2 = The Tahoe REST-ful Web API =
4 1. Enabling the web-API port
5 2. Basic Concepts: GET, PUT, DELETE, POST
6 3. URLs, Machine-Oriented Interfaces
7 4. Browser Operations: Human-Oriented Interfaces
8 5. Welcome / Debug / Status pages
9 6. Static Files in /public_html
10 7. Safety and security issues -- names vs. URIs
14 == Enabling the web-API port ==
16 Every Tahoe node is capable of running a built-in HTTP server. To enable
17 this, just write a port number into the "[node]web.port" line of your node's
18 tahoe.cfg file. For example, writing "web.port = 3456" into the "[node]"
19 section of $NODEDIR/tahoe.cfg will cause the node to run a webserver on port
22 This string is actually a Twisted "strports" specification, meaning you can
23 get more control over the interface to which the server binds by supplying
24 additional arguments. For more details, see the documentation on
25 twisted.application.strports:
26 http://twistedmatrix.com/documents/current/api/twisted.application.strports.html
28 Writing "tcp:3456:interface=127.0.0.1" into the web.port line does the same
29 but binds to the loopback interface, ensuring that only the programs on the
30 local host can connect. Using
31 "ssl:3456:privateKey=mykey.pem:certKey=cert.pem" runs an SSL server.
33 This webport can be set when the node is created by passing a --webport
34 option to the 'tahoe create-client' command. By default, the node listens on
35 port 3456, on the loopback (127.0.0.1) interface.
39 As described in architecture.txt, each file and directory in a Tahoe virtual
40 filesystem is referenced by an identifier that combines the designation of
41 the object with the authority to do something with it (such as read or modify
42 the contents). This identifier is called a "read-cap" or "write-cap",
43 depending upon whether it enables read-only or read-write access. These
44 "caps" are also referred to as URIs.
46 The Tahoe web-based API is "REST-ful", meaning it implements the concepts of
47 "REpresentational State Transfer": the original scheme by which the World
48 Wide Web was intended to work. Each object (file or directory) is referenced
49 by a URL that includes the read- or write- cap. HTTP methods (GET, PUT, and
50 DELETE) are used to manipulate these objects. You can think of the URL as a
51 noun, and the method as a verb.
53 In REST, the GET method is used to retrieve information about an object, or
54 to retrieve some representation of the object itself. When the object is a
55 file, the basic GET method will simply return the contents of that file.
56 Other variations (generally implemented by adding query parameters to the
57 URL) will return information about the object, such as metadata. GET
58 operations are required to have no side-effects.
60 PUT is used to upload new objects into the filesystem, or to replace an
61 existing object. DELETE it used to delete objects from the filesystem. Both
62 PUT and DELETE are required to be idempotent: performing the same operation
63 multiple times must have the same side-effects as only performing it once.
65 POST is used for more complicated actions that cannot be expressed as a GET,
66 PUT, or DELETE. POST operations can be thought of as a method call: sending
67 some message to the object referenced by the URL. In Tahoe, POST is also used
68 for operations that must be triggered by an HTML form (including upload and
69 delete), because otherwise a regular web browser has no way to accomplish
70 these tasks. In general, everything that can be done with a PUT or DELETE can
71 also be done with a POST.
73 Tahoe's web API is designed for two different consumers. The first is a
74 program that needs to manipulate the virtual file system. Such programs are
75 expected to use the RESTful interface described above. The second is a human
76 using a standard web browser to work with the filesystem. This user is given
77 a series of HTML pages with links to download files, and forms that use POST
78 actions to upload, rename, and delete files.
80 When an error occurs, the HTTP response code will be set to an appropriate
81 400-series code (like 404 for an unknown childname, or 400 Gone when a file
82 is unrecoverable due to insufficient shares), and the HTTP response body will
83 usually contain a few lines of explanation as to the cause of the error and
84 possible responses. Unusual exceptions may result in a 500 Internal Server
85 Error as a catch-all, with a default response body will contain a
86 Nevow-generated HTML-ized representation of the Python exception stack trace
87 that caused the problem. CLI programs which want to copy the response body to
88 stderr should provide an "Accept: text/plain" header to their requests to get
89 a plain text stack trace instead. If the Accept header contains */*, or
90 text/*, or text/html (or if there is no Accept header), HTML tracebacks will
95 Tahoe uses a variety of read- and write- caps to identify files and
96 directories. The most common of these is the "immutable file read-cap", which
97 is used for most uploaded files. These read-caps look like the following:
99 URI:CHK:ime6pvkaxuetdfah2p2f35pe54:4btz54xk3tew6nd4y2ojpxj4m6wxjqqlwnztgre6gnjgtucd5r4a:3:10:202
101 The next most common is a "directory write-cap", which provides both read and
102 write access to a directory, and look like this:
104 URI:DIR2:djrdkfawoqihigoett4g6auz6a:jx5mplfpwexnoqff7y5e4zjus4lidm76dcuarpct7cckorh2dpgq
106 There are also "directory read-caps", which start with "URI:DIR2-RO:", and
107 give read-only access to a directory. Finally there are also mutable file
108 read- and write- caps, which start with "URI:SSK", and give access to mutable
111 (later versions of Tahoe will make these strings shorter, and will remove the
112 unfortunate colons, which must be escaped when these caps are embedded in
115 To refer to any Tahoe object through the web API, you simply need to combine
116 a prefix (which indicates the HTTP server to use) with the cap (which
117 indicates which object inside that server to access). Since the default Tahoe
118 webport is 3456, the most common prefix is one that will use a local node
119 listening on this port:
121 http://127.0.0.1:3456/uri/ + $CAP
123 So, to access the directory named above (which happens to be the
124 publically-writable sample directory on the Tahoe test grid, described at
125 http://allmydata.org/trac/tahoe/wiki/TestGrid), the URL would be:
127 http://127.0.0.1:3456/uri/URI%3ADIR2%3Adjrdkfawoqihigoett4g6auz6a%3Ajx5mplfpwexnoqff7y5e4zjus4lidm76dcuarpct7cckorh2dpgq/
129 (note that the colons in the directory-cap are url-encoded into "%3A"
132 Likewise, to access the file named above, use:
134 http://127.0.0.1:3456/uri/URI%3ACHK%3Aime6pvkaxuetdfah2p2f35pe54%3A4btz54xk3tew6nd4y2ojpxj4m6wxjqqlwnztgre6gnjgtucd5r4a%3A3%3A10%3A202
136 In the rest of this document, we'll use "$DIRCAP" as shorthand for a read-cap
137 or write-cap that refers to a directory, and "$FILECAP" to abbreviate a cap
138 that refers to a file (whether mutable or immutable). So those URLs above can
141 http://127.0.0.1:3456/uri/$DIRCAP/
142 http://127.0.0.1:3456/uri/$FILECAP
144 The operation summaries below will abbreviate these further, by eliding the
145 server prefix. They will be displayed like this:
153 Tahoe directories contain named children, just like directories in a regular
154 local filesystem. These children can be either files or subdirectories.
156 If you have a Tahoe URL that refers to a directory, and want to reference a
157 named child inside it, just append the child name to the URL. For example, if
158 our sample directory contains a file named "welcome.txt", we can refer to
161 http://127.0.0.1:3456/uri/$DIRCAP/welcome.txt
163 (or http://127.0.0.1:3456/uri/URI%3ADIR2%3Adjrdkfawoqihigoett4g6auz6a%3Ajx5mplfpwexnoqff7y5e4zjus4lidm76dcuarpct7cckorh2dpgq/welcome.txt)
165 Multiple levels of subdirectories can be handled this way:
167 http://127.0.0.1:3456/uri/$DIRCAP/tahoe-source/docs/webapi.txt
169 In this document, when we need to refer to a URL that references a file using
170 this child-of-some-directory format, we'll use the following string:
172 /uri/$DIRCAP/[SUBDIRS../]FILENAME
174 The "[SUBDIRS../]" part means that there are zero or more (optional)
175 subdirectory names in the middle of the URL. The "FILENAME" at the end means
176 that this whole URL refers to a file of some sort, rather than to a
179 When we need to refer specifically to a directory in this way, we'll write:
181 /uri/$DIRCAP/[SUBDIRS../]SUBDIR
184 Note that all components of pathnames in URLs are required to be UTF-8
185 encoded, so "resume.doc" (with an acute accent on both E's) would be accessed
188 http://127.0.0.1:3456/uri/$DIRCAP/r%C3%A9sum%C3%A9.doc
190 Also note that the filenames inside upload POST forms are interpreted using
191 whatever character set was provided in the conventional '_charset' field, and
192 defaults to UTF-8 if not otherwise specified. The JSON representation of each
193 directory contains native unicode strings. Tahoe directories are specified to
194 contain unicode filenames, and cannot contain binary strings that are not
195 representable as such.
197 All Tahoe operations that refer to existing files or directories must include
198 a suitable read- or write- cap in the URL: the wapi server won't add one
199 for you. If you don't know the cap, you can't access the file. This allows
200 the security properties of Tahoe caps to be extended across the wapi
203 == Slow Operations, Progress, and Cancelling ==
205 Certain operations can be expected to take a long time. The "t=deep-check",
206 described below, will recursively visit every file and directory reachable
207 from a given starting point, which can take minutes or even hours for
208 extremely large directory structures. A single long-running HTTP request is a
209 fragile thing: proxies, NAT boxes, browsers, and users may all grow impatient
210 with waiting and give up on the connection.
212 For this reason, long-running operations have an "operation handle", which
213 can be used to poll for status/progress messages while the operation
214 proceeds. This handle can also be used to cancel the operation. These handles
215 are created by the client, and passed in as a an "ophandle=" query argument
216 to the POST or PUT request which starts the operation. The following
217 operations can then be used to retrieve status:
219 GET /operations/$HANDLE?output=HTML (with or without t=status)
220 GET /operations/$HANDLE?output=JSON (same)
222 These two retrieve the current status of the given operation. Each operation
223 presents a different sort of information, but in general the page retrieved
226 * whether the operation is complete, or if it is still running
227 * how much of the operation is complete, and how much is left, if possible
229 Note that the final status output can be quite large: a deep-manifest of a
230 directory structure with 300k directories and 200k unique files is about
231 275MB of JSON, and might take two minutes to generate. For this reason, the
232 full status is not provided until the operation has completed.
234 The HTML form will include a meta-refresh tag, which will cause a regular
235 web browser to reload the status page about 60 seconds later. This tag will
236 be removed once the operation has completed.
238 There may be more status information available under
239 /operations/$HANDLE/$ETC : i.e., the handle forms the root of a URL space.
241 POST /operations/$HANDLE?t=cancel
243 This terminates the operation, and returns an HTML page explaining what was
244 cancelled. If the operation handle has already expired (see below), this
245 POST will return a 404, which indicates that the operation is no longer
246 running (either it was completed or terminated). The response body will be
247 the same as a GET /operations/$HANDLE on this operation handle, and the
248 handle will be expired immediately afterwards.
250 The operation handle will eventually expire, to avoid consuming an unbounded
251 amount of memory. The handle's time-to-live can be reset at any time, by
252 passing a retain-for= argument (with a count of seconds) to either the
253 initial POST that starts the operation, or the subsequent GET request which
254 asks about the operation. For example, if a 'GET
255 /operations/$HANDLE?output=JSON&retain-for=600' query is performed, the
256 handle will remain active for 600 seconds (10 minutes) after the GET was
259 In addition, if the GET includes a release-after-complete=True argument, and
260 the operation has completed, the operation handle will be released
263 If a retain-for= argument is not used, the default handle lifetimes are:
265 * handles will remain valid at least until their operation finishes
266 * uncollected handles for finished operations (i.e. handles for operations
267 which have finished but for which the GET page has not been accessed since
268 completion) will remain valid for one hour, or for the total time consumed
269 by the operation, whichever is greater.
270 * collected handles (i.e. the GET page has been retrieved at least once
271 since the operation completed) will remain valid for ten minutes.
273 Many "slow" operations can begin to use unacceptable amounts of memory when
274 operation on large directory structures. The memory usage increases when the
275 ophandle is polled, as the results must be copied into a JSON string, sent
276 over the wire, then parsed by a client. So, as an alternative, many "slow"
277 operations have streaming equivalents. These equivalents do not use operation
278 handles. Instead, they emit line-oriented status results immediately. Client
279 code can cancel the operation by simply closing the HTTP connection.
281 == Programmatic Operations ==
283 Now that we know how to build URLs that refer to files and directories in a
284 Tahoe virtual filesystem, what sorts of operations can we do with those URLs?
285 This section contains a catalog of GET, PUT, DELETE, and POST operations that
286 can be performed on these URLs. This set of operations are aimed at programs
287 that use HTTP to communicate with a Tahoe node. A later section describes
288 operations that are intended for web browsers.
290 === Reading A File ===
293 GET /uri/$DIRCAP/[SUBDIRS../]FILENAME
295 This will retrieve the contents of the given file. The HTTP response body
296 will contain the sequence of bytes that make up the file.
298 To view files in a web browser, you may want more control over the
299 Content-Type and Content-Disposition headers. Please see the next section
300 "Browser Operations", for details on how to modify these URLs for that
303 === Writing/Uploading A File ===
306 PUT /uri/$DIRCAP/[SUBDIRS../]FILENAME
308 Upload a file, using the data from the HTTP request body, and add whatever
309 child links and subdirectories are necessary to make the file available at
310 the given location. Once this operation succeeds, a GET on the same URL will
311 retrieve the same contents that were just uploaded. This will create any
312 necessary intermediate subdirectories.
314 To use the /uri/$FILECAP form, $FILECAP be a write-cap for a mutable file.
316 In the /uri/$DIRCAP/[SUBDIRS../]FILENAME form, if the target file is a
317 writable mutable file, that files contents will be overwritten in-place. If
318 it is a read-cap for a mutable file, an error will occur. If it is an
319 immutable file, the old file will be discarded, and a new one will be put in
322 When creating a new file, if "mutable=true" is in the query arguments, the
323 operation will create a mutable file instead of an immutable one.
325 This returns the file-cap of the resulting file. If a new file was created
326 by this method, the HTTP response code (as dictated by rfc2616) will be set
327 to 201 CREATED. If an existing file was replaced or modified, the response
330 Note that the 'curl -T localfile http://127.0.0.1:3456/uri/$DIRCAP/foo.txt'
331 command can be used to invoke this operation.
335 This uploads a file, and produces a file-cap for the contents, but does not
336 attach the file into the virtual drive. No directories will be modified by
337 this operation. The file-cap is returned as the body of the HTTP response.
339 If "mutable=true" is in the query arguments, the operation will create a
340 mutable file, and return its write-cap in the HTTP respose. The default is
341 to create an immutable file, returning the read-cap as a response.
343 === Creating A New Directory ===
348 Create a new empty directory and return its write-cap as the HTTP response
349 body. This does not make the newly created directory visible from the
350 virtual drive. The "PUT" operation is provided for backwards compatibility:
351 new code should use POST.
353 POST /uri?t=mkdir-with-children
355 Create a new directory, populated with a set of child nodes, and return its
356 write-cap as the HTTP response body. The new directory is not attached to
357 any other directory: the returned write-cap is the only reference to it.
359 Initial children are provided as the body of the POST form (this is more
360 efficient than doing separate mkdir and set_children operations). If the
361 body is empty, the new directory will be empty. If not empty, the body will
362 be interpreted as a JSON-encoded dictionary of children with which the new
363 directory should be populated, using the same format as would be returned in
364 the 'children' value of the t=json GET request, described below. Each
365 dictionary key should be a child name, and each value should be a list of
366 [TYPE, PROPDICT], where PROPDICT contains "rw_uri", "ro_uri", and "metadata"
367 keys (all others are ignored). For example, the PUT request body could be:
370 "Fran\u00e7ais": [ "filenode", {
371 "ro_uri": "URI:CHK:...",
374 "ctime": 1202777696.7564139,
375 "mtime": 1202777696.7564139,
377 "linkcrtime": 1202777696.7564139,
378 "linkmotime": 1202777696.7564139,
380 "subdir": [ "dirnode", {
381 "rw_uri": "URI:DIR2:...",
382 "ro_uri": "URI:DIR2-RO:...",
384 "ctime": 1202778102.7589991,
385 "mtime": 1202778111.2160511,
387 "linkcrtime": 1202777696.7564139,
388 "linkmotime": 1202777696.7564139,
392 Note that the webapi-using client application must not provide the
393 "Content-Type: multipart/form-data" header that usually accompanies HTML
394 form submissions, since the body is not formatted this way. Doing so will
395 cause a server error as the lower-level code misparses the request body.
397 POST /uri?t=mkdir-immutable
399 Like t=mkdir-with-children above, but the new directory will be
400 deep-immutable. This means that the directory itself is immutable, and that
401 it can only contain deep-immutable objects, like immutable files, literal
402 files, and deep-immutable directories. A non-empty request body is
403 mandatory, since after the directory is created, it will not be possible to
404 add more children to it.
406 POST /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=mkdir
407 PUT /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=mkdir
409 Create new directories as necessary to make sure that the named target
410 ($DIRCAP/SUBDIRS../SUBDIR) is a directory. This will create additional
411 intermediate directories as necessary. If the named target directory already
412 exists, this will make no changes to it.
414 If the final directory is created, it will be empty.
416 This will return an error if a blocking file is present at any of the parent
417 names, preventing the server from creating the necessary parent directory.
419 The write-cap of the new directory will be returned as the HTTP response
422 POST /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=mkdir-with-children
424 Like above, but if the final directory is created, it will be populated with
425 initial children from the POST request body, as described above in the
426 /uri?t=mkdir-with-children operation.
428 POST /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=mkdir-immutable
430 Like above, but the final directory will be deep-immutable, with the
431 children specified as a JSON dictionary in the POST request body.
433 POST /uri/$DIRCAP/[SUBDIRS../]?t=mkdir&name=NAME
435 Create a new empty directory and attach it to the given existing directory.
436 This will create additional intermediate directories as necessary.
438 The URL of this form points to the parent of the bottom-most new directory,
439 whereas the previous form has a URL that points directly to the bottom-most
442 POST /uri/$DIRCAP/[SUBDIRS../]?t=mkdir-with-children&name=NAME
444 As above, but the new directory will be populated with initial children via
445 the POST request body, as described in /uri?t=mkdir-with-children above.
446 Note that the name= argument must be passed as a queryarg, because the POST
447 request body is used for the initial children JSON.
449 POST /uri/$DIRCAP/[SUBDIRS../]?t=mkdir-immutable&name=NAME
451 As above, but the new directory will be deep-immutable, with the children
452 specified as a JSON dictionary in the POST request body. Again, the name=
453 argument must be passed as a queryarg.
455 === Get Information About A File Or Directory (as JSON) ===
457 GET /uri/$FILECAP?t=json
458 GET /uri/$DIRCAP?t=json
459 GET /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=json
460 GET /uri/$DIRCAP/[SUBDIRS../]FILENAME?t=json
462 This returns a machine-parseable JSON-encoded description of the given
463 object. The JSON always contains a list, and the first element of the list is
464 always a flag that indicates whether the referenced object is a file or a
465 directory. If it is a capability to a file, then the information includes
466 file size and URI, like this:
468 GET /uri/$FILECAP?t=json :
472 "verify_uri": verify_uri,
477 If it is a capability to a directory followed by a path from that directory
478 to a file, then the information also includes metadata from the link to the
479 file in the parent directory, like this:
481 GET /uri/$DIRCAP/[SUBDIRS../]FILENAME?t=json :
485 "verify_uri": verify_uri,
489 "ctime": 1202777696.7564139,
490 "mtime": 1202777696.7564139,
492 "linkcrtime": 1202777696.7564139,
493 "linkmotime": 1202777696.7564139,
496 If it is a directory, then it includes information about the children of
497 this directory, as a mapping from child name to a set of data about the
498 child (the same data that would appear in a corresponding GET?t=json of the
499 child itself). The child entries also include metadata about each child,
500 including link-creation- and link-change- timestamps. The output looks like
503 GET /uri/$DIRCAP?t=json :
504 GET /uri/$DIRCAP/[SUBDIRS../]SUBDIR?t=json :
507 "rw_uri": read_write_uri,
508 "ro_uri": read_only_uri,
509 "verify_uri": verify_uri,
512 "foo.txt": [ "filenode", {
516 "ctime": 1202777696.7564139,
517 "mtime": 1202777696.7564139,
519 "linkcrtime": 1202777696.7564139,
520 "linkmotime": 1202777696.7564139,
522 "subdir": [ "dirnode", {
526 "ctime": 1202778102.7589991,
527 "mtime": 1202778111.2160511,
529 "linkcrtime": 1202777696.7564139,
530 "linkmotime": 1202777696.7564139,
534 In the above example, note how 'children' is a dictionary in which the keys
535 are child names and the values depend upon whether the child is a file or a
536 directory. The value is mostly the same as the JSON representation of the
537 child object (except that directories do not recurse -- the "children"
538 entry of the child is omitted, and the directory view includes the metadata
539 that is stored on the directory edge).
541 Then the rw_uri field will be present in the information about a directory
542 if and only if you have read-write access to that directory. The verify_uri
543 field will be presend if and only if the object has a verify-cap
544 (non-distributed LIT files do not have verify-caps).
546 ==== About the metadata ====
548 The value of the 'mtime' key and of the 'tahoe':'linkmotime' is updated
549 whenever a link to a child is set. The value of the 'ctime' key and of the
550 'tahoe':'linkcrtime' key is updated whenever a link to a child is created --
551 i.e. when there was not previously a link under that name.
553 In Tahoe earlier than v1.4.0, only the 'mtime'/'ctime' keys were populated.
554 Starting in Tahoe v1.4.0, the 'linkmotime'/'linkcrtime' keys in the 'tahoe'
555 sub-dict are also populated.
557 The reason we added the new values in Tahoe v1.4.0 is that there is a
558 "set_children" API (described below) which you can use to overwrite the
559 values of the 'mtime'/'ctime' pair, and this API is used by the "tahoe
560 backup" command (both in Tahoe v1.3.0 and in Tahoe v1.4.0) to set the
561 'mtime' and 'ctime' values when backing up files from a local filesystem
562 into the Tahoe filesystem. As of Tahoe v1.4.0, the set_children API cannot
563 be used to set anything under the 'tahoe' key of the metadata dict -- if
564 you include 'tahoe' keys in your 'metadata' arguments then it will silently
567 Therefore, if the 'tahoe' sub-dict is present, you can rely on the
568 'linkcrtime' and 'linkmotime' values therein to have the semantics described
569 above. (This is assuming that only official Tahoe clients have been used to
570 write those links, and that their system clocks were set to what you expected
571 -- there is nothing preventing someone from editing their Tahoe client or
572 writing their own Tahoe client which would overwrite those values however
573 they like, and there is nothing to constrain their system clock from taking
576 The meaning of the 'ctime'/'mtime' fields are slightly more complex.
578 The meaning of the 'mtime' field is: whenever the edge is updated (by an HTTP
579 PUT or POST, as is done by the "tahoe cp" command), then the mtime is set to
580 the current time on the clock of the updating client. Whenever the edge is
581 updated by "tahoe backup" then the mtime is instead set to the value which
582 the updating client read from its local filesystem for the "mtime" of the
583 local file in question, which means the last time the contents of that file
584 were changed. Note however, that if the edge in the Tahoe filesystem points
585 to a mutable file and the contents of that mutable file is changed then the
586 "mtime" value on that edge will *not* be updated, since the edge itself
587 wasn't updated -- only the mutable file was.
589 The meaning of the 'ctime' field is even more complex. Whenever a new edge is
590 created (by an HTTP PUT or POST, as is done by "tahoe cp") then the ctime is
591 set to the current time on the clock of the updating client. Whenever the
592 edge is created *or updated* by "tahoe backup" then the ctime is instead set
593 to the value which the updating client read from its local filesystem. On
594 Windows, it reads the timestamp of when the local file was created and puts
595 that into the "ctime", and on other platforms it reads the timestamp of the
596 most recent time that either the contents or the metadata of the local file
597 was changed and puts that into the ctime. Again, if the edge points to a
598 mutable file and the content of that mutable file is changed then the ctime
599 will not be updated in any case.
601 Therefore there are several ways that the 'ctime' field could be confusing:
603 1. You might be confused about whether it reflects the time of the creation
604 of a link in the Tahoe filesystem or a timestamp copied in from a local
607 2. You might be confused about whether it is a copy of the file creation time
608 (if "tahoe backup" was run on a Windows system) or of the last
609 contents-or-metadata change (if "tahoe backup" was run on a different
612 3. You might be confused by the fact that changing the contents of a mutable
613 file in Tahoe don't have any effect on any links pointing at that file in any
614 directories, although "tahoe backup" sets the link 'ctime'/'mtime' to reflect
615 timestamps about the local file corresponding to the Tahoe file to which the
618 4. Also, quite apart from Tahoe, you might be confused about the meaning of
619 the 'ctime' in unix local filesystems, which people sometimes think means
620 file creation time, but which actually means, in unix local filesystems, the
621 most recent time that the file contents or the file metadata (such as owner,
622 permission bits, extended attributes, etc.) has changed. Note that although
623 'ctime' does not mean file creation time in Unix, it does mean link creation
624 time in Tahoe, unless the "tahoe backup" command has been used on that link,
625 in which case it means something about the local filesystem file which
626 corresponds to the Tahoe file which is pointed at by the link. It means
627 either file creation time of the local file (if "tahoe backup" was run on
628 Windows) or file-contents-or-metadata-update-time of the local file (if
629 "tahoe backup" was run on a different operating system).
632 === Attaching an existing File or Directory by its read- or write- cap ===
634 PUT /uri/$DIRCAP/[SUBDIRS../]CHILDNAME?t=uri
636 This attaches a child object (either a file or directory) to a specified
637 location in the virtual filesystem. The child object is referenced by its
638 read- or write- cap, as provided in the HTTP request body. This will create
639 intermediate directories as necessary.
641 This is similar to a UNIX hardlink: by referencing a previously-uploaded file
642 (or previously-created directory) instead of uploading/creating a new one,
643 you can create two references to the same object.
645 The read- or write- cap of the child is provided in the body of the HTTP
646 request, and this same cap is returned in the response body.
648 The default behavior is to overwrite any existing object at the same
649 location. To prevent this (and make the operation return an error instead
650 of overwriting), add a "replace=false" argument, as "?t=uri&replace=false".
651 With replace=false, this operation will return an HTTP 409 "Conflict" error
652 if there is already an object at the given location, rather than
653 overwriting the existing object. To allow the operation to overwrite a
654 file, but return an error when trying to overwrite a directory, use
655 "replace=only-files" (this behavior is closer to the traditional unix "mv"
656 command). Note that "true", "t", and "1" are all synonyms for "True", and
657 "false", "f", and "0" are synonyms for "False", and the parameter is
660 === Adding multiple files or directories to a parent directory at once ===
662 POST /uri/$DIRCAP/[SUBDIRS..]?t=set_children
664 This command adds multiple children to a directory in a single operation.
665 It reads the request body and interprets it as a JSON-encoded description
666 of the child names and read/write-caps that should be added.
668 The body should be a JSON-encoded dictionary, in the same format as the
669 "children" value returned by the "GET /uri/$DIRCAP?t=json" operation
670 described above. In this format, each key is a child names, and the
671 corresponding value is a tuple of (type, childinfo). "type" is ignored, and
672 "childinfo" is a dictionary that contains "rw_uri", "ro_uri", and
673 "metadata" keys. You can take the output of "GET /uri/$DIRCAP1?t=json" and
674 use it as the input to "POST /uri/$DIRCAP2?t=set_children" to make DIR2
675 look very much like DIR1.
677 When the set_children request contains a child name that already exists in
678 the target directory, this command defaults to overwriting that child with
679 the new value (both child cap and metadata, but if the JSON data does not
680 contain a "metadata" key, the old child's metadata is preserved). The
681 command takes a boolean "overwrite=" query argument to control this
682 behavior. If you use "?t=set_children&overwrite=false", then an attempt to
683 replace an existing child will instead cause an error.
685 Any "tahoe" key in the new child's "metadata" value is ignored. Any
686 existing "tahoe" metadata is preserved. The metadata["tahoe"] value is
687 reserved for metadata generated by the tahoe node itself. The only two keys
688 currently placed here are "linkcrtime" and "linkmotime". For details, see
689 the section above entitled "Get Information About A File Or Directory (as
690 JSON)", in the "About the metadata" subsection.
693 === Deleting a File or Directory ===
695 DELETE /uri/$DIRCAP/[SUBDIRS../]CHILDNAME
697 This removes the given name from its parent directory. CHILDNAME is the
698 name to be removed, and $DIRCAP/SUBDIRS.. indicates the directory that will
701 Note that this does not actually delete the file or directory that the name
702 points to from the tahoe grid -- it only removes the named reference from
703 this directory. If there are other names in this directory or in other
704 directories that point to the resource, then it will remain accessible
705 through those paths. Even if all names pointing to this object are removed
706 from their parent directories, then someone with possession of its read-cap
707 can continue to access the object through that cap.
709 The object will only become completely unreachable once 1: there are no
710 reachable directories that reference it, and 2: nobody is holding a read-
711 or write- cap to the object. (This behavior is very similar to the way
712 hardlinks and anonymous files work in traditional unix filesystems).
714 This operation will not modify more than a single directory. Intermediate
715 directories which were implicitly created by PUT or POST methods will *not*
716 be automatically removed by DELETE.
718 This method returns the file- or directory- cap of the object that was just
721 == Browser Operations ==
723 This section describes the HTTP operations that provide support for humans
724 running a web browser. Most of these operations use HTML forms that use POST
725 to drive the Tahoe node. This section is intended for HTML authors who want
726 to write web pages that contain forms and buttons which manipulate the Tahoe
729 Note that for all POST operations, the arguments listed can be provided
730 either as URL query arguments or as form body fields. URL query arguments are
731 separated from the main URL by "?", and from each other by "&". For example,
732 "POST /uri/$DIRCAP?t=upload&mutable=true". Form body fields are usually
733 specified by using <input type="hidden"> elements. For clarity, the
734 descriptions below display the most significant arguments as URL query args.
736 === Viewing A Directory (as HTML) ===
738 GET /uri/$DIRCAP/[SUBDIRS../]
740 This returns an HTML page, intended to be displayed to a human by a web
741 browser, which contains HREF links to all files and directories reachable
742 from this directory. These HREF links do not have a t= argument, meaning
743 that a human who follows them will get pages also meant for a human. It also
744 contains forms to upload new files, and to delete files and directories.
745 Those forms use POST methods to do their job.
747 === Viewing/Downloading a File ===
750 GET /uri/$DIRCAP/[SUBDIRS../]FILENAME
752 This will retrieve the contents of the given file. The HTTP response body
753 will contain the sequence of bytes that make up the file.
755 If you want the HTTP response to include a useful Content-Type header,
756 either use the second form (which starts with a $DIRCAP), or add a
757 "filename=foo" query argument, like "GET /uri/$FILECAP?filename=foo.jpg".
758 The bare "GET /uri/$FILECAP" does not give the Tahoe node enough information
759 to determine a Content-Type (since Tahoe immutable files are merely
760 sequences of bytes, not typed+named file objects).
762 If the URL has both filename= and "save=true" in the query arguments, then
763 the server to add a "Content-Disposition: attachment" header, along with a
764 filename= parameter. When a user clicks on such a link, most browsers will
765 offer to let the user save the file instead of displaying it inline (indeed,
766 most browsers will refuse to display it inline). "true", "t", "1", and other
767 case-insensitive equivalents are all treated the same.
769 Character-set handling in URLs and HTTP headers is a dubious art[1]. For
770 maximum compatibility, Tahoe simply copies the bytes from the filename=
771 argument into the Content-Disposition header's filename= parameter, without
772 trying to interpret them in any particular way.
775 GET /named/$FILECAP/FILENAME
777 This is an alternate download form which makes it easier to get the correct
778 filename. The Tahoe server will provide the contents of the given file, with
779 a Content-Type header derived from the given filename. This form is used to
780 get browsers to use the "Save Link As" feature correctly, and also helps
781 command-line tools like "wget" and "curl" use the right filename. Note that
782 this form can *only* be used with file caps; it is an error to use a
783 directory cap after the /named/ prefix.
785 === Get Information About A File Or Directory (as HTML) ===
787 GET /uri/$FILECAP?t=info
788 GET /uri/$DIRCAP/?t=info
789 GET /uri/$DIRCAP/[SUBDIRS../]SUBDIR/?t=info
790 GET /uri/$DIRCAP/[SUBDIRS../]FILENAME?t=info
792 This returns a human-oriented HTML page with more detail about the selected
793 file or directory object. This page contains the following items:
798 raw contents (text/plain)
799 access caps (URIs): verify-cap, read-cap, write-cap (for mutable objects)
800 check/verify/repair form
801 deep-check/deep-size/deep-stats/manifest (for directories)
802 replace-conents form (for mutable files)
804 === Creating a Directory ===
808 This creates a new empty directory, but does not attach it to the virtual
811 If a "redirect_to_result=true" argument is provided, then the HTTP response
812 will cause the web browser to be redirected to a /uri/$DIRCAP page that
813 gives access to the newly-created directory. If you bookmark this page,
814 you'll be able to get back to the directory again in the future. This is the
815 recommended way to start working with a Tahoe server: create a new unlinked
816 directory (using redirect_to_result=true), then bookmark the resulting
817 /uri/$DIRCAP page. There is a "create directory" button on the Welcome page
818 to invoke this action.
820 If "redirect_to_result=true" is not provided (or is given a value of
821 "false"), then the HTTP response body will simply be the write-cap of the
824 POST /uri/$DIRCAP/[SUBDIRS../]?t=mkdir&name=CHILDNAME
826 This creates a new empty directory as a child of the designated SUBDIR. This
827 will create additional intermediate directories as necessary.
829 If a "when_done=URL" argument is provided, the HTTP response will cause the
830 web browser to redirect to the given URL. This provides a convenient way to
831 return the browser to the directory that was just modified. Without a
832 when_done= argument, the HTTP response will simply contain the write-cap of
833 the directory that was just created.
836 === Uploading a File ===
840 This uploads a file, and produces a file-cap for the contents, but does not
841 attach the file into the virtual drive. No directories will be modified by
844 The file must be provided as the "file" field of an HTML encoded form body,
845 produced in response to an HTML form like this:
846 <form action="/uri" method="POST" enctype="multipart/form-data">
847 <input type="hidden" name="t" value="upload" />
848 <input type="file" name="file" />
849 <input type="submit" value="Upload Unlinked" />
852 If a "when_done=URL" argument is provided, the response body will cause the
853 browser to redirect to the given URL. If the when_done= URL has the string
854 "%(uri)s" in it, that string will be replaced by a URL-escaped form of the
855 newly created file-cap. (Note that without this substitution, there is no
856 way to access the file that was just uploaded).
858 The default (in the absence of when_done=) is to return an HTML page that
859 describes the results of the upload. This page will contain information
860 about which storage servers were used for the upload, how long each
863 If a "mutable=true" argument is provided, the operation will create a
864 mutable file, and the response body will contain the write-cap instead of
865 the upload results page. The default is to create an immutable file,
866 returning the upload results page as a response.
869 POST /uri/$DIRCAP/[SUBDIRS../]?t=upload
871 This uploads a file, and attaches it as a new child of the given directory.
872 The file must be provided as the "file" field of an HTML encoded form body,
873 produced in response to an HTML form like this:
874 <form action="." method="POST" enctype="multipart/form-data">
875 <input type="hidden" name="t" value="upload" />
876 <input type="file" name="file" />
877 <input type="submit" value="Upload" />
880 A "name=" argument can be provided to specify the new child's name,
881 otherwise it will be taken from the "filename" field of the upload form
882 (most web browsers will copy the last component of the original file's
883 pathname into this field). To avoid confusion, name= is not allowed to
886 If there is already a child with that name, and it is a mutable file, then
887 its contents are replaced with the data being uploaded. If it is not a
888 mutable file, the default behavior is to remove the existing child before
889 creating a new one. To prevent this (and make the operation return an error
890 instead of overwriting the old child), add a "replace=false" argument, as
891 "?t=upload&replace=false". With replace=false, this operation will return an
892 HTTP 409 "Conflict" error if there is already an object at the given
893 location, rather than overwriting the existing object. Note that "true",
894 "t", and "1" are all synonyms for "True", and "false", "f", and "0" are
895 synonyms for "False". the parameter is case-insensitive.
897 This will create additional intermediate directories as necessary, although
898 since it is expected to be triggered by a form that was retrieved by "GET
899 /uri/$DIRCAP/[SUBDIRS../]", it is likely that the parent directory will
902 If a "mutable=true" argument is provided, any new file that is created will
903 be a mutable file instead of an immutable one. <input type="checkbox"
904 name="mutable" /> will give the user a way to set this option.
906 If a "when_done=URL" argument is provided, the HTTP response will cause the
907 web browser to redirect to the given URL. This provides a convenient way to
908 return the browser to the directory that was just modified. Without a
909 when_done= argument, the HTTP response will simply contain the file-cap of
910 the file that was just uploaded (a write-cap for mutable files, or a
911 read-cap for immutable files).
913 POST /uri/$DIRCAP/[SUBDIRS../]FILENAME?t=upload
915 This also uploads a file and attaches it as a new child of the given
916 directory. It is a slight variant of the previous operation, as the URL
917 refers to the target file rather than the parent directory. It is otherwise
918 identical: this accepts mutable= and when_done= arguments too.
920 POST /uri/$FILECAP?t=upload
922 This modifies the contents of an existing mutable file in-place. An error is
923 signalled if $FILECAP does not refer to a mutable file. It behaves just like
924 the "PUT /uri/$FILECAP" form, but uses a POST for the benefit of HTML forms
927 === Attaching An Existing File Or Directory (by URI) ===
929 POST /uri/$DIRCAP/[SUBDIRS../]?t=uri&name=CHILDNAME&uri=CHILDCAP
931 This attaches a given read- or write- cap "CHILDCAP" to the designated
932 directory, with a specified child name. This behaves much like the PUT t=uri
933 operation, and is a lot like a UNIX hardlink.
935 This will create additional intermediate directories as necessary, although
936 since it is expected to be triggered by a form that was retrieved by "GET
937 /uri/$DIRCAP/[SUBDIRS../]", it is likely that the parent directory will
940 This accepts the same replace= argument as POST t=upload.
942 === Deleting A Child ===
944 POST /uri/$DIRCAP/[SUBDIRS../]?t=delete&name=CHILDNAME
946 This instructs the node to delete a child object (file or subdirectory) from
947 the given directory. Note that the entire subtree is removed. This is
948 somewhat like "rm -rf" (from the point of view of the parent), but other
949 references into the subtree will see that the child subdirectories are not
950 modified by this operation. Only the link from the given directory to its
953 === Renaming A Child ===
955 POST /uri/$DIRCAP/[SUBDIRS../]?t=rename&from_name=OLD&to_name=NEW
957 This instructs the node to rename a child of the given directory. This is
958 exactly the same as removing the child, then adding the same child-cap under
959 the new name. This operation cannot move the child to a different directory.
961 This operation will replace any existing child of the new name, making it
962 behave like the UNIX "mv -f" command.
964 === Other Utilities ===
968 This causes a redirect to /uri/$CAP, and retains any additional query
969 arguments (like filename= or save=). This is for the convenience of web
970 forms which allow the user to paste in a read- or write- cap (obtained
971 through some out-of-band channel, like IM or email).
973 Note that this form merely redirects to the specific file or directory
974 indicated by the $CAP: unlike the GET /uri/$DIRCAP form, you cannot
975 traverse to children by appending additional path segments to the URL.
977 GET /uri/$DIRCAP/[SUBDIRS../]?t=rename-form&name=$CHILDNAME
979 This provides a useful facility to browser-based user interfaces. It
980 returns a page containing a form targetting the "POST $DIRCAP t=rename"
981 functionality described above, with the provided $CHILDNAME present in the
982 'from_name' field of that form. I.e. this presents a form offering to
983 rename $CHILDNAME, requesting the new name, and submitting POST rename.
985 GET /uri/$DIRCAP/[SUBDIRS../]CHILDNAME?t=uri
987 This returns the file- or directory- cap for the specified object.
989 GET /uri/$DIRCAP/[SUBDIRS../]CHILDNAME?t=readonly-uri
991 This returns a read-only file- or directory- cap for the specified object.
992 If the object is an immutable file, this will return the same value as
995 === Debugging and Testing Features ===
997 These URLs are less-likely to be helpful to the casual Tahoe user, and are
998 mainly intended for developers.
1002 This triggers the FileChecker to determine the current "health" of the
1003 given file or directory, by counting how many shares are available. The
1004 page that is returned will display the results. This can be used as a "show
1005 me detailed information about this file" page.
1007 If a verify=true argument is provided, the node will perform a more
1008 intensive check, downloading and verifying every single bit of every share.
1010 If an add-lease=true argument is provided, the node will also add (or
1011 renew) a lease to every share it encounters. Each lease will keep the share
1012 alive for a certain period of time (one month by default). Once the last
1013 lease expires or is explicitly cancelled, the storage server is allowed to
1016 If an output=JSON argument is provided, the response will be
1017 machine-readable JSON instead of human-oriented HTML. The data is a
1018 dictionary with the following keys:
1020 storage-index: a base32-encoded string with the objects's storage index,
1021 or an empty string for LIT files
1022 summary: a string, with a one-line summary of the stats of the file
1023 results: a dictionary that describes the state of the file. For LIT files,
1024 this dictionary has only the 'healthy' key, which will always be
1025 True. For distributed files, this dictionary has the following
1027 count-shares-good: the number of good shares that were found
1028 count-shares-needed: 'k', the number of shares required for recovery
1029 count-shares-expected: 'N', the number of total shares generated
1030 count-good-share-hosts: the number of distinct storage servers with
1031 good shares. If this number is less than
1032 count-shares-good, then some shares are doubled
1033 up, increasing the correlation of failures. This
1034 indicates that one or more shares should be
1035 moved to an otherwise unused server, if one is
1037 count-wrong-shares: for mutable files, the number of shares for
1038 versions other than the 'best' one (highest
1039 sequence number, highest roothash). These are
1041 count-recoverable-versions: for mutable files, the number of
1042 recoverable versions of the file. For
1043 a healthy file, this will equal 1.
1044 count-unrecoverable-versions: for mutable files, the number of
1045 unrecoverable versions of the file.
1046 For a healthy file, this will be 0.
1047 count-corrupt-shares: the number of shares with integrity failures
1048 list-corrupt-shares: a list of "share locators", one for each share
1049 that was found to be corrupt. Each share locator
1050 is a list of (serverid, storage_index, sharenum).
1051 needs-rebalancing: (bool) True if there are multiple shares on a single
1052 storage server, indicating a reduction in reliability
1053 that could be resolved by moving shares to new
1055 servers-responding: list of base32-encoded storage server identifiers,
1056 one for each server which responded to the share
1058 healthy: (bool) True if the file is completely healthy, False otherwise.
1059 Healthy files have at least N good shares. Overlapping shares
1060 (indicated by count-good-share-hosts < count-shares-good) do not
1061 currently cause a file to be marked unhealthy. If there are at
1062 least N good shares, then corrupt shares do not cause the file to
1063 be marked unhealthy, although the corrupt shares will be listed
1064 in the results (list-corrupt-shares) and should be manually
1065 removed to wasting time in subsequent downloads (as the
1066 downloader rediscovers the corruption and uses alternate shares).
1067 sharemap: dict mapping share identifier to list of serverids
1068 (base32-encoded strings). This indicates which servers are
1069 holding which shares. For immutable files, the shareid is
1070 an integer (the share number, from 0 to N-1). For
1071 immutable files, it is a string of the form
1072 'seq%d-%s-sh%d', containing the sequence number, the
1073 roothash, and the share number.
1075 POST $URL?t=start-deep-check (must add &ophandle=XYZ)
1077 This initiates a recursive walk of all files and directories reachable from
1078 the target, performing a check on each one just like t=check. The result
1079 page will contain a summary of the results, including details on any
1080 file/directory that was not fully healthy.
1082 t=start-deep-check can only be invoked on a directory. An error (400
1083 BAD_REQUEST) will be signalled if it is invoked on a file. The recursive
1084 walker will deal with loops safely.
1086 This accepts the same verify= and add-lease= arguments as t=check.
1088 Since this operation can take a long time (perhaps a second per object),
1089 the ophandle= argument is required (see "Slow Operations, Progress, and
1090 Cancelling" above). The response to this POST will be a redirect to the
1091 corresponding /operations/$HANDLE page (with output=HTML or output=JSON to
1092 match the output= argument given to the POST). The deep-check operation
1093 will continue to run in the background, and the /operations page should be
1094 used to find out when the operation is done.
1096 Detailed check results for non-healthy files and directories will be
1097 available under /operations/$HANDLE/$STORAGEINDEX, and the HTML status will
1098 contain links to these detailed results.
1100 The HTML /operations/$HANDLE page for incomplete operations will contain a
1101 meta-refresh tag, set to 60 seconds, so that a browser which uses
1102 deep-check will automatically poll until the operation has completed.
1104 The JSON page (/options/$HANDLE?output=JSON) will contain a
1105 machine-readable JSON dictionary with the following keys:
1107 finished: a boolean, True if the operation is complete, else False. Some
1108 of the remaining keys may not be present until the operation
1110 root-storage-index: a base32-encoded string with the storage index of the
1111 starting point of the deep-check operation
1112 count-objects-checked: count of how many objects were checked. Note that
1113 non-distributed objects (i.e. small immutable LIT
1114 files) are not checked, since for these objects,
1115 the data is contained entirely in the URI.
1116 count-objects-healthy: how many of those objects were completely healthy
1117 count-objects-unhealthy: how many were damaged in some way
1118 count-corrupt-shares: how many shares were found to have corruption,
1119 summed over all objects examined
1120 list-corrupt-shares: a list of "share identifiers", one for each share
1121 that was found to be corrupt. Each share identifier
1122 is a list of (serverid, storage_index, sharenum).
1123 list-unhealthy-files: a list of (pathname, check-results) tuples, for
1124 each file that was not fully healthy. 'pathname' is
1125 a list of strings (which can be joined by "/"
1126 characters to turn it into a single string),
1127 relative to the directory on which deep-check was
1128 invoked. The 'check-results' field is the same as
1129 that returned by t=check&output=JSON, described
1131 stats: a dictionary with the same keys as the t=start-deep-stats command
1134 POST $URL?t=stream-deep-check
1136 This initiates a recursive walk of all files and directories reachable from
1137 the target, performing a check on each one just like t=check. For each
1138 unique object (duplicates are skipped), a single line of JSON is emitted to
1139 the HTTP response channel (or an error indication, see below). When the walk
1140 is complete, a final line of JSON is emitted which contains the accumulated
1141 file-size/count "deep-stats" data.
1143 This command takes the same arguments as t=start-deep-check.
1145 A CLI tool can split the response stream on newlines into "response units",
1146 and parse each response unit as JSON. Each such parsed unit will be a
1147 dictionary, and will contain at least the "type" key: a string, one of
1148 "file", "directory", or "stats".
1150 For all units that have a type of "file" or "directory", the dictionary will
1151 contain the following keys:
1153 "path": a list of strings, with the path that is traversed to reach the
1155 "cap": a writecap for the file or directory, if available, else a readcap
1156 "verifycap": a verifycap for the file or directory
1157 "repaircap": the weakest cap which can still be used to repair the object
1158 "storage-index": a base32 storage index for the object
1159 "check-results": a copy of the dictionary which would be returned by
1160 t=check&output=json, with three top-level keys:
1161 "storage-index", "summary", and "results", and a variety
1162 of counts and sharemaps in the "results" value.
1164 Note that non-distributed files (i.e. LIT files) will have values of None
1165 for verifycap, repaircap, and storage-index, since these files can neither
1166 be verified nor repaired, and are not stored on the storage servers.
1167 Likewise the check-results dictionary will be limited: an empty string for
1168 storage-index, and a results dictionary with only the "healthy" key.
1170 The last unit in the stream will have a type of "stats", and will contain
1171 the keys described in the "start-deep-stats" operation, below.
1173 If any errors occur during the traversal (specifically if a directory is
1174 unrecoverable, such that further traversal is not possible), an error
1175 indication is written to the response body, instead of the usual line of
1176 JSON. This error indication line will begin with the string "ERROR:" (in all
1177 caps), and contain a summary of the error on the rest of the line. The
1178 remaining lines of the response body will be a python exception. The client
1179 application should look for the ERROR: and stop processing JSON as soon as
1180 it is seen. Note that neither a file being unrecoverable nor a directory
1181 merely being unhealthy will cause traversal to stop. The line just before
1182 the ERROR: will describe the directory that was untraversable, since the
1183 unit is emitted to the HTTP response body before the child is traversed.
1186 POST $URL?t=check&repair=true
1188 This performs a health check of the given file or directory, and if the
1189 checker determines that the object is not healthy (some shares are missing
1190 or corrupted), it will perform a "repair". During repair, any missing
1191 shares will be regenerated and uploaded to new servers.
1193 This accepts the same verify=true and add-lease= arguments as t=check. When
1194 an output=JSON argument is provided, the machine-readable JSON response
1195 will contain the following keys:
1197 storage-index: a base32-encoded string with the objects's storage index,
1198 or an empty string for LIT files
1199 repair-attempted: (bool) True if repair was attempted
1200 repair-successful: (bool) True if repair was attempted and the file was
1201 fully healthy afterwards. False if no repair was
1202 attempted, or if a repair attempt failed.
1203 pre-repair-results: a dictionary that describes the state of the file
1204 before any repair was performed. This contains exactly
1205 the same keys as the 'results' value of the t=check
1206 response, described above.
1207 post-repair-results: a dictionary that describes the state of the file
1208 after any repair was performed. If no repair was
1209 performed, post-repair-results and pre-repair-results
1210 will be the same. This contains exactly the same keys
1211 as the 'results' value of the t=check response,
1214 POST $URL?t=start-deep-check&repair=true (must add &ophandle=XYZ)
1216 This triggers a recursive walk of all files and directories, performing a
1217 t=check&repair=true on each one.
1219 Like t=start-deep-check without the repair= argument, this can only be
1220 invoked on a directory. An error (400 BAD_REQUEST) will be signalled if it
1221 is invoked on a file. The recursive walker will deal with loops safely.
1223 This accepts the same verify= and add-lease= arguments as
1224 t=start-deep-check. It uses the same ophandle= mechanism as
1225 start-deep-check. When an output=JSON argument is provided, the response
1226 will contain the following keys:
1228 finished: (bool) True if the operation has completed, else False
1229 root-storage-index: a base32-encoded string with the storage index of the
1230 starting point of the deep-check operation
1231 count-objects-checked: count of how many objects were checked
1233 count-objects-healthy-pre-repair: how many of those objects were completely
1234 healthy, before any repair
1235 count-objects-unhealthy-pre-repair: how many were damaged in some way
1236 count-objects-healthy-post-repair: how many of those objects were completely
1237 healthy, after any repair
1238 count-objects-unhealthy-post-repair: how many were damaged in some way
1240 count-repairs-attempted: repairs were attempted on this many objects.
1241 count-repairs-successful: how many repairs resulted in healthy objects
1242 count-repairs-unsuccessful: how many repairs resulted did not results in
1243 completely healthy objects
1244 count-corrupt-shares-pre-repair: how many shares were found to have
1245 corruption, summed over all objects
1246 examined, before any repair
1247 count-corrupt-shares-post-repair: how many shares were found to have
1248 corruption, summed over all objects
1249 examined, after any repair
1250 list-corrupt-shares: a list of "share identifiers", one for each share
1251 that was found to be corrupt (before any repair).
1252 Each share identifier is a list of (serverid,
1253 storage_index, sharenum).
1254 list-remaining-corrupt-shares: like list-corrupt-shares, but mutable shares
1255 that were successfully repaired are not
1256 included. These are shares that need
1257 manual processing. Since immutable shares
1258 cannot be modified by clients, all corruption
1259 in immutable shares will be listed here.
1260 list-unhealthy-files: a list of (pathname, check-results) tuples, for
1261 each file that was not fully healthy. 'pathname' is
1262 relative to the directory on which deep-check was
1263 invoked. The 'check-results' field is the same as
1264 that returned by t=check&repair=true&output=JSON,
1266 stats: a dictionary with the same keys as the t=start-deep-stats command
1269 POST $URL?t=stream-deep-check&repair=true
1271 This triggers a recursive walk of all files and directories, performing a
1272 t=check&repair=true on each one. For each unique object (duplicates are
1273 skipped), a single line of JSON is emitted to the HTTP response channel (or
1274 an error indication). When the walk is complete, a final line of JSON is
1275 emitted which contains the accumulated file-size/count "deep-stats" data.
1277 This emits the same data as t=stream-deep-check (without the repair=true),
1278 except that the "check-results" field is replaced with a
1279 "check-and-repair-results" field, which contains the keys returned by
1280 t=check&repair=true&output=json (i.e. repair-attempted, repair-successful,
1281 pre-repair-results, and post-repair-results). The output does not contain
1282 the summary dictionary that is provied by t=start-deep-check&repair=true
1283 (the one with count-objects-checked and list-unhealthy-files), since the
1284 receiving client is expected to calculate those values itself from the
1285 stream of per-object check-and-repair-results.
1287 Note that the "ERROR:" indication will only be emitted if traversal stops,
1288 which will only occur if an unrecoverable directory is encountered. If a
1289 file or directory repair fails, the traversal will continue, and the repair
1290 failure will be indicated in the JSON data (in the "repair-successful" key).
1292 POST $DIRURL?t=start-manifest (must add &ophandle=XYZ)
1294 This operation generates a "manfest" of the given directory tree, mostly
1295 for debugging. This is a table of (path, filecap/dircap), for every object
1296 reachable from the starting directory. The path will be slash-joined, and
1297 the filecap/dircap will contain a link to the object in question. This page
1298 gives immediate access to every object in the virtual filesystem subtree.
1300 This operation uses the same ophandle= mechanism as deep-check. The
1301 corresponding /operations/$HANDLE page has three different forms. The
1302 default is output=HTML.
1304 If output=text is added to the query args, the results will be a text/plain
1305 list. The first line is special: it is either "finished: yes" or "finished:
1306 no"; if the operation is not finished, you must periodically reload the
1307 page until it completes. The rest of the results are a plaintext list, with
1308 one file/dir per line, slash-separated, with the filecap/dircap separated
1311 If output=JSON is added to the queryargs, then the results will be a
1312 JSON-formatted dictionary with six keys. Note that because large directory
1313 structures can result in very large JSON results, the full results will not
1314 be available until the operation is complete (i.e. until output["finished"]
1317 finished (bool): if False then you must reload the page until True
1318 origin_si (base32 str): the storage index of the starting point
1319 manifest: list of (path, cap) tuples, where path is a list of strings.
1320 verifycaps: list of (printable) verify cap strings
1321 storage-index: list of (base32) storage index strings
1322 stats: a dictionary with the same keys as the t=start-deep-stats command
1325 POST $DIRURL?t=start-deep-size (must add &ophandle=XYZ)
1327 This operation generates a number (in bytes) containing the sum of the
1328 filesize of all directories and immutable files reachable from the given
1329 directory. This is a rough lower bound of the total space consumed by this
1330 subtree. It does not include space consumed by mutable files, nor does it
1331 take expansion or encoding overhead into account. Later versions of the
1332 code may improve this estimate upwards.
1334 The /operations/$HANDLE status output consists of two lines of text:
1339 POST $DIRURL?t=start-deep-stats (must add &ophandle=XYZ)
1341 This operation performs a recursive walk of all files and directories
1342 reachable from the given directory, and generates a collection of
1343 statistics about those objects.
1345 The result (obtained from the /operations/$OPHANDLE page) is a
1346 JSON-serialized dictionary with the following keys (note that some of these
1347 keys may be missing until 'finished' is True):
1349 finished: (bool) True if the operation has finished, else False
1350 count-immutable-files: count of how many CHK files are in the set
1351 count-mutable-files: same, for mutable files (does not include directories)
1352 count-literal-files: same, for LIT files (data contained inside the URI)
1353 count-files: sum of the above three
1354 count-directories: count of directories
1355 count-unknown: count of unrecognized objects (perhaps from the future)
1356 size-immutable-files: total bytes for all CHK files in the set, =deep-size
1357 size-mutable-files (TODO): same, for current version of all mutable files
1358 size-literal-files: same, for LIT files
1359 size-directories: size of directories (includes size-literal-files)
1360 size-files-histogram: list of (minsize, maxsize, count) buckets,
1361 with a histogram of filesizes, 5dB/bucket,
1362 for both literal and immutable files
1363 largest-directory: number of children in the largest directory
1364 largest-immutable-file: number of bytes in the largest CHK file
1366 size-mutable-files is not implemented, because it would require extra
1367 queries to each mutable file to get their size. This may be implemented in
1370 Assuming no sharing, the basic space consumed by a single root directory is
1371 the sum of size-immutable-files, size-mutable-files, and size-directories.
1372 The actual disk space used by the shares is larger, because of the
1373 following sources of overhead:
1376 expansion due to erasure coding
1377 share management data (leases)
1378 backend (ext3) minimum block size
1380 POST $URL?t=stream-manifest
1382 This operation performs a recursive walk of all files and directories
1383 reachable from the given starting point. For each such unique object
1384 (duplicates are skipped), a single line of JSON is emitted to the HTTP
1385 response channel (or an error indication, see below). When the walk is
1386 complete, a final line of JSON is emitted which contains the accumulated
1387 file-size/count "deep-stats" data.
1389 A CLI tool can split the response stream on newlines into "response units",
1390 and parse each response unit as JSON. Each such parsed unit will be a
1391 dictionary, and will contain at least the "type" key: a string, one of
1392 "file", "directory", or "stats".
1394 For all units that have a type of "file" or "directory", the dictionary will
1395 contain the following keys:
1397 "path": a list of strings, with the path that is traversed to reach the
1399 "cap": a writecap for the file or directory, if available, else a readcap
1400 "verifycap": a verifycap for the file or directory
1401 "repaircap": the weakest cap which can still be used to repair the object
1402 "storage-index": a base32 storage index for the object
1404 Note that non-distributed files (i.e. LIT files) will have values of None
1405 for verifycap, repaircap, and storage-index, since these files can neither
1406 be verified nor repaired, and are not stored on the storage servers.
1408 The last unit in the stream will have a type of "stats", and will contain
1409 the keys described in the "start-deep-stats" operation, below.
1411 If any errors occur during the traversal (specifically if a directory is
1412 unrecoverable, such that further traversal is not possible), an error
1413 indication is written to the response body, instead of the usual line of
1414 JSON. This error indication line will begin with the string "ERROR:" (in all
1415 caps), and contain a summary of the error on the rest of the line. The
1416 remaining lines of the response body will be a python exception. The client
1417 application should look for the ERROR: and stop processing JSON as soon as
1418 it is seen. The line just before the ERROR: will describe the directory that
1419 was untraversable, since the manifest entry is emitted to the HTTP response
1420 body before the child is traversed.
1422 == Other Useful Pages ==
1424 The portion of the web namespace that begins with "/uri" (and "/named") is
1425 dedicated to giving users (both humans and programs) access to the Tahoe
1426 virtual filesystem. The rest of the namespace provides status information
1427 about the state of the Tahoe node.
1429 GET / (the root page)
1431 This is the "Welcome Page", and contains a few distinct sections:
1433 Node information: library versions, local nodeid, services being provided.
1435 Filesystem Access Forms: create a new directory, view a file/directory by
1436 URI, upload a file (unlinked), download a file by
1439 Grid Status: introducer information, helper information, connected storage
1444 This page lists all active uploads and downloads, and contains a short list
1445 of recent upload/download operations. Each operation has a link to a page
1446 that describes file sizes, servers that were involved, and the time consumed
1447 in each phase of the operation.
1449 A GET of /status/?t=json will contain a machine-readable subset of the same
1450 data. It returns a JSON-encoded dictionary. The only key defined at this
1451 time is "active", with a value that is a list of operation dictionaries, one
1452 for each active operation. Once an operation is completed, it will no longer
1453 appear in data["active"] .
1455 Each op-dict contains a "type" key, one of "upload", "download",
1456 "mapupdate", "publish", or "retrieve" (the first two are for immutable
1457 files, while the latter three are for mutable files and directories).
1459 The "upload" op-dict will contain the following keys:
1461 type (string): "upload"
1462 storage-index-string (string): a base32-encoded storage index
1463 total-size (int): total size of the file
1464 status (string): current status of the operation
1465 progress-hash (float): 1.0 when the file has been hashed
1466 progress-ciphertext (float): 1.0 when the file has been encrypted.
1467 progress-encode-push (float): 1.0 when the file has been encoded and
1468 pushed to the storage servers. For helper
1469 uploads, the ciphertext value climbs to 1.0
1470 first, then encoding starts. For unassisted
1471 uploads, ciphertext and encode-push progress
1472 will climb at the same pace.
1474 The "download" op-dict will contain the following keys:
1476 type (string): "download"
1477 storage-index-string (string): a base32-encoded storage index
1478 total-size (int): total size of the file
1479 status (string): current status of the operation
1480 progress (float): 1.0 when the file has been fully downloaded
1482 Front-ends which want to report progress information are advised to simply
1483 average together all the progress-* indicators. A slightly more accurate
1484 value can be found by ignoring the progress-hash value (since the current
1485 implementation hashes synchronously, so clients will probably never see
1486 progress-hash!=1.0).
1490 This page provides a basic tool to predict the likely storage and bandwidth
1491 requirements of a large Tahoe grid. It provides forms to input things like
1492 total number of users, number of files per user, average file size, number
1493 of servers, expansion ratio, hard drive failure rate, etc. It then provides
1494 numbers like how many disks per server will be needed, how many read
1495 operations per second should be expected, and the likely MTBF for files in
1496 the grid. This information is very preliminary, and the model upon which it
1497 is based still needs a lot of work.
1501 If the node is running a helper (i.e. if [helper]enabled is set to True in
1502 tahoe.cfg), then this page will provide a list of all the helper operations
1503 currently in progress. If "?t=json" is added to the URL, it will return a
1504 JSON-formatted list of helper statistics, which can then be used to produce
1505 graphs to indicate how busy the helper is.
1509 This page provides "node statistics", which are collected from a variety of
1512 load_monitor: every second, the node schedules a timer for one second in
1513 the future, then measures how late the subsequent callback
1514 is. The "load_average" is this tardiness, measured in
1515 seconds, averaged over the last minute. It is an indication
1516 of a busy node, one which is doing more work than can be
1517 completed in a timely fashion. The "max_load" value is the
1518 highest value that has been seen in the last 60 seconds.
1520 cpu_monitor: every minute, the node uses time.clock() to measure how much
1521 CPU time it has used, and it uses this value to produce
1522 1min/5min/15min moving averages. These values range from 0%
1523 (0.0) to 100% (1.0), and indicate what fraction of the CPU
1524 has been used by the Tahoe node. Not all operating systems
1525 provide meaningful data to time.clock(): they may report 100%
1526 CPU usage at all times.
1528 uploader: this counts how many immutable files (and bytes) have been
1529 uploaded since the node was started
1531 downloader: this counts how many immutable files have been downloaded
1532 since the node was started
1534 publishes: this counts how many mutable files (including directories) have
1535 been modified since the node was started
1537 retrieves: this counts how many mutable files (including directories) have
1538 been read since the node was started
1540 There are other statistics that are tracked by the node. The "raw stats"
1541 section shows a formatted dump of all of them.
1543 By adding "?t=json" to the URL, the node will return a JSON-formatted
1544 dictionary of stats values, which can be used by other tools to produce
1545 graphs of node behavior. The misc/munin/ directory in the source
1546 distribution provides some tools to produce these graphs.
1548 GET / (introducer status)
1550 For Introducer nodes, the welcome page displays information about both
1551 clients and servers which are connected to the introducer. Servers make
1552 "service announcements", and these are listed in a table. Clients will
1553 subscribe to hear about service announcements, and these subscriptions are
1554 listed in a separate table. Both tables contain information about what
1555 version of Tahoe is being run by the remote node, their advertised and
1556 outbound IP addresses, their nodeid and nickname, and how long they have
1559 By adding "?t=json" to the URL, the node will return a JSON-formatted
1560 dictionary of stats values, which can be used to produce graphs of connected
1561 clients over time. This dictionary has the following keys:
1563 ["subscription_summary"] : a dictionary mapping service name (like
1564 "storage") to an integer with the number of
1565 clients that have subscribed to hear about that
1567 ["announcement_summary"] : a dictionary mapping service name to an integer
1568 with the number of servers which are announcing
1570 ["announcement_distinct_hosts"] : a dictionary mapping service name to an
1571 integer which represents the number of
1572 distinct hosts that are providing that
1573 service. If two servers have announced
1574 FURLs which use the same hostnames (but
1575 different ports and tubids), they are
1576 considered to be on the same host.
1579 == Static Files in /public_html ==
1581 The wapi server will take any request for a URL that starts with /static
1582 and serve it from a configurable directory which defaults to
1583 $BASEDIR/public_html . This is configured by setting the "[node]web.static"
1584 value in $BASEDIR/tahoe.cfg . If this is left at the default value of
1585 "public_html", then http://localhost:3456/static/subdir/foo.html will be
1586 served with the contents of the file $BASEDIR/public_html/subdir/foo.html .
1588 This can be useful to serve a javascript application which provides a
1589 prettier front-end to the rest of the Tahoe wapi.
1592 == safety and security issues -- names vs. URIs ==
1594 Summary: use explicit file- and dir- caps whenever possible, to reduce the
1595 potential for surprises when the filesystem structure is changed.
1597 Tahoe provides a mutable filesystem, but the ways that the filesystem can
1598 change are limited. The only thing that can change is that the mapping from
1599 child names to child objects that each directory contains can be changed by
1600 adding a new child name pointing to an object, removing an existing child name,
1601 or changing an existing child name to point to a different object.
1603 Obviously if you query Tahoe for information about the filesystem and then act
1604 to change the filesystem (such as by getting a listing of the contents of a
1605 directory and then adding a file to the directory), then the filesystem might
1606 have been changed after you queried it and before you acted upon it. However,
1607 if you use the URI instead of the pathname of an object when you act upon the
1608 object, then the only change that can happen is if the object is a directory
1609 then the set of child names it has might be different. If, on the other hand,
1610 you act upon the object using its pathname, then a different object might be in
1611 that place, which can result in more kinds of surprises.
1613 For example, suppose you are writing code which recursively downloads the
1614 contents of a directory. The first thing your code does is fetch the listing
1615 of the contents of the directory. For each child that it fetched, if that
1616 child is a file then it downloads the file, and if that child is a directory
1617 then it recurses into that directory. Now, if the download and the recurse
1618 actions are performed using the child's name, then the results might be
1619 wrong, because for example a child name that pointed to a sub-directory when
1620 you listed the directory might have been changed to point to a file (in which
1621 case your attempt to recurse into it would result in an error and the file
1622 would be skipped), or a child name that pointed to a file when you listed the
1623 directory might now point to a sub-directory (in which case your attempt to
1624 download the child would result in a file containing HTML text describing the
1627 If your recursive algorithm uses the uri of the child instead of the name of
1628 the child, then those kinds of mistakes just can't happen. Note that both the
1629 child's name and the child's URI are included in the results of listing the
1630 parent directory, so it isn't any harder to use the URI for this purpose.
1632 In general, use names if you want "whatever object (whether file or
1633 directory) is found by following this name (or sequence of names) when my
1634 request reaches the server". Use URIs if you want "this particular object".
1636 == Concurrency Issues ==
1638 Tahoe uses both mutable and immutable files. Mutable files can be created
1639 explicitly by doing an upload with ?mutable=true added, or implicitly by
1640 creating a new directory (since a directory is just a special way to
1641 interpret a given mutable file).
1643 Mutable files suffer from the same consistency-vs-availability tradeoff that
1644 all distributed data storage systems face. It is not possible to
1645 simultaneously achieve perfect consistency and perfect availability in the
1646 face of network partitions (servers being unreachable or faulty).
1648 Tahoe tries to achieve a reasonable compromise, but there is a basic rule in
1649 place, known as the Prime Coordination Directive: "Don't Do That". What this
1650 means is that if write-access to a mutable file is available to several
1651 parties, then those parties are responsible for coordinating their activities
1652 to avoid multiple simultaneous updates. This could be achieved by having
1653 these parties talk to each other and using some sort of locking mechanism, or
1654 by serializing all changes through a single writer.
1656 The consequences of performing uncoordinated writes can vary. Some of the
1657 writers may lose their changes, as somebody else wins the race condition. In
1658 many cases the file will be left in an "unhealthy" state, meaning that there
1659 are not as many redundant shares as we would like (reducing the reliability
1660 of the file against server failures). In the worst case, the file can be left
1661 in such an unhealthy state that no version is recoverable, even the old ones.
1662 It is this small possibility of data loss that prompts us to issue the Prime
1663 Coordination Directive.
1665 Tahoe nodes implement internal serialization to make sure that a single Tahoe
1666 node cannot conflict with itself. For example, it is safe to issue two
1667 directory modification requests to a single tahoe node's wapi server at the
1668 same time, because the Tahoe node will internally delay one of them until
1669 after the other has finished being applied. (This feature was introduced in
1670 Tahoe-1.1; back with Tahoe-1.0 the web client was responsible for serializing
1671 web requests themselves).
1673 For more details, please see the "Consistency vs Availability" and "The Prime
1674 Coordination Directive" sections of mutable.txt, in the same directory as
1678 [1]: URLs and HTTP and UTF-8, Oh My
1680 HTTP does not provide a mechanism to specify the character set used to
1681 encode non-ascii names in URLs (rfc2396#2.1). We prefer the convention that
1682 the filename= argument shall be a URL-encoded UTF-8 encoded unicode object.
1683 For example, suppose we want to provoke the server into using a filename of
1684 "f i a n c e-acute e" (i.e. F I A N C U+00E9 E). The UTF-8 encoding of this
1685 is 0x66 0x69 0x61 0x6e 0x63 0xc3 0xa9 0x65 (or "fianc\xC3\xA9e", as python's
1686 repr() function would show). To encode this into a URL, the non-printable
1687 characters must be escaped with the urlencode '%XX' mechansim, giving us
1688 "fianc%C3%A9e". Thus, the first line of the HTTP request will be "GET
1689 /uri/CAP...?save=true&filename=fianc%C3%A9e HTTP/1.1". Not all browsers
1690 provide this: IE7 uses the Latin-1 encoding, which is fianc%E9e.
1692 The response header will need to indicate a non-ASCII filename. The actual
1693 mechanism to do this is not clear. For ASCII filenames, the response header
1696 Content-Disposition: attachment; filename="english.txt"
1698 If Tahoe were to enforce the utf-8 convention, it would need to decode the
1699 URL argument into a unicode string, and then encode it back into a sequence
1700 of bytes when creating the response header. One possibility would be to use
1701 unencoded utf-8. Developers suggest that IE7 might accept this:
1703 #1: Content-Disposition: attachment; filename="fianc\xC3\xA9e"
1704 (note, the last four bytes of that line, not including the newline, are
1705 0xC3 0xA9 0x65 0x22)
1707 RFC2231#4 (dated 1997): suggests that the following might work, and some
1708 developers (http://markmail.org/message/dsjyokgl7hv64ig3) have reported that
1709 it is supported by firefox (but not IE7):
1711 #2: Content-Disposition: attachment; filename*=utf-8''fianc%C3%A9e
1713 My reading of RFC2616#19.5.1 (which defines Content-Disposition) says that
1714 the filename= parameter is defined to be wrapped in quotes (presumeably to
1715 allow spaces without breaking the parsing of subsequent parameters), which
1718 #3: Content-Disposition: attachment; filename*=utf-8''"fianc%C3%A9e"
1720 However this is contrary to the examples in the email thread listed above.
1722 Developers report that IE7 (when it is configured for UTF-8 URL encoding,
1723 which is not the default in asian countries), will accept:
1725 #4: Content-Disposition: attachment; filename=fianc%C3%A9e
1727 However, for maximum compatibility, Tahoe simply copies bytes from the URL
1728 into the response header, rather than enforcing the utf-8 convention. This
1729 means it does not try to decode the filename from the URL argument, nor does
1730 it encode the filename into the response header.