blob: 424a1f8e24c22fb61fe32de9b51930f8927615dd [file] [log] [blame]
Willy Tarreaub1a34b62010-05-09 22:37:12 +02001 ----------------------
2 HAProxy how-to
3 ----------------------
Willy Tarreau9229f122014-06-19 21:01:06 +02004 version 1.5
Willy Tarreaub1a34b62010-05-09 22:37:12 +02005 willy tarreau
Willy Tarreau9229f122014-06-19 21:01:06 +02006 2014/06/19
willy tarreau78345332005-12-18 01:33:16 +01007
8
Willy Tarreaub1a34b62010-05-09 22:37:12 +020091) How to build it
10------------------
11
Willy Tarreau869f3512014-06-19 15:26:32 +020012First, please note that this version is a stable version, so in general if you
13are not used to build from sources, it is recommended that instead you follow
14the packaged updates provided by your software vendor or Linux distribution.
15Most of them are taking this task seriously and are doing a good job. If for
16any reason you'd prefer a different version than the one packaged for your
17system, or to get some commercial support, other choices are available at :
18
19 http://www.haproxy.com/
20
willy tarreau78345332005-12-18 01:33:16 +010021To build haproxy, you will need :
Willy Tarreaub1a34b62010-05-09 22:37:12 +020022 - GNU make. Neither Solaris nor OpenBSD's make work with the GNU Makefile.
Willy Tarreau3543cdb2014-05-10 09:12:46 +020023 If you get many syntax errors when running "make", you may want to retry
24 with "gmake" which is the name commonly used for GNU make on BSD systems.
Willy Tarreau869f3512014-06-19 15:26:32 +020025 - GCC between 2.95 and 4.8. Others may work, but not tested.
willy tarreau78345332005-12-18 01:33:16 +010026 - GNU ld
27
28Also, you might want to build with libpcre support, which will provide a very
Willy Tarreaub1a34b62010-05-09 22:37:12 +020029efficient regex implementation and will also fix some badness on Solaris' one.
willy tarreau78345332005-12-18 01:33:16 +010030
31To build haproxy, you have to choose your target OS amongst the following ones
32and assign it to the TARGET variable :
33
Willy Tarreau83b30c12008-05-25 10:32:50 +020034 - linux22 for Linux 2.2
35 - linux24 for Linux 2.4 and above (default)
36 - linux24e for Linux 2.4 with support for a working epoll (> 0.21)
Willy Tarreau83b30c12008-05-25 10:32:50 +020037 - linux26 for Linux 2.6 and above
Willy Tarreau869f3512014-06-19 15:26:32 +020038 - linux2628 for Linux 2.6.28, 3.x, and above (enables splice and tproxy)
Willy Tarreau83b30c12008-05-25 10:32:50 +020039 - solaris for Solaris 8 or 10 (others untested)
Willy Tarreau869f3512014-06-19 15:26:32 +020040 - freebsd for FreeBSD 5 to 10 (others untested)
Willy Tarreau8624cab2013-04-02 08:17:43 +020041 - osx for Mac OS/X
Willy Tarreau3b8e9792012-11-22 00:43:09 +010042 - openbsd for OpenBSD 3.1 to 5.2 (others untested)
Willy Tarreau50abe302014-04-02 20:44:43 +020043 - aix51 for AIX 5.1
Willy Tarreau7dec9652012-06-06 16:15:03 +020044 - aix52 for AIX 5.2
Yitzhak Sapir32087312009-06-14 18:27:54 +020045 - cygwin for Cygwin
Willy Tarreau869f3512014-06-19 15:26:32 +020046 - generic for any other OS or version.
Willy Tarreau83b30c12008-05-25 10:32:50 +020047 - custom to manually adjust every setting
willy tarreau78345332005-12-18 01:33:16 +010048
49You may also choose your CPU to benefit from some optimizations. This is
50particularly important on UltraSparc machines. For this, you can assign
51one of the following choices to the CPU variable :
52
53 - i686 for intel PentiumPro, Pentium 2 and above, AMD Athlon
54 - i586 for intel Pentium, AMD K6, VIA C3.
55 - ultrasparc : Sun UltraSparc I/II/III/IV processor
Willy Tarreaua5899aa2010-11-28 07:41:00 +010056 - native : use the build machine's specific processor optimizations
willy tarreau78345332005-12-18 01:33:16 +010057 - generic : any other processor or no specific optimization. (default)
58
Willy Tarreau83b30c12008-05-25 10:32:50 +020059Alternatively, you may just set the CPU_CFLAGS value to the optimal GCC options
60for your platform.
61
Willy Tarreauef7341d2009-04-11 19:45:50 +020062You may want to build specific target binaries which do not match your native
63compiler's target. This is particularly true on 64-bit systems when you want
64to build a 32-bit binary. Use the ARCH variable for this purpose. Right now
Willy Tarreaua5899aa2010-11-28 07:41:00 +010065it only knows about a few x86 variants (i386,i486,i586,i686,x86_64), two
66generic ones (32,64) and sets -m32/-m64 as well as -march=<arch> accordingly.
Willy Tarreauef7341d2009-04-11 19:45:50 +020067
willy tarreau78345332005-12-18 01:33:16 +010068If your system supports PCRE (Perl Compatible Regular Expressions), then you
69really should build with libpcre which is between 2 and 10 times faster than
70other libc implementations. Regex are used for header processing (deletion,
71rewriting, allow, deny). The only inconvenient of libpcre is that it is not
72yet widely spread, so if you build for other systems, you might get into
73trouble if they don't have the dynamic library. In this situation, you should
74statically link libpcre into haproxy so that it will not be necessary to
Willy Tarreau83b30c12008-05-25 10:32:50 +020075install it on target systems. Available build options for PCRE are :
willy tarreau78345332005-12-18 01:33:16 +010076
Willy Tarreau83b30c12008-05-25 10:32:50 +020077 - USE_PCRE=1 to use libpcre, in whatever form is available on your system
willy tarreau78345332005-12-18 01:33:16 +010078 (shared or static)
79
Willy Tarreau83b30c12008-05-25 10:32:50 +020080 - USE_STATIC_PCRE=1 to use a static version of libpcre even if the dynamic
81 one is available. This will enhance portability.
82
Willy Tarreau663148c2012-12-12 00:38:22 +010083 - with no option, use your OS libc's standard regex implementation (default).
Willy Tarreau83b30c12008-05-25 10:32:50 +020084 Warning! group references on Solaris seem broken. Use static-pcre whenever
85 possible.
willy tarreau78345332005-12-18 01:33:16 +010086
Willy Tarreau64bc40b2011-03-23 20:00:53 +010087Recent systems can resolve IPv6 host names using getaddrinfo(). This primitive
88is not present in all libcs and does not work in all of them either. Support in
89glibc was broken before 2.3. Some embedded libs may not properly work either,
90thus, support is disabled by default, meaning that some host names which only
91resolve as IPv6 addresses will not resolve and configs might emit an error
92during parsing. If you know that your OS libc has reliable support for
93getaddrinfo(), you can add USE_GETADDRINFO=1 on the make command line to enable
94it. This is the recommended option for most Linux distro packagers since it's
95working fine on all recent mainstream distros. It is automatically enabled on
96Solaris 8 and above, as it's known to work.
97
Willy Tarreau3543cdb2014-05-10 09:12:46 +020098It is possible to add native support for SSL using the GNU makefile, by passing
99"USE_OPENSSL=1" on the make command line. The libssl and libcrypto will
100automatically be linked with haproxy. Some systems also require libz, so if the
101build fails due to missing symbols such as deflateInit(), then try again with
102"ADDLIB=-lz".
Willy Tarreaud4508812012-09-10 09:07:41 +0200103
Lukas Tribus3fe9f1e2013-05-19 16:28:17 +0200104To link OpenSSL statically against haproxy, build OpenSSL with the no-shared
105keyword and install it to a local directory, so your system is not affected :
106
107 $ export STATICLIBSSL=/tmp/staticlibssl
108 $ ./config --prefix=$STATICLIBSSL no-shared
109 $ make && make install_sw
110
Lukas Tribus130ddf72013-10-01 00:28:03 +0200111When building haproxy, pass that path via SSL_INC and SSL_LIB to make and
112include additional libs with ADDLIB if needed (in this case for example libdl):
Willy Tarreau3543cdb2014-05-10 09:12:46 +0200113
Lukas Tribus130ddf72013-10-01 00:28:03 +0200114 $ make TARGET=linux26 USE_OPENSSL=1 SSL_INC=$STATICLIBSSL/include SSL_LIB=$STATICLIBSSL/lib ADDLIB=-ldl
Lukas Tribus3fe9f1e2013-05-19 16:28:17 +0200115
William Lallemand82fe75c2012-10-23 10:25:10 +0200116It is also possible to include native support for ZLIB to benefit from HTTP
117compression. For this, pass "USE_ZLIB=1" on the "make" command line and ensure
118that zlib is present on the system.
119
willy tarreau78345332005-12-18 01:33:16 +0100120By default, the DEBUG variable is set to '-g' to enable debug symbols. It is
121not wise to disable it on uncommon systems, because it's often the only way to
122get a complete core when you need one. Otherwise, you can set DEBUG to '-s' to
123strip the binary.
124
125For example, I use this to build for Solaris 8 :
126
Willy Tarreau83b30c12008-05-25 10:32:50 +0200127 $ make TARGET=solaris CPU=ultrasparc USE_STATIC_PCRE=1
willy tarreau78345332005-12-18 01:33:16 +0100128
Willy Tarreau83b30c12008-05-25 10:32:50 +0200129And I build it this way on OpenBSD or FreeBSD :
willy tarreaud38e72d2006-03-19 20:56:52 +0100130
Willy Tarreau3543cdb2014-05-10 09:12:46 +0200131 $ gmake TARGET=freebsd USE_PCRE=1 USE_OPENSSL=1 USE_ZLIB=1
willy tarreaud38e72d2006-03-19 20:56:52 +0100132
Willy Tarreau663148c2012-12-12 00:38:22 +0100133And on a classic Linux with SSL and ZLIB support (eg: Red Hat 5.x) :
134
135 $ make TARGET=linux26 CPU=native USE_PCRE=1 USE_OPENSSL=1 USE_ZLIB=1
136
137And on a recent Linux >= 2.6.28 with SSL and ZLIB support :
Willy Tarreaud4508812012-09-10 09:07:41 +0200138
William Lallemand82fe75c2012-10-23 10:25:10 +0200139 $ make TARGET=linux2628 CPU=native USE_PCRE=1 USE_OPENSSL=1 USE_ZLIB=1
Willy Tarreaud4508812012-09-10 09:07:41 +0200140
William Lallemand82fe75c2012-10-23 10:25:10 +0200141In order to build a 32-bit binary on an x86_64 Linux system with SSL support
142without support for compression but when OpenSSL requires ZLIB anyway :
Willy Tarreauef7341d2009-04-11 19:45:50 +0200143
Willy Tarreaud4508812012-09-10 09:07:41 +0200144 $ make TARGET=linux26 ARCH=i386 USE_OPENSSL=1 ADDLIB=-lz
Willy Tarreauef7341d2009-04-11 19:45:50 +0200145
Willy Tarreaub1efede2014-05-09 00:44:48 +0200146The SSL stack supports session cache synchronization between all running
147processes. This involves some atomic operations and synchronization operations
148which come in multiple flavors depending on the system and architecture :
149
150 Atomic operations :
151 - internal assembler versions for x86/x86_64 architectures
152
153 - gcc builtins for other architectures. Some architectures might not
154 be fully supported or might require a more recent version of gcc.
155 If your architecture is not supported, you willy have to either use
156 pthread if supported, or to disable the shared cache.
157
158 - pthread (posix threads). Pthreads are very common but inter-process
159 support is not that common, and some older operating systems did not
160 report an error when enabling multi-process mode, so they used to
161 silently fail, possibly causing crashes. Linux's implementation is
162 fine. OpenBSD doesn't support them and doesn't build. FreeBSD 9 builds
163 and reports an error at runtime, while certain older versions might
164 silently fail. Pthreads are enabled using USE_PTHREAD_PSHARED=1.
165
166 Synchronization operations :
167 - internal spinlock : this mode is OS-independant, light but will not
168 scale well to many processes. However, accesses to the session cache
169 are rare enough that this mode could certainly always be used. This
170 is the default mode.
171
172 - Futexes, which are Linux-specific highly scalable light weight mutexes
173 implemented in user-space with some limited assistance from the kernel.
174 This is the default on Linux 2.6 and above and is enabled by passing
175 USE_FUTEX=1
176
177 - pthread (posix threads). See above.
178
179If none of these mechanisms is supported by your platform, you may need to
180build with USE_PRIVATE_CACHE=1 to totally disable SSL cache sharing. Then
181it is better not to run SSL on multiple processes.
182
willy tarreau78345332005-12-18 01:33:16 +0100183If you need to pass other defines, includes, libraries, etc... then please
184check the Makefile to see which ones will be available in your case, and
Willy Tarreau3543cdb2014-05-10 09:12:46 +0200185use the USE_* variables in the Makefile.
willy tarreau78345332005-12-18 01:33:16 +0100186
Willy Tarreau97ec9692010-01-28 20:52:05 +0100187AIX 5.3 is known to work with the generic target. However, for the binary to
188also run on 5.2 or earlier, you need to build with DEFINE="-D_MSGQSUPPORT",
Willy Tarreau869f3512014-06-19 15:26:32 +0200189otherwise __fd_select() will be used while not being present in the libc, but
190this is easily addressed using the "aix52" target. If you get build errors
191because of strange symbols or section mismatches, simply remove -g from
192DEBUG_CFLAGS.
Willy Tarreau97ec9692010-01-28 20:52:05 +0100193
Willy Tarreau32e65ef2013-04-02 08:14:29 +0200194You can easily define your own target with the GNU Makefile. Unknown targets
195are processed with no default option except USE_POLL=default. So you can very
196well use that property to define your own set of options. USE_POLL can even be
197disabled by setting USE_POLL="". For example :
198
199 $ gmake TARGET=tiny USE_POLL="" TARGET_CFLAGS=-fomit-frame-pointer
200
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200201
2022) How to install it
203--------------------
204
205To install haproxy, you can either copy the single resulting binary to the
206place you want, or run :
207
208 $ sudo make install
209
210If you're packaging it for another system, you can specify its root directory
211in the usual DESTDIR variable.
212
213
2143) How to set it up
215-------------------
216
217There is some documentation in the doc/ directory :
218
219 - architecture.txt : this is the architecture manual. It is quite old and
220 does not tell about the nice new features, but it's still a good starting
221 point when you know what you want but don't know how to do it.
222
223 - configuration.txt : this is the configuration manual. It recalls a few
224 essential HTTP basic concepts, and details all the configuration file
225 syntax (keywords, units). It also describes the log and stats format. It
226 is normally always up to date. If you see that something is missing from
Willy Tarreau74774c02014-04-23 00:57:08 +0200227 it, please report it as this is a bug. Please note that this file is
228 huge and that it's generally more convenient to review Cyril Bonté's
229 HTML translation online here :
230
231 http://cbonte.github.io/haproxy-dconv/configuration-1.5.html
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200232
233 - haproxy-en.txt / haproxy-fr.txt : these are the old outdated docs. You
234 should never need them. If you do, then please report what you didn't
235 find in the other ones.
236
237 - gpl.txt / lgpl.txt : the copy of the licenses covering the software. See
238 the 'LICENSE' file at the top for more information.
239
240 - the rest is mainly for developers.
241
242There are also a number of nice configuration examples in the "examples"
243directory as well as on several sites and articles on the net which are linked
244to from the haproxy web site.
245
246
2474) How to report a bug
248----------------------
249
250It is possible that from time to time you'll find a bug. A bug is a case where
251what you see is not what is documented. Otherwise it can be a misdesign. If you
252find that something is stupidly design, please discuss it on the list (see the
253"how to contribute" section below). If you feel like you're proceeding right
254and haproxy doesn't obey, then first ask yourself if it is possible that nobody
255before you has even encountered this issue. If it's unlikely, the you probably
256have an issue in your setup. Just in case of doubt, please consult the mailing
257list archives :
258
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200259 http://marc.info/?l=haproxy
260
261Otherwise, please try to gather the maximum amount of information to help
262reproduce the issue and send that to the mailing list :
263
264 haproxy@formilux.org
265
266Please include your configuration and logs. You can mask your IP addresses and
267passwords, we don't need them. But it's essential that you post your config if
268you want people to guess what is happening.
269
270Also, keep in mind that haproxy is designed to NEVER CRASH. If you see it die
271without any reason, then it definitely is a critical bug that must be reported
272and urgently fixed. It has happened a couple of times in the past, essentially
273on development versions running on new architectures. If you think your setup
274is fairly common, then it is possible that the issue is totally unrelated.
275Anyway, if that happens, feel free to contact me directly, as I will give you
276instructions on how to collect a usable core file, and will probably ask for
277other captures that you'll not want to share with the list.
278
279
2805) How to contribute
281--------------------
282
283It is possible that you'll want to add a specific feature to satisfy your needs
284or one of your customers'. Contributions are welcome, however I'm often very
285picky about changes. I will generally reject patches that change massive parts
286of the code, or that touch the core parts without any good reason if those
287changes have not been discussed first.
288
289The proper place to discuss your changes is the HAProxy Mailing List. There are
290enough skilled readers to catch hazardous mistakes and to suggest improvements.
Willy Tarreau9a639a12011-09-10 22:48:36 +0200291I trust a number of them enough to merge a patch if they say it's OK, so using
292the list is the fastest way to get your code reviewed and merged. You can
293subscribe to it by sending an empty e-mail at the following address :
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200294
295 haproxy+subscribe@formilux.org
296
Willy Tarreau9a639a12011-09-10 22:48:36 +0200297If you have an idea about something to implement, *please* discuss it on the
298list first. It has already happened several times that two persons did the same
299thing simultaneously. This is a waste of time for both of them. It's also very
300common to see some changes rejected because they're done in a way that will
301conflict with future evolutions, or that does not leave a good feeling. It's
302always unpleasant for the person who did the work, and it is unpleasant for me
303too because I value people's time and efforts. That would not happen if these
304were discussed first. There is no problem posting work in progress to the list,
305it happens quite often in fact. Also, don't waste your time with the doc when
306submitting patches for review, only add the doc with the patch you consider
307ready to merge.
308
Willy Tarreau2ddccb72013-05-01 10:07:21 +0200309Another important point concerns code portability. Haproxy requires gcc as the
310C compiler, and may or may not work with other compilers. However it's known
311to build using gcc 2.95 or any later version. As such, it is important to keep
312in mind that certain facilities offered by recent versions must not be used in
313the code :
314
315 - declarations mixed in the code (requires gcc >= 3.x)
316 - GCC builtins without checking for their availability based on version and
317 architecture ;
318 - assembly code without any alternate portable form for other platforms
319 - use of stdbool.h, "bool", "false", "true" : simply use "int", "0", "1"
320 - in general, anything which requires C99 (such as declaring variables in
321 "for" statements)
322
323Since most of these restrictions are just a matter of coding style, it is
324normally not a problem to comply.
325
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200326If your work is very confidential and you can't publicly discuss it, you can
327also mail me directly about it, but your mail may be waiting several days in
328the queue before you get a response.
329
330If you'd like a feature to be added but you think you don't have the skills to
331implement it yourself, you should follow these steps :
332
333 1. discuss the feature on the mailing list. It is possible that someone
334 else has already implemented it, or that someone will tell you how to
335 proceed without it, or even why not to do it. It is also possible that
336 in fact it's quite easy to implement and people will guide you through
337 the process. That way you'll finally have YOUR patch merged, providing
338 the feature YOU need.
339
340 2. if you really can't code it yourself after discussing it, then you may
341 consider contacting someone to do the job for you. Some people on the
Willy Tarreau869f3512014-06-19 15:26:32 +0200342 list might sometimes be OK with trying to do it.
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200343
344Note to contributors: it's very handy when patches comes with a properly
Willy Tarreau9a639a12011-09-10 22:48:36 +0200345formated subject. There are 3 criteria of particular importance in any patch :
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200346
Willy Tarreau9a639a12011-09-10 22:48:36 +0200347 - its nature (is it a fix for a bug, a new feature, an optimization, ...)
348 - its importance, which generally reflects the risk of merging/not merging it
349 - what area it applies to (eg: http, stats, startup, config, doc, ...)
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200350
Willy Tarreau9a639a12011-09-10 22:48:36 +0200351It's important to make these 3 criteria easy to spot in the patch's subject,
352because it's the first (and sometimes the only) thing which is read when
353reviewing patches to find which ones need to be backported to older versions.
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200354
Willy Tarreau9a639a12011-09-10 22:48:36 +0200355Specifically, bugs must be clearly easy to spot so that they're never missed.
356Any patch fixing a bug must have the "BUG" tag in its subject. Most common
357patch types include :
358
359 - BUG fix for a bug. The severity of the bug should also be indicated
360 when known. Similarly, if a backport is needed to older versions,
361 it should be indicated on the last line of the commit message. If
362 the bug has been identified as a regression brought by a specific
363 patch or version, this indication will be appreciated too. New
364 maintenance releases are generally emitted when a few of these
365 patches are merged.
366
367 - CLEANUP code cleanup, silence of warnings, etc... theorically no impact.
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200368 These patches will rarely be seen in stable branches, though they
Willy Tarreau9a639a12011-09-10 22:48:36 +0200369 may appear when they remove some annoyance or when they make
370 backporting easier. By nature, a cleanup is always minor.
371
372 - REORG code reorganization. Some blocks may be moved to other places,
373 some important checks might be swapped, etc... These changes
374 always present a risk of regression. For this reason, they should
375 never be mixed with any bug fix nor functional change. Code is
376 only moved as-is. Indicating the risk of breakage is highly
377 recommended.
378
379 - BUILD updates or fixes for build issues. Changes to makefiles also fall
380 into this category. The risk of breakage should be indicated if
381 known. It is also appreciated to indicate what platforms and/or
382 configurations were tested after the change.
383
384 - OPTIM some code was optimised. Sometimes if the regression risk is very
385 low and the gains significant, such patches may be merged in the
386 stable branch. Depending on the amount of code changed or replaced
387 and the level of trust the author has in the change, the risk of
388 regression should be indicated.
389
390 - RELEASE release of a new version (development or stable).
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200391
Willy Tarreau9a639a12011-09-10 22:48:36 +0200392 - LICENSE licensing updates (may impact distro packagers).
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200393
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200394
Willy Tarreau9a639a12011-09-10 22:48:36 +0200395When the patch cannot be categorized, it's best not to put any tag. This is
396commonly the case for new features, which development versions are mostly made
397of.
398
399Additionally, the importance of the patch should be indicated when known. A
400single upper-case word is preferred, among :
401
402 - MINOR minor change, very low risk of impact. It is often the case for
403 code additions that don't touch live code. For a bug, it generally
404 indicates an annoyance, nothing more.
405
406 - MEDIUM medium risk, may cause unexpected regressions of low importance or
407 which may quickly be discovered. For a bug, it generally indicates
408 something odd which requires changing the configuration in an
409 undesired way to work around the issue.
410
411 - MAJOR major risk of hidden regression. This happens when I rearrange
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200412 large parts of code, when I play with timeouts, with variable
413 initializations, etc... We should only exceptionally find such
Willy Tarreau9a639a12011-09-10 22:48:36 +0200414 patches in stable branches. For a bug, it indicates severe
415 reliability issues for which workarounds are identified with or
416 without performance impacts.
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200417
Willy Tarreau9a639a12011-09-10 22:48:36 +0200418 - CRITICAL medium-term reliability or security is at risk and workarounds,
419 if they exist, might not always be acceptable. An upgrade is
420 absolutely required. A maintenance release may be emitted even if
421 only one of these bugs are fixed. Note that this tag is only used
422 with bugs. Such patches must indicate what is the first version
423 affected, and if known, the commit ID which introduced the issue.
424
425If this criterion doesn't apply, it's best not to put it. For instance, most
426doc updates and most examples or test files are just added or updated without
427any need to qualify a level of importance.
428
429The area the patch applies to is quite important, because some areas are known
430to be similar in older versions, suggesting a backport might be desirable, and
431conversely, some areas are known to be specific to one version. When the tag is
432used alone, uppercase is preferred for readability, otherwise lowercase is fine
433too. The following tags are suggested but not limitative :
434
435 - doc documentation updates or fixes. No code is affected, no need to
436 upgrade. These patches can also be sent right after a new feature,
437 to document it.
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200438
Willy Tarreau9a639a12011-09-10 22:48:36 +0200439 - examples example files. Be careful, sometimes these files are packaged.
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200440
Willy Tarreau9a639a12011-09-10 22:48:36 +0200441 - tests regression test files. No code is affected, no need to upgrade.
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200442
Willy Tarreau9a639a12011-09-10 22:48:36 +0200443 - init initialization code, arguments parsing, etc...
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200444
Willy Tarreau9a639a12011-09-10 22:48:36 +0200445 - config configuration parser, mostly used when adding new config keywords
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200446
Willy Tarreau9a639a12011-09-10 22:48:36 +0200447 - http the HTTP engine
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200448
Willy Tarreau9a639a12011-09-10 22:48:36 +0200449 - stats the stats reporting engine as well as the stats socket CLI
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200450
Willy Tarreau9a639a12011-09-10 22:48:36 +0200451 - checks the health checks engine (eg: when adding new checks)
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200452
Willy Tarreau9a639a12011-09-10 22:48:36 +0200453 - acl the ACL processing core or some ACLs from other areas
454
455 - peers the peer synchronization engine
456
457 - listeners everything related to incoming connection settings
458
459 - frontend everything related to incoming connection processing
460
461 - backend everything related to LB algorithms and server farm
462
463 - session session processing and flags (very sensible, be careful)
464
465 - server server connection management, queueing
466
467 - proxy proxy maintenance (start/stop)
468
469 - log log management
470
471 - poll any of the pollers
472
473 - halog the halog sub-component in the contrib directory
474
475 - contrib any addition to the contrib directory
476
477Other names may be invented when more precise indications are meaningful, for
478instance : "cookie" which indicates cookie processing in the HTTP core. Last,
479indicating the name of the affected file is also a good way to quickly spot
480changes. Many commits were already tagged with "stream_sock" or "cfgparse" for
481instance.
482
483It is desired that AT LEAST one of the 3 criteria tags is reported in the patch
484subject. Ideally, we would have the 3 most often. The two first criteria should
485be present before a first colon (':'). If both are present, then they should be
486delimited with a slash ('/'). The 3rd criterion (area) should appear next, also
487followed by a colon. Thus, all of the following messages are valid :
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200488
489Examples of messages :
Willy Tarreau9a639a12011-09-10 22:48:36 +0200490 - DOC: document options forwardfor to logasap
491 - DOC/MAJOR: reorganize the whole document and change indenting
492 - BUG: stats: connection reset counters must be plain ascii, not HTML
493 - BUG/MINOR: stats: connection reset counters must be plain ascii, not HTML
494 - MEDIUM: checks: support multi-packet health check responses
495 - RELEASE: Released version 1.4.2
496 - BUILD: stats: stdint is not present on solaris
497 - OPTIM/MINOR: halog: make fgets parse more bytes by blocks
498 - REORG/MEDIUM: move syscall redefinition to specific places
499
500Please do not use square brackets anymore around the tags, because they give me
501more work when merging patches. By default I'm asking Git to keep them but this
502causes trouble when patches are prefixed with the [PATCH] tag because in order
503not to store it, I have to hand-edit the patches. So as of now, I will ask Git
504to remove whatever is located between square brackets, which implies that any
505subject formatted the old way will have its tag stripped out.
506
507In fact, one of the only square bracket tags that still makes sense is '[RFC]'
508at the beginning of the subject, when you're asking for someone to review your
509change before getting it merged. If the patch is OK to be merged, then I can
510merge it as-is and the '[RFC]' tag will automatically be removed. If you don't
511want it to be merged at all, you can simply state it in the message, or use an
512alternate '[WIP]' tag ("work in progress").
513
514The tags are not rigid, follow your intuition first, anyway I reserve the right
515to change them when merging the patch. It may happen that a same patch has a
516different tag in two distinct branches. The reason is that a bug in one branch
517may just be a cleanup in the other one because the code cannot be triggered.
518
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200519
520For a more efficient interaction between the mainline code and your code, I can
521only strongly encourage you to try the Git version control system :
522
523 http://git-scm.com/
524
525It's very fast, lightweight and lets you undo/redo your work as often as you
526want, without making your mistakes visible to the rest of the world. It will
527definitely help you contribute quality code and take other people's feedback
528in consideration. In order to clone the HAProxy Git repository :
529
Willy Tarreau869f3512014-06-19 15:26:32 +0200530 $ git clone http://git.haproxy.org/git/haproxy-1.5.git (stable 1.5)
Willy Tarreau6346f0a2014-05-10 11:04:39 +0200531 $ git clone http://git.haproxy.org/git/haproxy.git/ (development)
Willy Tarreau663148c2012-12-12 00:38:22 +0100532
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200533If you decide to use Git for your developments, then your commit messages will
534have the subject line in the format described above, then the whole description
535of your work (mainly why you did it) will be in the body. You can directly send
536your commits to the mailing list, the format is convenient to read and process.
537
willy tarreau78345332005-12-18 01:33:16 +0100538-- end