|
1 ============================== |
|
2 How to use Django with FastCGI |
|
3 ============================== |
|
4 |
|
5 Although the `current preferred setup`_ for running Django is Apache_ with |
|
6 `mod_python`_, many people use shared hosting, on which FastCGI is the only |
|
7 viable option. In some setups, FastCGI also allows better security -- and, |
|
8 possibly, better performance -- than mod_python. |
|
9 |
|
10 Essentially, FastCGI is an efficient way of letting an external application |
|
11 serve pages to a Web server. The Web server delegates the incoming Web requests |
|
12 (via a socket) to FastCGI, which executes the code and passes the response back |
|
13 to the Web server, which, in turn, passes it back to the client's Web browser. |
|
14 |
|
15 Like mod_python, FastCGI allows code to stay in memory, allowing requests to be |
|
16 served with no startup time. Unlike mod_python (or `mod_perl`_), a FastCGI |
|
17 process doesn't run inside the Web server process, but in a separate, |
|
18 persistent process. |
|
19 |
|
20 .. _current preferred setup: ../modpython/ |
|
21 .. _Apache: http://httpd.apache.org/ |
|
22 .. _mod_python: http://www.modpython.org/ |
|
23 .. _mod_perl: http://perl.apache.org/ |
|
24 |
|
25 .. admonition:: Why run code in a separate process? |
|
26 |
|
27 The traditional ``mod_*`` arrangements in Apache embed various scripting |
|
28 languages (most notably PHP, Python and Perl) inside the process space of |
|
29 your Web server. Although this lowers startup time -- because code doesn't |
|
30 have to be read off disk for every request -- it comes at the cost of |
|
31 memory use. For mod_python, for example, every Apache process gets its own |
|
32 Python interpreter, which uses up a considerable amount of RAM. |
|
33 |
|
34 Due to the nature of FastCGI, it's even possible to have processes that run |
|
35 under a different user account than the Web server process. That's a nice |
|
36 security benefit on shared systems, because it means you can secure your |
|
37 code from other users. |
|
38 |
|
39 Prerequisite: flup |
|
40 ================== |
|
41 |
|
42 Before you can start using FastCGI with Django, you'll need to install flup_, |
|
43 which is a Python library for dealing with FastCGI. Make sure to use the latest |
|
44 Subversion snapshot of flup, as some users have reported stalled pages with |
|
45 older flup versions. |
|
46 |
|
47 .. _flup: http://www.saddi.com/software/flup/ |
|
48 |
|
49 Starting your FastCGI server |
|
50 ============================ |
|
51 |
|
52 FastCGI operates on a client-server model, and in most cases you'll be starting |
|
53 the FastCGI process on your own. Your Web server (be it Apache, lighttpd, or |
|
54 otherwise) only contacts your Django-FastCGI process when the server needs a |
|
55 dynamic page to be loaded. Because the daemon is already running with the code |
|
56 in memory, it's able to serve the response very quickly. |
|
57 |
|
58 .. admonition:: Note |
|
59 |
|
60 If you're on a shared hosting system, you'll probably be forced to use |
|
61 Web server-managed FastCGI processes. See the section below on running |
|
62 Django with Web server-managed processes for more information. |
|
63 |
|
64 A Web server can connect to a FastCGI server in one of two ways: It can use |
|
65 either a Unix domain socket (a "named pipe" on Win32 systems), or it can use a |
|
66 TCP socket. What you choose is a manner of preference; a TCP socket is usually |
|
67 easier due to permissions issues. |
|
68 |
|
69 To start your server, first change into the directory of your project (wherever |
|
70 your ``manage.py`` is), and then run ``manage.py`` with the ``runfcgi`` option:: |
|
71 |
|
72 ./manage.py runfcgi [options] |
|
73 |
|
74 If you specify ``help`` as the only option after ``runfcgi``, it'll display a |
|
75 list of all the available options. |
|
76 |
|
77 You'll need to specify either a ``socket`` or both ``host`` and ``port``. Then, |
|
78 when you set up your Web server, you'll just need to point it at the host/port |
|
79 or socket you specified when starting the FastCGI server. |
|
80 |
|
81 Examples |
|
82 -------- |
|
83 |
|
84 Running a threaded server on a TCP port:: |
|
85 |
|
86 ./manage.py runfcgi method=threaded host=127.0.0.1 port=3033 |
|
87 |
|
88 Running a preforked server on a Unix domain socket:: |
|
89 |
|
90 ./manage.py runfcgi method=prefork socket=/home/user/mysite.sock pidfile=django.pid |
|
91 |
|
92 Run without daemonizing (backgrounding) the process (good for debugging):: |
|
93 |
|
94 ./manage.py runfcgi daemonize=false socket=/tmp/mysite.sock |
|
95 |
|
96 Stopping the FastCGI daemon |
|
97 --------------------------- |
|
98 |
|
99 If you have the process running in the foreground, it's easy enough to stop it: |
|
100 Simply hitting ``Ctrl-C`` will stop and quit the FastCGI server. However, when |
|
101 you're dealing with background processes, you'll need to resort to the Unix |
|
102 ``kill`` command. |
|
103 |
|
104 If you specify the ``pidfile`` option to your ``manage.py runfcgi``, you can |
|
105 kill the running FastCGI daemon like this:: |
|
106 |
|
107 kill `cat $PIDFILE` |
|
108 |
|
109 ...where ``$PIDFILE`` is the ``pidfile`` you specified. |
|
110 |
|
111 To easily restart your FastCGI daemon on Unix, try this small shell script:: |
|
112 |
|
113 #!/bin/bash |
|
114 |
|
115 # Replace these three settings. |
|
116 PROJDIR="/home/user/myproject" |
|
117 PIDFILE="$PROJDIR/mysite.pid" |
|
118 SOCKET="$PROJDIR/mysite.sock" |
|
119 |
|
120 cd $PROJDIR |
|
121 if [ -f $PIDFILE ]; then |
|
122 kill `cat -- $PIDFILE` |
|
123 rm -f -- $PIDFILE |
|
124 fi |
|
125 |
|
126 exec /usr/bin/env - \ |
|
127 PYTHONPATH="../python:.." \ |
|
128 ./manage.py runfcgi socket=$SOCKET pidfile=$PIDFILE |
|
129 |
|
130 Apache setup |
|
131 ============ |
|
132 |
|
133 To use Django with Apache and FastCGI, you'll need Apache installed and |
|
134 configured, with `mod_fastcgi`_ installed and enabled. Consult the Apache |
|
135 documentation for instructions. |
|
136 |
|
137 Once you've got that set up, point Apache at your Django FastCGI instance by |
|
138 editing the ``httpd.conf`` (Apache configuration) file. You'll need to do two |
|
139 things: |
|
140 |
|
141 * Use the ``FastCGIExternalServer`` directive to specify the location of |
|
142 your FastCGI server. |
|
143 * Use ``mod_rewrite`` to point URLs at FastCGI as appropriate. |
|
144 |
|
145 .. _mod_fastcgi: http://www.fastcgi.com/mod_fastcgi/docs/mod_fastcgi.html |
|
146 |
|
147 Specifying the location of the FastCGI server |
|
148 --------------------------------------------- |
|
149 |
|
150 The ``FastCGIExternalServer`` directive tells Apache how to find your FastCGI |
|
151 server. As the `FastCGIExternalServer docs`_ explain, you can specify either a |
|
152 ``socket`` or a ``host``. Here are examples of both:: |
|
153 |
|
154 # Connect to FastCGI via a socket / named pipe. |
|
155 FastCGIExternalServer /home/user/public_html/mysite.fcgi -socket /home/user/mysite.sock |
|
156 |
|
157 # Connect to FastCGI via a TCP host/port. |
|
158 FastCGIExternalServer /home/user/public_html/mysite.fcgi -host 127.0.0.1:3033 |
|
159 |
|
160 In either case, the file ``/home/user/public_html/mysite.fcgi`` doesn't |
|
161 actually have to exist. It's just a URL used by the Web server internally -- a |
|
162 hook for signifying which requests at a URL should be handled by FastCGI. (More |
|
163 on this in the next section.) |
|
164 |
|
165 .. _FastCGIExternalServer docs: http://www.fastcgi.com/mod_fastcgi/docs/mod_fastcgi.html#FastCgiExternalServer |
|
166 |
|
167 Using mod_rewrite to point URLs at FastCGI |
|
168 ------------------------------------------ |
|
169 |
|
170 The second step is telling Apache to use FastCGI for URLs that match a certain |
|
171 pattern. To do this, use the `mod_rewrite`_ module and rewrite URLs to |
|
172 ``mysite.fcgi`` (or whatever you specified in the ``FastCGIExternalServer`` |
|
173 directive, as explained in the previous section). |
|
174 |
|
175 In this example, we tell Apache to use FastCGI to handle any request that |
|
176 doesn't represent a file on the filesystem and doesn't start with ``/media/``. |
|
177 This is probably the most common case, if you're using Django's admin site:: |
|
178 |
|
179 <VirtualHost 12.34.56.78> |
|
180 ServerName example.com |
|
181 DocumentRoot /home/user/public_html |
|
182 Alias /media /home/user/python/django/contrib/admin/media |
|
183 RewriteEngine On |
|
184 RewriteRule ^/(media.*)$ /$1 [QSA,L] |
|
185 RewriteCond %{REQUEST_FILENAME} !-f |
|
186 RewriteRule ^/(.*)$ /mysite.fcgi/$1 [QSA,L] |
|
187 </VirtualHost> |
|
188 |
|
189 .. _mod_rewrite: http://httpd.apache.org/docs/2.0/mod/mod_rewrite.html |
|
190 |
|
191 lighttpd setup |
|
192 ============== |
|
193 |
|
194 lighttpd is a lightweight Web server commonly used for serving static files. It |
|
195 supports FastCGI natively and, thus, is a good choice for serving both static |
|
196 and dynamic pages, if your site doesn't have any Apache-specific needs. |
|
197 |
|
198 Make sure ``mod_fastcgi`` is in your modules list, somewhere after |
|
199 ``mod_rewrite`` and ``mod_access``, but not after ``mod_accesslog``. You'll |
|
200 probably want ``mod_alias`` as well, for serving admin media. |
|
201 |
|
202 Add the following to your lighttpd config file:: |
|
203 |
|
204 server.document-root = "/home/user/public_html" |
|
205 fastcgi.server = ( |
|
206 "/mysite.fcgi" => ( |
|
207 "main" => ( |
|
208 # Use host / port instead of socket for TCP fastcgi |
|
209 # "host" => "127.0.0.1", |
|
210 # "port" => 3033, |
|
211 "socket" => "/home/user/mysite.sock", |
|
212 "check-local" => "disable", |
|
213 ) |
|
214 ), |
|
215 ) |
|
216 alias.url = ( |
|
217 "/media/" => "/home/user/django/contrib/admin/media/", |
|
218 ) |
|
219 |
|
220 url.rewrite-once = ( |
|
221 "^(/media.*)$" => "$1", |
|
222 "^/favicon\.ico$" => "/media/favicon.ico", |
|
223 "^(/.*)$" => "/mysite.fcgi$1", |
|
224 ) |
|
225 |
|
226 Running multiple Django sites on one lighttpd |
|
227 --------------------------------------------- |
|
228 |
|
229 lighttpd lets you use "conditional configuration" to allow configuration to be |
|
230 customized per host. To specify multiple FastCGI sites, just add a conditional |
|
231 block around your FastCGI config for each site:: |
|
232 |
|
233 # If the hostname is 'www.example1.com'... |
|
234 $HTTP["host"] == "www.example1.com" { |
|
235 server.document-root = "/foo/site1" |
|
236 fastcgi.server = ( |
|
237 ... |
|
238 ) |
|
239 ... |
|
240 } |
|
241 |
|
242 # If the hostname is 'www.example2.com'... |
|
243 $HTTP["host"] == "www.example2.com" { |
|
244 server.document-root = "/foo/site2" |
|
245 fastcgi.server = ( |
|
246 ... |
|
247 ) |
|
248 ... |
|
249 } |
|
250 |
|
251 You can also run multiple Django installations on the same site simply by |
|
252 specifying multiple entries in the ``fastcgi.server`` directive. Add one |
|
253 FastCGI host for each. |
|
254 |
|
255 Running Django on a shared-hosting provider with Apache |
|
256 ======================================================= |
|
257 |
|
258 Many shared-hosting providers don't allow you to run your own server daemons or |
|
259 edit the ``httpd.conf`` file. In these cases, it's still possible to run Django |
|
260 using Web server-spawned processes. |
|
261 |
|
262 .. admonition:: Note |
|
263 |
|
264 If you're using Web server-spawned processes, as explained in this section, |
|
265 there's no need for you to start the FastCGI server on your own. Apache |
|
266 will spawn a number of processes, scaling as it needs to. |
|
267 |
|
268 In your Web root directory, add this to a file named ``.htaccess`` :: |
|
269 |
|
270 AddHandler fastcgi-script .fcgi |
|
271 RewriteEngine On |
|
272 RewriteCond %{REQUEST_FILENAME} !-f |
|
273 RewriteRule ^(.*)$ mysite.fcgi/$1 [QSA,L] |
|
274 |
|
275 Then, create a small script that tells Apache how to spawn your FastCGI |
|
276 program. Create a file ``mysite.fcgi`` and place it in your Web directory, and |
|
277 be sure to make it executable:: |
|
278 |
|
279 #!/usr/bin/python |
|
280 import sys, os |
|
281 |
|
282 # Add a custom Python path. |
|
283 sys.path.insert(0, "/home/user/python") |
|
284 |
|
285 # Switch to the directory of your project. (Optional.) |
|
286 # os.chdir("/home/user/myproject") |
|
287 |
|
288 # Set the DJANGO_SETTINGS_MODULE environment variable. |
|
289 os.environ['DJANGO_SETTINGS_MODULE'] = "myproject.settings" |
|
290 |
|
291 from django.core.servers.fastcgi import runfastcgi |
|
292 runfastcgi(method="threaded", daemonize="false") |
|
293 |
|
294 Restarting the spawned server |
|
295 ----------------------------- |
|
296 |
|
297 If you change any Python code on your site, you'll need to tell FastCGI the |
|
298 code has changed. But there's no need to restart Apache in this case. Rather, |
|
299 just reupload ``mysite.fcgi``, or edit the file, so that the timestamp on the |
|
300 file will change. When Apache sees the file has been updated, it will restart |
|
301 your Django application for you. |
|
302 |
|
303 If you have access to a command shell on a Unix system, you can accomplish this |
|
304 easily by using the ``touch`` command:: |
|
305 |
|
306 touch mysite.fcgi |
|
307 |
|
308 Serving admin media files |
|
309 ========================= |
|
310 |
|
311 Regardless of the server and configuration you eventually decide to use, you will |
|
312 also need to give some thought to how to serve the admin media files. The |
|
313 advice given in the modpython_ documentation is also applicable in the setups |
|
314 detailed above. |
|
315 |
|
316 .. _modpython: ../modpython/#serving-the-admin-files |
|
317 |