage and bandwidth
1714 requirements of a large Tahoe grid. It provides forms to input things like
1715 total number of users, number of files per user, average file size, number
1716 of servers, expansion ratio, hard drive failure rate, etc. It then provides
1717 numbers like how many disks per server will be needed, how many read
1718 operations per second should be expected, and the likely MTBF for files in
1719 the grid. This information is very preliminary, and the model upon which it
1720 is based still needs a lot of work.
1722 ``GET /helper_status/``
1724 If the node is running a helper (i.e. if [helper]enabled is set to True in
1725 tahoe.cfg), then this page will provide a list of all the helper operations
1726 currently in progress. If "?t=json" is added to the URL, it will return a
1727 JSON-formatted list of helper statistics, which can then be used to produce
1728 graphs to indicate how busy the helper is.
1730 ``GET /statistics/``
1732 This page provides "node statistics", which are collected from a variety of
1735 load_monitor: every second, the node schedules a timer for one second in
1736 the future, then measures how late the subsequent callback
1737 is. The "load_average" is this tardiness, measured in
1738 seconds, averaged over the last minute. It is an indication
1739 of a busy node, one which is doing more work than can be
1740 completed in a timely fashion. The "max_load" value is the
1741 highest value that has been seen in the last 60 seconds.
1743 cpu_monitor: every minute, the node uses time.clock() to measure how much
1744 CPU time it has used, and it uses this value to produce
1745 1min/5min/15min moving averages. These values range from 0%
1746 (0.0) to 100% (1.0), and indicate what fraction of the CPU
1747 has been used by the Tahoe node. Not all operating systems
1748 provide meaningful data to time.clock(): they may report 100%
1749 CPU usage at all times.
1751 uploader: this counts how many immutable files (and bytes) have been
1752 uploaded since the node was started
1754 downloader: this counts how many immutable files have been downloaded
1755 since the node was started
1757 publishes: this counts how many mutable files (including directories) have
1758 been modified since the node was started
1760 retrieves: this counts how many mutable files (including directories) have
1761 been read since the node was started
1763 There are other statistics that are tracked by the node. The "raw stats"
1764 section shows a formatted dump of all of them.
1766 By adding "?t=json" to the URL, the node will return a JSON-formatted
1767 dictionary of stats values, which can be used by other tools to produce
1768 graphs of node behavior. The misc/munin/ directory in the source
1769 distribution provides some tools to produce these graphs.
1771 ``GET /`` (introducer status)
1773 For Introducer nodes, the welcome page displays information about both
1774 clients and servers which are connected to the introducer. Servers make
1775 "service announcements", and these are listed in a table. Clients will
1776 subscribe to hear about service announcements, and these subscriptions are
1777 listed in a separate table. Both tables contain information about what
1778 version of Tahoe is being run by the remote node, their advertised and
1779 outbound IP addresses, their nodeid and nickname, and how long they have
1782 By adding "?t=json" to the URL, the node will return a JSON-formatted
1783 dictionary of stats values, which can be used to produce graphs of connected
1784 clients over time. This dictionary has the following keys::
1786 ["subscription_summary"] : a dictionary mapping service name (like
1787 "storage") to an integer with the number of
1788 clients that have subscribed to hear about that
1790 ["announcement_summary"] : a dictionary mapping service name to an integer
1791 with the number of servers which are announcing
1793 ["announcement_distinct_hosts"] : a dictionary mapping service name to an
1794 integer which represents the number of
1795 distinct hosts that are providing that
1796 service. If two servers have announced
1797 FURLs which use the same hostnames (but
1798 different ports and tubids), they are
1799 considered to be on the same host.
1802 Static Files in /public_html
1803 ============================
1805 The web-API server will take any request for a URL that starts with /static
1806 and serve it from a configurable directory which defaults to
1807 $BASEDIR/public_html . This is configured by setting the "[node]web.static"
1808 value in $BASEDIR/tahoe.cfg . If this is left at the default value of
1809 "public_html", then http://localhost:3456/static/subdir/foo.html will be
1810 served with the contents of the file $BASEDIR/public_html/subdir/foo.html .
1812 This can be useful to serve a javascript application which provides a
1813 prettier front-end to the rest of the Tahoe web-API.
1816 Safety and security issues -- names vs. URIs
1817 ============================================
1819 Summary: use explicit file- and dir- caps whenever possible, to reduce the
1820 potential for surprises when the filesystem structure is changed.
1822 Tahoe provides a mutable filesystem, but the ways that the filesystem can
1823 change are limited. The only thing that can change is that the mapping from
1824 child names to child objects that each directory contains can be changed by
1825 adding a new child name pointing to an object, removing an existing child name,
1826 or changing an existing child name to point to a different object.
1828 Obviously if you query Tahoe for information about the filesystem and then act
1829 to change the filesystem (such as by getting a listing of the contents of a
1830 directory and then adding a file to the directory), then the filesystem might
1831 have been changed after you queried it and before you acted upon it. However,
1832 if you use the URI instead of the pathname of an object when you act upon the
1833 object, then the only change that can happen is if the object is a directory
1834 then the set of child names it has might be different. If, on the other hand,
1835 you act upon the object using its pathname, then a different object might be in
1836 that place, which can result in more kinds of surprises.
1838 For example, suppose you are writing code which recursively downloads the
1839 contents of a directory. The first thing your code does is fetch the listing
1840 of the contents of the directory. For each child that it fetched, if that
1841 child is a file then it downloads the file, and if that child is a directory
1842 then it recurses into that directory. Now, if the download and the recurse
1843 actions are performed using the child's name, then the results might be
1844 wrong, because for example a child name that pointed to a sub-directory when
1845 you listed the directory might have been changed to point to a file (in which
1846 case your attempt to recurse into it would result in an error and the file
1847 would be skipped), or a child name that pointed to a file when you listed the
1848 directory might now point to a sub-directory (in which case your attempt to
1849 download the child would result in a file containing HTML text describing the
1852 If your recursive algorithm uses the uri of the child instead of the name of
1853 the child, then those kinds of mistakes just can't happen. Note that both the
1854 child's name and the child's URI are included in the results of listing the
1855 parent directory, so it isn't any harder to use the URI for this purpose.
1857 The read and write caps in a given directory node are separate URIs, and
1858 can't be assumed to point to the same object even if they were retrieved in
1859 the same operation (although the web-API server attempts to ensure this
1860 in most cases). If you need to rely on that property, you should explicitly
1861 verify it. More generally, you should not make assumptions about the
1862 internal consistency of the contents of mutable directories. As a result
1863 of the signatures on mutable object versions, it is guaranteed that a given
1864 version was written in a single update, but -- as in the case of a file --
1865 the contents may have been chosen by a malicious writer in a way that is
1866 designed to confuse applications that rely on their consistency.
1868 In general, use names if you want "whatever object (whether file or
1869 directory) is found by following this name (or sequence of names) when my
1870 request reaches the server". Use URIs if you want "this particular object".
1875 Tahoe uses both mutable and immutable files. Mutable files can be created
1876 explicitly by doing an upload with ?mutable=true added, or implicitly by
1877 creating a new directory (since a directory is just a special way to
1878 interpret a given mutable file).
1880 Mutable files suffer from the same consistency-vs-availability tradeoff that
1881 all distributed data storage systems face. It is not possible to
1882 simultaneously achieve perfect consistency and perfect availability in the
1883 face of network partitions (servers being unreachable or faulty).
1885 Tahoe tries to achieve a reasonable compromise, but there is a basic rule in
1886 place, known as the Prime Coordination Directive: "Don't Do That". What this
1887 means is that if write-access to a mutable file is available to several
1888 parties, then those parties are responsible for coordinating their activities
1889 to avoid multiple simultaneous updates. This could be achieved by having
1890 these parties talk to each other and using some sort of locking mechanism, or
1891 by serializing all changes through a single writer.
1893 The consequences of performing uncoordinated writes can vary. Some of the
1894 writers may lose their changes, as somebody else wins the race condition. In
1895 many cases the file will be left in an "unhealthy" state, meaning that there
1896 are not as many redundant shares as we would like (reducing the reliability
1897 of the file against server failures). In the worst case, the file can be left
1898 in such an unhealthy state that no version is recoverable, even the old ones.
1899 It is this small possibility of data loss that prompts us to issue the Prime
1900 Coordination Directive.
1902 Tahoe nodes implement internal serialization to make sure that a single Tahoe
1903 node cannot conflict with itself. For example, it is safe to issue two
1904 directory modification requests to a single tahoe node's web-API server at the
1905 same time, because the Tahoe node will internally delay one of them until
1906 after the other has finished being applied. (This feature was introduced in
1907 Tahoe-1.1; back with Tahoe-1.0 the web client was responsible for serializing
1908 web requests themselves).
1910 For more details, please see the "Consistency vs Availability" and "The Prime
1911 Coordination Directive" sections of `mutable.rst <../specifications/mutable.rst>`_.
1914 .. [1] URLs and HTTP and UTF-8, Oh My
1916 HTTP does not provide a mechanism to specify the character set used to
1917 encode non-ASCII names in URLs
1918 (`RFC3986#2.1 <http://tools.ietf.org/html/rfc3986#section-2.1>`_).
1919 We prefer the convention that the ``filename=`` argument shall be a
1920 URL-escaped UTF-8 encoded Unicode string.
1921 For example, suppose we want to provoke the server into using a filename of
1922 "f i a n c e-acute e" (i.e. f i a n c U+00E9 e). The UTF-8 encoding of this
1923 is 0x66 0x69 0x61 0x6e 0x63 0xc3 0xa9 0x65 (or "fianc\\xC3\\xA9e", as python's
1924 ``repr()`` function would show). To encode this into a URL, the non-printable
1925 characters must be escaped with the urlencode ``%XX`` mechanism, giving
1926 us "fianc%C3%A9e". Thus, the first line of the HTTP request will be
1927 "``GET /uri/CAP...?save=true&filename=fianc%C3%A9e HTTP/1.1``". Not all
1928 browsers provide this: IE7 by default uses the Latin-1 encoding, which is
1929 "fianc%E9e" (although it has a configuration option to send URLs as UTF-8).
1931 The response header will need to indicate a non-ASCII filename. The actual
1932 mechanism to do this is not clear. For ASCII filenames, the response header
1935 Content-Disposition: attachment; filename="english.txt"
1937 If Tahoe were to enforce the UTF-8 convention, it would need to decode the
1938 URL argument into a Unicode string, and then encode it back into a sequence
1939 of bytes when creating the response header. One possibility would be to use
1940 unencoded UTF-8. Developers suggest that IE7 might accept this::
1942 #1: Content-Disposition: attachment; filename="fianc\xC3\xA9e"
1943 (note, the last four bytes of that line, not including the newline, are
1944 0xC3 0xA9 0x65 0x22)
1946 `RFC2231#4 <http://tools.ietf.org/html/rfc2231#section-4>`_
1947 (dated 1997): suggests that the following might work, and
1948 `some developers have reported <http://markmail.org/message/dsjyokgl7hv64ig3>`_
1949 that it is supported by Firefox (but not IE7)::
1951 #2: Content-Disposition: attachment; filename*=utf-8''fianc%C3%A9e
1953 My reading of `RFC2616#19.5.1 <http://tools.ietf.org/html/rfc2616#section-19.5.1>`_
1954 (which defines Content-Disposition) says that the filename= parameter is
1955 defined to be wrapped in quotes (presumably to allow spaces without breaking
1956 the parsing of subsequent parameters), which would give us::
1958 #3: Content-Disposition: attachment; filename*=utf-8''"fianc%C3%A9e"
1960 However this is contrary to the examples in the email thread listed above.
1962 Developers report that IE7 (when it is configured for UTF-8 URL encoding,
1963 which is not the default in Asian countries), will accept::
1965 #4: Content-Disposition: attachment; filename=fianc%C3%A9e
1967 However, for maximum compatibility, Tahoe simply copies bytes from the URL
1968 into the response header, rather than enforcing the UTF-8 convention. This
1969 means it does not try to decode the filename from the URL argument, nor does
1970 it encode the filename into the response header.