2 <!DOCTYPE modulesynopsis SYSTEM "../style/modulesynopsis.dtd">
3 <?xml-stylesheet type="text/xsl" href="../style/manual.en.xsl"?>
4 <!-- $LastChangedRevision$ -->
7 Licensed to the Apache Software Foundation (ASF) under one or more
8 contributor license agreements. See the NOTICE file distributed with
9 this work for additional information regarding copyright ownership.
10 The ASF licenses this file to You under the Apache License, Version 2.0
11 (the "License"); you may not use this file except in compliance with
12 the License. You may obtain a copy of the License at
14 http://www.apache.org/licenses/LICENSE-2.0
16 Unless required by applicable law or agreed to in writing, software
17 distributed under the License is distributed on an "AS IS" BASIS,
18 WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
19 See the License for the specific language governing permissions and
20 limitations under the License.
23 <modulesynopsis metafile="mod_proxy_ajp.xml.meta">
25 <name>mod_proxy_ajp</name>
26 <description>AJP support module for
27 <module>mod_proxy</module></description>
28 <status>Extension</status>
29 <sourcefile>mod_proxy_ajp.c</sourcefile>
30 <identifier>proxy_ajp_module</identifier>
31 <compatibility>Available in version 2.1 and later</compatibility>
34 <p>This module <em>requires</em> the service of <module
35 >mod_proxy</module>. It provides support for the
36 <code>Apache JServ Protocol version 1.3</code> (hereafter
39 <p>Thus, in order to get the ability of handling <code>AJP13</code>
40 protocol, <module>mod_proxy</module> and
41 <module>mod_proxy_ajp</module> have to be present in the server.</p>
43 <note type="warning"><title>Warning</title>
44 <p>Do not enable proxying until you have <a
45 href="mod_proxy.html#access">secured your server</a>. Open proxy
46 servers are dangerous both to your network and to the Internet at
51 <seealso><module>mod_proxy</module></seealso>
52 <seealso><a href="../env.html">Environment Variable documentation</a></seealso>
54 <section id="usage"><title>Usage</title>
55 <p>This module is used to reverse proxy to a backend application server
56 (e.g. Apache Tomcat) using the AJP13 protocol. The usage is similar to
57 an HTTP reverse proxy, but uses the <code>ajp://</code> prefix:</p>
59 <example><title>Simple Reverse Proxy</title>
60 <highlight language="config">
61 ProxyPass /app ajp://backend.example.com:8009/app
65 <p>Balancers may also be used:</p>
66 <example><title>Balancer Reverse Proxy</title>
67 <highlight language="config">
68 <Proxy balancer://cluster>
69 BalancerMember ajp://app1.example.com:8009 loadfactor=1
70 BalancerMember ajp://app2.example.com:8009 loadfactor=2
71 ProxySet lbmethod=bytraffic
73 ProxyPass /app balancer://cluster/app
77 <p>Note that usually no
78 <directive module="mod_proxy">ProxyPassReverse</directive>
79 directive is necessary. The AJP request includes the original host
80 header given to the proxy, and the application server can be expected
81 to generate self-referential headers relative to this host, so no
82 rewriting is necessary.</p>
84 <p>The main exception is when the URL path on the proxy differs from that
86 backend. In this case, a redirect header can be rewritten relative to the
87 original host URL (not the backend <code>ajp://</code> URL), for
89 <example><title>Rewriting Proxied Path</title>
90 <highlight language="config">
91 ProxyPass /apps/foo ajp://backend.example.com:8009/foo
92 ProxyPassReverse /apps/foo http://www.example.com/foo
95 <p>However, it is usually better to deploy the application on the backend
96 server at the same path as the proxy rather than to take this approach.
100 <section id="env"><title>Environment Variables</title>
101 <p>Environment variables whose names have the prefix <code>AJP_</code>
102 are forwarded to the origin server as AJP request attributes
103 (with the AJP_ prefix removed from the name of the key).</p>
106 <section id="overviewprotocol"><title>Overview of the protocol</title>
107 <p>The <code>AJP13</code> protocol is packet-oriented. A binary format
108 was presumably chosen over the more readable plain text for reasons of
109 performance. The web server communicates with the servlet container over
110 TCP connections. To cut down on the expensive process of socket creation,
111 the web server will attempt to maintain persistent TCP connections to the
112 servlet container, and to reuse a connection for multiple request/response
114 <p>Once a connection is assigned to a particular request, it will not be
115 used for any others until the request-handling cycle has terminated. In
116 other words, requests are not multiplexed over connections. This makes
117 for much simpler code at either end of the connection, although it does
118 cause more connections to be open at once.</p>
119 <p>Once the web server has opened a connection to the servlet container,
120 the connection can be in one of the following states:</p>
122 <li> Idle <br/> No request is being handled over this connection. </li>
123 <li> Assigned <br/> The connection is handling a specific request.</li>
125 <p>Once a connection is assigned to handle a particular request, the basic
126 request information (e.g. HTTP headers, etc) is sent over the connection in
127 a highly condensed form (e.g. common strings are encoded as integers).
128 Details of that format are below in Request Packet Structure. If there is a
129 body to the request <code>(content-length > 0)</code>, that is sent in a
130 separate packet immediately after.</p>
131 <p>At this point, the servlet container is presumably ready to start
132 processing the request. As it does so, it can send the
133 following messages back to the web server:</p>
135 <li>SEND_HEADERS <br/>Send a set of headers back to the browser.</li>
136 <li>SEND_BODY_CHUNK <br/>Send a chunk of body data back to the browser.
138 <li>GET_BODY_CHUNK <br/>Get further data from the request if it hasn't all
139 been transferred yet. This is necessary because the packets have a fixed
140 maximum size and arbitrary amounts of data can be included the body of a
141 request (for uploaded files, for example). (Note: this is unrelated to
142 HTTP chunked transfer).</li>
143 <li>END_RESPONSE <br/> Finish the request-handling cycle.</li>
145 <p>Each message is accompanied by a differently formatted packet of data.
146 See Response Packet Structures below for details.</p>
149 <section id="basppacketstruct"><title>Basic Packet Structure</title>
150 <p>There is a bit of an XDR heritage to this protocol, but it differs
151 in lots of ways (no 4 byte alignment, for example).</p>
152 <p>Byte order: I am not clear about the endian-ness of the individual
153 bytes. I'm guessing the bytes are little-endian, because that's what
154 XDR specifies, and I'm guessing that sys/socket library is magically
155 making that so (on the C side). If anyone with a better knowledge of
156 socket calls can step in, that would be great.</p>
157 <p>There are four data types in the protocol: bytes, booleans,
158 integers and strings.</p>
160 <dt><strong>Byte</strong></dt><dd>A single byte.</dd>
161 <dt><strong>Boolean</strong></dt>
162 <dd>A single byte, <code>1 = true</code>, <code>0 = false</code>.
163 Using other non-zero values as true (i.e. C-style) may work in some places,
164 but it won't in others.</dd>
165 <dt><strong>Integer</strong></dt>
166 <dd>A number in the range of <code>0 to 2^16 (32768)</code>. Stored in
167 2 bytes with the high-order byte first.</dd>
168 <dt><strong>String</strong></dt>
169 <dd>A variable-sized string (length bounded by 2^16). Encoded with
170 the length packed into two bytes first, followed by the string
171 (including the terminating '\0'). Note that the encoded length does
172 <strong>not</strong> include the trailing '\0' -- it is like
173 <code>strlen</code>. This is a touch confusing on the Java side, which
174 is littered with odd autoincrement statements to skip over these
175 terminators. I believe the reason this was done was to allow the C
176 code to be extra efficient when reading strings which the servlet
177 container is sending back -- with the terminating \0 character, the
178 C code can pass around references into a single buffer, without copying.
179 if the \0 was missing, the C code would have to copy things out in order
180 to get its notion of a string.</dd>
183 <section><title>Packet Size</title>
184 <p>According to much of the code, the max packet size is <code>
185 8 * 1024 bytes (8K)</code>. The actual length of the packet is encoded in
188 <section><title>Packet Headers</title>
189 <p>Packets sent from the server to the container begin with
190 <code>0x1234</code>. Packets sent from the container to the server
191 begin with <code>AB</code> (that's the ASCII code for A followed by the
192 ASCII code for B). After those first two bytes, there is an integer
193 (encoded as above) with the length of the payload. Although this might
194 suggest that the maximum payload could be as large as 2^16, in fact, the
195 code sets the maximum to be 8K.</p>
198 <td colspan="6"><em>Packet Format (Server->Container)</em></td>
212 <td colspan="2">Data Length (n)</td>
218 <td colspan="6"><em>Packet Format (Container->Server)</em></td>
232 <td colspan="2">Data Length (n)</td>
236 <p>For most packets, the first byte of the payload encodes the type of
237 message. The exception is for request body packets sent from the server to
238 the container -- they are sent with a standard packet header (<code>
239 0x1234</code> and then length of the packet), but without any prefix code
241 <p>The web server can send the following messages to the servlet
246 <td>Type of Packet</td>
251 <td>Forward Request</td>
252 <td>Begin the request-processing cycle with the following data</td>
257 <td>The web server asks the container to shut itself down.</td>
262 <td>The web server asks the container to take control
263 (secure login phase).</td>
268 <td>The web server asks the container to respond quickly with a CPong.
274 <td>Size (2 bytes) and corresponding body data.</td>
277 <p>To ensure some basic security, the container will only actually do the
278 <code>Shutdown</code> if the request comes from the same machine on which
280 <p>The first <code>Data</code> packet is send immediately after the
281 <code>Forward Request</code> by the web server.</p>
282 <p>The servlet container can send the following types of messages to the
287 <td>Type of Packet</td>
292 <td>Send Body Chunk</td>
293 <td>Send a chunk of the body from the servlet container to the web
294 server (and presumably, onto the browser). </td>
298 <td>Send Headers</td>
299 <td>Send the response headers from the servlet container to the web
300 server (and presumably, onto the browser).</td>
304 <td>End Response</td>
305 <td>Marks the end of the response (and thus the request-handling cycle).
310 <td>Get Body Chunk</td>
311 <td>Get further data from the request if it hasn't all been
312 transferred yet.</td>
317 <td>The reply to a CPing request</td>
320 <p>Each of the above messages has a different internal structure, detailed
324 <section id="rpacetstruct"><title>Request Packet Structure</title>
325 <p>For messages from the server to the container of type
326 <em>Forward Request</em>:</p>
328 AJP13_FORWARD_REQUEST :=
329 prefix_code (byte) 0x02 = JK_AJP13_FORWARD_REQUEST
336 server_port (integer)
338 num_headers (integer)
339 request_headers *(req_header_name req_header_value)
340 attributes *(attribut_name attribute_value)
341 request_terminator (byte) OxFF
343 <p>The <code>request_headers</code> have the following structure:
346 sc_req_header_name | (string) [see below for how this is parsed]
348 sc_req_header_name := 0xA0xx (integer)
350 req_header_value := (string)
352 <p>The <code>attributes</code> are optional and have the following
355 attribute_name := sc_a_name | (sc_a_req_attribute string)
357 attribute_value := (string)
360 <p>Not that the all-important header is <code>content-length</code>,
361 because it determines whether or not the container looks for another
362 packet immediately.</p>
363 <section><title>Detailed description of the elements of Forward Request
365 <section><title>Request prefix</title>
366 <p>For all requests, this will be 2. See above for details on other Prefix
369 <section><title>Method</title>
370 <p>The HTTP method, encoded as a single byte:</p>
372 <tr><td>Command Name</td><td>Code</td></tr>
373 <tr><td>OPTIONS</td><td>1</td></tr>
374 <tr><td>GET</td><td>2</td></tr>
375 <tr><td>HEAD</td><td>3</td></tr>
376 <tr><td>POST</td><td>4</td></tr>
377 <tr><td>PUT</td><td>5</td></tr>
378 <tr><td>DELETE</td><td>6</td></tr>
379 <tr><td>TRACE</td><td>7</td></tr>
380 <tr><td>PROPFIND</td><td>8</td></tr>
381 <tr><td>PROPPATCH</td><td>9</td></tr>
382 <tr><td>MKCOL</td><td>10</td></tr>
383 <tr><td>COPY</td><td>11</td></tr>
384 <tr><td>MOVE</td><td>12</td></tr>
385 <tr><td>LOCK</td><td>13</td></tr>
386 <tr><td>UNLOCK</td><td>14</td></tr>
387 <tr><td>ACL</td><td>15</td></tr>
388 <tr><td>REPORT</td><td>16</td></tr>
389 <tr><td>VERSION-CONTROL</td><td>17</td></tr>
390 <tr><td>CHECKIN</td><td>18</td></tr>
391 <tr><td>CHECKOUT</td><td>19</td></tr>
392 <tr><td>UNCHECKOUT</td><td>20</td></tr>
393 <tr><td>SEARCH</td><td>21</td></tr>
394 <tr><td>MKWORKSPACE</td><td>22</td></tr>
395 <tr><td>UPDATE</td><td>23</td></tr>
396 <tr><td>LABEL</td><td>24</td></tr>
397 <tr><td>MERGE</td><td>25</td></tr>
398 <tr><td>BASELINE_CONTROL</td><td>26</td></tr>
399 <tr><td>MKACTIVITY</td><td>27</td></tr>
401 <p>Later version of ajp13, will transport
402 additional methods, even if they are not in this list.</p>
404 <section><title>protocol, req_uri, remote_addr, remote_host, server_name,
405 server_port, is_ssl</title>
406 <p>These are all fairly self-explanatory. Each of these is required, and
407 will be sent for every request.</p>
409 <section><title>Headers</title>
410 <p>The structure of <code>request_headers</code> is the following:
411 First, the number of headers <code>num_headers</code> is encoded.
412 Then, a series of header name <code>req_header_name</code> / value
413 <code>req_header_value</code> pairs follows.
414 Common header names are encoded as integers,
415 to save space. If the header name is not in the list of basic headers,
416 it is encoded normally (as a string, with prefixed length). The list of
417 common headers <code>sc_req_header_name</code>and their codes
418 is as follows (all are case-sensitive):</p>
420 <tr><td>Name</td><td>Code value</td><td>Code name</td></tr>
421 <tr><td>accept</td><td>0xA001</td><td>SC_REQ_ACCEPT</td></tr>
422 <tr><td>accept-charset</td><td>0xA002</td><td>SC_REQ_ACCEPT_CHARSET
424 <tr><td>accept-encoding</td><td>0xA003</td><td>SC_REQ_ACCEPT_ENCODING
426 <tr><td>accept-language</td><td>0xA004</td><td>SC_REQ_ACCEPT_LANGUAGE
428 <tr><td>authorization</td><td>0xA005</td><td>SC_REQ_AUTHORIZATION</td>
430 <tr><td>connection</td><td>0xA006</td><td>SC_REQ_CONNECTION</td></tr>
431 <tr><td>content-type</td><td>0xA007</td><td>SC_REQ_CONTENT_TYPE</td>
433 <tr><td>content-length</td><td>0xA008</td><td>SC_REQ_CONTENT_LENGTH</td>
435 <tr><td>cookie</td><td>0xA009</td><td>SC_REQ_COOKIE</td></tr>
436 <tr><td>cookie2</td><td>0xA00A</td><td>SC_REQ_COOKIE2</td></tr>
437 <tr><td>host</td><td>0xA00B</td><td>SC_REQ_HOST</td></tr>
438 <tr><td>pragma</td><td>0xA00C</td><td>SC_REQ_PRAGMA</td></tr>
439 <tr><td>referer</td><td>0xA00D</td><td>SC_REQ_REFERER</td></tr>
440 <tr><td>user-agent</td><td>0xA00E</td><td>SC_REQ_USER_AGENT</td></tr>
442 <p>The Java code that reads this grabs the first two-byte integer and if
443 it sees an <code>'0xA0'</code> in the most significant
444 byte, it uses the integer in the second byte as an index into an array of
445 header names. If the first byte is not <code>0xA0</code>, it assumes that
446 the two-byte integer is the length of a string, which is then read in.</p>
447 <p>This works on the assumption that no header names will have length
448 greater than <code>0x9999 (==0xA000 - 1)</code>, which is perfectly
449 reasonable, though somewhat arbitrary.</p>
450 <note><title>Note:</title>
451 The <code>content-length</code> header is extremely
452 important. If it is present and non-zero, the container assumes that
453 the request has a body (a POST request, for example), and immediately
454 reads a separate packet off the input stream to get that body.
457 <section><title>Attributes</title>
458 <p>The attributes prefixed with a <code>?</code>
459 (e.g. <code>?context</code>) are all optional. For each, there is a
460 single byte code to indicate the type of attribute, and then its value
461 (string or integer). They can be sent in any order (though the C code
462 always sends them in the order listed below). A special terminating code
463 is sent to signal the end of the list of optional attributes. The list of
466 <tr><td>Information</td><td>Code Value</td><td>Type Of Value</td><td>Note</td></tr>
467 <tr><td>?context</td><td>0x01</td><td>-</td><td>Not currently implemented
469 <tr><td>?servlet_path</td><td>0x02</td><td>-</td><td>Not currently implemented
471 <tr><td>?remote_user</td><td>0x03</td><td>String</td><td></td></tr>
472 <tr><td>?auth_type</td><td>0x04</td><td>String</td><td></td></tr>
473 <tr><td>?query_string</td><td>0x05</td><td>String</td><td></td></tr>
474 <tr><td>?jvm_route</td><td>0x06</td><td>String</td><td></td></tr>
475 <tr><td>?ssl_cert</td><td>0x07</td><td>String</td><td></td></tr>
476 <tr><td>?ssl_cipher</td><td>0x08</td><td>String</td><td></td></tr>
477 <tr><td>?ssl_session</td><td>0x09</td><td>String</td><td></td></tr>
478 <tr><td>?req_attribute</td><td>0x0A</td><td>String</td><td>Name (the name of the
479 attribute follows)</td></tr>
480 <tr><td>?ssl_key_size</td><td>0x0B</td><td>Integer</td><td></td></tr>
481 <tr><td>are_done</td><td>0xFF</td><td>-</td><td>request_terminator</td></tr>
483 <p>The <code>context</code> and <code>servlet_path</code> are not
484 currently set by the C code, and most of the Java code completely ignores
485 whatever is sent over for those fields (and some of it will actually break
486 if a string is sent along after one of those codes). I don't know if this
487 is a bug or an unimplemented feature or just vestigial code, but it's
488 missing from both sides of the connection.</p>
489 <p>The <code>remote_user</code> and <code>auth_type</code> presumably
490 refer to HTTP-level authentication, and communicate the remote user's
491 username and the type of authentication used to establish their identity
492 (e.g. Basic, Digest).</p>
493 <p>The <code>query_string</code>, <code>ssl_cert</code>,
494 <code>ssl_cipher</code>, and <code>ssl_session</code> refer to the
495 corresponding pieces of HTTP and HTTPS.</p>
496 <p>The <code>jvm_route</code>, is used to support sticky
497 sessions -- associating a user's sesson with a particular Tomcat instance
498 in the presence of multiple, load-balancing servers.</p>
499 <p>Beyond this list of basic attributes, any number of other attributes
500 can be sent via the <code>req_attribute</code> code <code>0x0A</code>.
501 A pair of strings to represent the attribute name and value are sent
502 immediately after each instance of that code. Environment values are passed
503 in via this method.</p>
504 <p>Finally, after all the attributes have been sent, the attribute
505 terminator, <code>0xFF</code>, is sent. This signals both the end of the
506 list of attributes and also then end of the Request Packet.</p>
510 <section id="resppacketstruct"><title>Response Packet Structure</title>
511 <p>for messages which the container can send back to the server.</p>
513 AJP13_SEND_BODY_CHUNK :=
515 chunk_length (integer)
517 chunk_terminator (byte) Ox00
520 AJP13_SEND_HEADERS :=
522 http_status_code (integer)
523 http_status_msg (string)
524 num_headers (integer)
525 response_headers *(res_header_name header_value)
528 sc_res_header_name | (string) [see below for how this is parsed]
530 sc_res_header_name := 0xA0 (byte)
532 header_value := (string)
534 AJP13_END_RESPONSE :=
539 AJP13_GET_BODY_CHUNK :=
541 requested_length (integer)
543 <section><title>Details:</title></section>
544 <section><title>Send Body Chunk</title>
545 <p>The chunk is basically binary data, and is sent directly back to the
548 <section><title>Send Headers</title>
549 <p>The status code and message are the usual HTTP things
550 (e.g. <code>200</code> and <code>OK</code>). The response header names are
551 encoded the same way the request header names are. See header_encoding above
552 for details about how the codes are distinguished from the strings.<br />
553 The codes for common headers are:</p>
555 <tr><td>Name</td><td>Code value</td></tr>
556 <tr><td>Content-Type</td><td>0xA001</td></tr>
557 <tr><td>Content-Language</td><td>0xA002</td></tr>
558 <tr><td>Content-Length</td><td>0xA003</td></tr>
559 <tr><td>Date</td><td>0xA004</td></tr>
560 <tr><td>Last-Modified</td><td>0xA005</td></tr>
561 <tr><td>Location</td><td>0xA006</td></tr>
562 <tr><td>Set-Cookie</td><td>0xA007</td></tr>
563 <tr><td>Set-Cookie2</td><td>0xA008</td></tr>
564 <tr><td>Servlet-Engine</td><td>0xA009</td></tr>
565 <tr><td>Status</td><td>0xA00A</td></tr>
566 <tr><td>WWW-Authenticate</td><td>0xA00B</td></tr>
568 <p> After the code or the string header name, the header value is
569 immediately encoded.</p>
571 <section><title>End Response</title>
572 <p>Signals the end of this request-handling cycle. If the
573 <code>reuse</code> flag is true <code>(==1)</code>, this TCP connection can
574 now be used to handle new incoming requests. If <code>reuse</code> is false
575 (anything other than 1 in the actual C code), the connection should
578 <section><title>Get Body Chunk</title>
579 <p>The container asks for more data from the request (If the body was
580 too large to fit in the first packet sent over or when the request is
581 chunked). The server will send a body packet back with an amount of data
582 which is the minimum of the <code>request_length</code>, the maximum send
583 body size <code>(8186 (8 Kbytes - 6))</code>, and the number of bytes
584 actually left to send from the request body.<br/>
585 If there is no more data in the body (i.e. the servlet container is
586 trying to read past the end of the body), the server will send back an
587 <em>empty</em> packet, which is a body packet with a payload length of 0.
588 <code>(0x12,0x34,0x00,0x00)</code></p>