blob: 5c084130ccaa4743c924964de36d31a7d003e35d [file] [log] [blame]
Willy Tarreaub1a34b62010-05-09 22:37:12 +02001 ----------------------
2 HAProxy how-to
3 ----------------------
Willy Tarreau11f64d62013-12-16 02:32:37 +01004 version 1.5-dev20
Willy Tarreaub1a34b62010-05-09 22:37:12 +02005 willy tarreau
Willy Tarreau11f64d62013-12-16 02:32:37 +01006 2013/12/16
willy tarreau78345332005-12-18 01:33:16 +01007
8
Willy Tarreaub1a34b62010-05-09 22:37:12 +020091) How to build it
10------------------
11
willy tarreau78345332005-12-18 01:33:16 +010012To build haproxy, you will need :
Willy Tarreaub1a34b62010-05-09 22:37:12 +020013 - GNU make. Neither Solaris nor OpenBSD's make work with the GNU Makefile.
Willy Tarreau83b30c12008-05-25 10:32:50 +020014 However, specific Makefiles for BSD and OSX are provided.
Willy Tarreau32e65ef2013-04-02 08:14:29 +020015 - GCC between 2.91 and 4.7. Others may work, but not tested.
willy tarreau78345332005-12-18 01:33:16 +010016 - GNU ld
17
18Also, you might want to build with libpcre support, which will provide a very
Willy Tarreaub1a34b62010-05-09 22:37:12 +020019efficient regex implementation and will also fix some badness on Solaris' one.
willy tarreau78345332005-12-18 01:33:16 +010020
21To build haproxy, you have to choose your target OS amongst the following ones
22and assign it to the TARGET variable :
23
Willy Tarreau83b30c12008-05-25 10:32:50 +020024 - linux22 for Linux 2.2
25 - linux24 for Linux 2.4 and above (default)
26 - linux24e for Linux 2.4 with support for a working epoll (> 0.21)
Willy Tarreau83b30c12008-05-25 10:32:50 +020027 - linux26 for Linux 2.6 and above
Willy Tarreaue0c623d2012-06-04 00:42:09 +020028 - linux2628 for Linux 2.6.28 and above (enables splice and tproxy)
Willy Tarreau83b30c12008-05-25 10:32:50 +020029 - solaris for Solaris 8 or 10 (others untested)
Willy Tarreaub1a34b62010-05-09 22:37:12 +020030 - freebsd for FreeBSD 5 to 8.0 (others untested)
Willy Tarreau8624cab2013-04-02 08:17:43 +020031 - osx for Mac OS/X
Willy Tarreau3b8e9792012-11-22 00:43:09 +010032 - openbsd for OpenBSD 3.1 to 5.2 (others untested)
Willy Tarreau7dec9652012-06-06 16:15:03 +020033 - aix52 for AIX 5.2
Yitzhak Sapir32087312009-06-14 18:27:54 +020034 - cygwin for Cygwin
Willy Tarreau83b30c12008-05-25 10:32:50 +020035 - generic for any other OS.
36 - custom to manually adjust every setting
willy tarreau78345332005-12-18 01:33:16 +010037
38You may also choose your CPU to benefit from some optimizations. This is
39particularly important on UltraSparc machines. For this, you can assign
40one of the following choices to the CPU variable :
41
42 - i686 for intel PentiumPro, Pentium 2 and above, AMD Athlon
43 - i586 for intel Pentium, AMD K6, VIA C3.
44 - ultrasparc : Sun UltraSparc I/II/III/IV processor
Willy Tarreaua5899aa2010-11-28 07:41:00 +010045 - native : use the build machine's specific processor optimizations
willy tarreau78345332005-12-18 01:33:16 +010046 - generic : any other processor or no specific optimization. (default)
47
Willy Tarreau83b30c12008-05-25 10:32:50 +020048Alternatively, you may just set the CPU_CFLAGS value to the optimal GCC options
49for your platform.
50
Willy Tarreauef7341d2009-04-11 19:45:50 +020051You may want to build specific target binaries which do not match your native
52compiler's target. This is particularly true on 64-bit systems when you want
53to build a 32-bit binary. Use the ARCH variable for this purpose. Right now
Willy Tarreaua5899aa2010-11-28 07:41:00 +010054it only knows about a few x86 variants (i386,i486,i586,i686,x86_64), two
55generic ones (32,64) and sets -m32/-m64 as well as -march=<arch> accordingly.
Willy Tarreauef7341d2009-04-11 19:45:50 +020056
willy tarreau78345332005-12-18 01:33:16 +010057If your system supports PCRE (Perl Compatible Regular Expressions), then you
58really should build with libpcre which is between 2 and 10 times faster than
59other libc implementations. Regex are used for header processing (deletion,
60rewriting, allow, deny). The only inconvenient of libpcre is that it is not
61yet widely spread, so if you build for other systems, you might get into
62trouble if they don't have the dynamic library. In this situation, you should
63statically link libpcre into haproxy so that it will not be necessary to
Willy Tarreau83b30c12008-05-25 10:32:50 +020064install it on target systems. Available build options for PCRE are :
willy tarreau78345332005-12-18 01:33:16 +010065
Willy Tarreau83b30c12008-05-25 10:32:50 +020066 - USE_PCRE=1 to use libpcre, in whatever form is available on your system
willy tarreau78345332005-12-18 01:33:16 +010067 (shared or static)
68
Willy Tarreau83b30c12008-05-25 10:32:50 +020069 - USE_STATIC_PCRE=1 to use a static version of libpcre even if the dynamic
70 one is available. This will enhance portability.
71
Willy Tarreau663148c2012-12-12 00:38:22 +010072 - with no option, use your OS libc's standard regex implementation (default).
Willy Tarreau83b30c12008-05-25 10:32:50 +020073 Warning! group references on Solaris seem broken. Use static-pcre whenever
74 possible.
willy tarreau78345332005-12-18 01:33:16 +010075
Willy Tarreau64bc40b2011-03-23 20:00:53 +010076Recent systems can resolve IPv6 host names using getaddrinfo(). This primitive
77is not present in all libcs and does not work in all of them either. Support in
78glibc was broken before 2.3. Some embedded libs may not properly work either,
79thus, support is disabled by default, meaning that some host names which only
80resolve as IPv6 addresses will not resolve and configs might emit an error
81during parsing. If you know that your OS libc has reliable support for
82getaddrinfo(), you can add USE_GETADDRINFO=1 on the make command line to enable
83it. This is the recommended option for most Linux distro packagers since it's
84working fine on all recent mainstream distros. It is automatically enabled on
85Solaris 8 and above, as it's known to work.
86
Willy Tarreaud4508812012-09-10 09:07:41 +020087It is possible to add native support for SSL using the GNU makefile only, and
88by passing "USE_OPENSSL=1" on the make commande line. The libssl and libcrypto
89will automatically be linked with haproxy. Some systems also require libz, so
90if the build fails due to missing symbols such as deflateInit(), then try again
91with "ADDLIB=-lz".
92
Lukas Tribus3fe9f1e2013-05-19 16:28:17 +020093To link OpenSSL statically against haproxy, build OpenSSL with the no-shared
94keyword and install it to a local directory, so your system is not affected :
95
96 $ export STATICLIBSSL=/tmp/staticlibssl
97 $ ./config --prefix=$STATICLIBSSL no-shared
98 $ make && make install_sw
99
Lukas Tribus130ddf72013-10-01 00:28:03 +0200100When building haproxy, pass that path via SSL_INC and SSL_LIB to make and
101include additional libs with ADDLIB if needed (in this case for example libdl):
102 $ make TARGET=linux26 USE_OPENSSL=1 SSL_INC=$STATICLIBSSL/include SSL_LIB=$STATICLIBSSL/lib ADDLIB=-ldl
Lukas Tribus3fe9f1e2013-05-19 16:28:17 +0200103
William Lallemand82fe75c2012-10-23 10:25:10 +0200104It is also possible to include native support for ZLIB to benefit from HTTP
105compression. For this, pass "USE_ZLIB=1" on the "make" command line and ensure
106that zlib is present on the system.
107
willy tarreau78345332005-12-18 01:33:16 +0100108By default, the DEBUG variable is set to '-g' to enable debug symbols. It is
109not wise to disable it on uncommon systems, because it's often the only way to
110get a complete core when you need one. Otherwise, you can set DEBUG to '-s' to
111strip the binary.
112
113For example, I use this to build for Solaris 8 :
114
Willy Tarreau83b30c12008-05-25 10:32:50 +0200115 $ make TARGET=solaris CPU=ultrasparc USE_STATIC_PCRE=1
willy tarreau78345332005-12-18 01:33:16 +0100116
Willy Tarreau83b30c12008-05-25 10:32:50 +0200117And I build it this way on OpenBSD or FreeBSD :
willy tarreaud38e72d2006-03-19 20:56:52 +0100118
119 $ make -f Makefile.bsd REGEX=pcre DEBUG= COPTS.generic="-Os -fomit-frame-pointer -mgnu"
120
Willy Tarreau663148c2012-12-12 00:38:22 +0100121And on a classic Linux with SSL and ZLIB support (eg: Red Hat 5.x) :
122
123 $ make TARGET=linux26 CPU=native USE_PCRE=1 USE_OPENSSL=1 USE_ZLIB=1
124
125And on a recent Linux >= 2.6.28 with SSL and ZLIB support :
Willy Tarreaud4508812012-09-10 09:07:41 +0200126
William Lallemand82fe75c2012-10-23 10:25:10 +0200127 $ make TARGET=linux2628 CPU=native USE_PCRE=1 USE_OPENSSL=1 USE_ZLIB=1
Willy Tarreaud4508812012-09-10 09:07:41 +0200128
William Lallemand82fe75c2012-10-23 10:25:10 +0200129In order to build a 32-bit binary on an x86_64 Linux system with SSL support
130without support for compression but when OpenSSL requires ZLIB anyway :
Willy Tarreauef7341d2009-04-11 19:45:50 +0200131
Willy Tarreaud4508812012-09-10 09:07:41 +0200132 $ make TARGET=linux26 ARCH=i386 USE_OPENSSL=1 ADDLIB=-lz
Willy Tarreauef7341d2009-04-11 19:45:50 +0200133
Willy Tarreau3b8e9792012-11-22 00:43:09 +0100134The BSD and OSX makefiles do not support build options for OpenSSL nor zlib.
135Also, at least on OpenBSD, pthread_mutexattr_setpshared() does not exist so
136the SSL session cache cannot be shared between multiple processes. If you want
137to enable these options, you need to use GNU make with the default makefile as
138follows :
139
140 $ gmake TARGET=openbsd USE_OPENSSL=1 USE_ZLIB=1 USE_PRIVATE_CACHE=1
141
willy tarreau78345332005-12-18 01:33:16 +0100142If you need to pass other defines, includes, libraries, etc... then please
143check the Makefile to see which ones will be available in your case, and
Willy Tarreau83b30c12008-05-25 10:32:50 +0200144use the USE_* variables in the GNU Makefile, or ADDINC, ADDLIB, and DEFINE
145variables in the BSD makefiles.
willy tarreau78345332005-12-18 01:33:16 +0100146
Willy Tarreau97ec9692010-01-28 20:52:05 +0100147AIX 5.3 is known to work with the generic target. However, for the binary to
148also run on 5.2 or earlier, you need to build with DEFINE="-D_MSGQSUPPORT",
149otherwise __fd_select() will be used while not being present in the libc.
Willy Tarreau7dec9652012-06-06 16:15:03 +0200150If you get build errors because of strange symbols or section mismatches,
151simply remove -g from DEBUG_CFLAGS.
Willy Tarreau97ec9692010-01-28 20:52:05 +0100152
Willy Tarreau32e65ef2013-04-02 08:14:29 +0200153You can easily define your own target with the GNU Makefile. Unknown targets
154are processed with no default option except USE_POLL=default. So you can very
155well use that property to define your own set of options. USE_POLL can even be
156disabled by setting USE_POLL="". For example :
157
158 $ gmake TARGET=tiny USE_POLL="" TARGET_CFLAGS=-fomit-frame-pointer
159
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200160
1612) How to install it
162--------------------
163
164To install haproxy, you can either copy the single resulting binary to the
165place you want, or run :
166
167 $ sudo make install
168
169If you're packaging it for another system, you can specify its root directory
170in the usual DESTDIR variable.
171
172
1733) How to set it up
174-------------------
175
176There is some documentation in the doc/ directory :
177
178 - architecture.txt : this is the architecture manual. It is quite old and
179 does not tell about the nice new features, but it's still a good starting
180 point when you know what you want but don't know how to do it.
181
182 - configuration.txt : this is the configuration manual. It recalls a few
183 essential HTTP basic concepts, and details all the configuration file
184 syntax (keywords, units). It also describes the log and stats format. It
185 is normally always up to date. If you see that something is missing from
186 it, please report it as this is a bug.
187
188 - haproxy-en.txt / haproxy-fr.txt : these are the old outdated docs. You
189 should never need them. If you do, then please report what you didn't
190 find in the other ones.
191
192 - gpl.txt / lgpl.txt : the copy of the licenses covering the software. See
193 the 'LICENSE' file at the top for more information.
194
195 - the rest is mainly for developers.
196
197There are also a number of nice configuration examples in the "examples"
198directory as well as on several sites and articles on the net which are linked
199to from the haproxy web site.
200
201
2024) How to report a bug
203----------------------
204
205It is possible that from time to time you'll find a bug. A bug is a case where
206what you see is not what is documented. Otherwise it can be a misdesign. If you
207find that something is stupidly design, please discuss it on the list (see the
208"how to contribute" section below). If you feel like you're proceeding right
209and haproxy doesn't obey, then first ask yourself if it is possible that nobody
210before you has even encountered this issue. If it's unlikely, the you probably
211have an issue in your setup. Just in case of doubt, please consult the mailing
212list archives :
213
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200214 http://marc.info/?l=haproxy
215
216Otherwise, please try to gather the maximum amount of information to help
217reproduce the issue and send that to the mailing list :
218
219 haproxy@formilux.org
220
221Please include your configuration and logs. You can mask your IP addresses and
222passwords, we don't need them. But it's essential that you post your config if
223you want people to guess what is happening.
224
225Also, keep in mind that haproxy is designed to NEVER CRASH. If you see it die
226without any reason, then it definitely is a critical bug that must be reported
227and urgently fixed. It has happened a couple of times in the past, essentially
228on development versions running on new architectures. If you think your setup
229is fairly common, then it is possible that the issue is totally unrelated.
230Anyway, if that happens, feel free to contact me directly, as I will give you
231instructions on how to collect a usable core file, and will probably ask for
232other captures that you'll not want to share with the list.
233
234
2355) How to contribute
236--------------------
237
238It is possible that you'll want to add a specific feature to satisfy your needs
239or one of your customers'. Contributions are welcome, however I'm often very
240picky about changes. I will generally reject patches that change massive parts
241of the code, or that touch the core parts without any good reason if those
242changes have not been discussed first.
243
244The proper place to discuss your changes is the HAProxy Mailing List. There are
245enough skilled readers to catch hazardous mistakes and to suggest improvements.
Willy Tarreau9a639a12011-09-10 22:48:36 +0200246I trust a number of them enough to merge a patch if they say it's OK, so using
247the list is the fastest way to get your code reviewed and merged. You can
248subscribe to it by sending an empty e-mail at the following address :
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200249
250 haproxy+subscribe@formilux.org
251
Willy Tarreau9a639a12011-09-10 22:48:36 +0200252If you have an idea about something to implement, *please* discuss it on the
253list first. It has already happened several times that two persons did the same
254thing simultaneously. This is a waste of time for both of them. It's also very
255common to see some changes rejected because they're done in a way that will
256conflict with future evolutions, or that does not leave a good feeling. It's
257always unpleasant for the person who did the work, and it is unpleasant for me
258too because I value people's time and efforts. That would not happen if these
259were discussed first. There is no problem posting work in progress to the list,
260it happens quite often in fact. Also, don't waste your time with the doc when
261submitting patches for review, only add the doc with the patch you consider
262ready to merge.
263
Willy Tarreau2ddccb72013-05-01 10:07:21 +0200264Another important point concerns code portability. Haproxy requires gcc as the
265C compiler, and may or may not work with other compilers. However it's known
266to build using gcc 2.95 or any later version. As such, it is important to keep
267in mind that certain facilities offered by recent versions must not be used in
268the code :
269
270 - declarations mixed in the code (requires gcc >= 3.x)
271 - GCC builtins without checking for their availability based on version and
272 architecture ;
273 - assembly code without any alternate portable form for other platforms
274 - use of stdbool.h, "bool", "false", "true" : simply use "int", "0", "1"
275 - in general, anything which requires C99 (such as declaring variables in
276 "for" statements)
277
278Since most of these restrictions are just a matter of coding style, it is
279normally not a problem to comply.
280
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200281If your work is very confidential and you can't publicly discuss it, you can
282also mail me directly about it, but your mail may be waiting several days in
283the queue before you get a response.
284
285If you'd like a feature to be added but you think you don't have the skills to
286implement it yourself, you should follow these steps :
287
288 1. discuss the feature on the mailing list. It is possible that someone
289 else has already implemented it, or that someone will tell you how to
290 proceed without it, or even why not to do it. It is also possible that
291 in fact it's quite easy to implement and people will guide you through
292 the process. That way you'll finally have YOUR patch merged, providing
293 the feature YOU need.
294
295 2. if you really can't code it yourself after discussing it, then you may
296 consider contacting someone to do the job for you. Some people on the
297 list might be OK with trying to do it. Otherwise, you can check the list
298 of contributors at the URL below, some of the regular contributors may
299 be able to do the work, probably not for free but their time is as much
300 valuable as yours after all, you can't eat the cake and have it too.
301
302The list of past and regular contributors is available below. It lists not only
303significant code contributions (features, fixes), but also time or money
304donations :
305
306 http://haproxy.1wt.eu/contrib.html
307
308Note to contributors: it's very handy when patches comes with a properly
Willy Tarreau9a639a12011-09-10 22:48:36 +0200309formated subject. There are 3 criteria of particular importance in any patch :
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200310
Willy Tarreau9a639a12011-09-10 22:48:36 +0200311 - its nature (is it a fix for a bug, a new feature, an optimization, ...)
312 - its importance, which generally reflects the risk of merging/not merging it
313 - what area it applies to (eg: http, stats, startup, config, doc, ...)
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200314
Willy Tarreau9a639a12011-09-10 22:48:36 +0200315It's important to make these 3 criteria easy to spot in the patch's subject,
316because it's the first (and sometimes the only) thing which is read when
317reviewing patches to find which ones need to be backported to older versions.
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200318
Willy Tarreau9a639a12011-09-10 22:48:36 +0200319Specifically, bugs must be clearly easy to spot so that they're never missed.
320Any patch fixing a bug must have the "BUG" tag in its subject. Most common
321patch types include :
322
323 - BUG fix for a bug. The severity of the bug should also be indicated
324 when known. Similarly, if a backport is needed to older versions,
325 it should be indicated on the last line of the commit message. If
326 the bug has been identified as a regression brought by a specific
327 patch or version, this indication will be appreciated too. New
328 maintenance releases are generally emitted when a few of these
329 patches are merged.
330
331 - CLEANUP code cleanup, silence of warnings, etc... theorically no impact.
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200332 These patches will rarely be seen in stable branches, though they
Willy Tarreau9a639a12011-09-10 22:48:36 +0200333 may appear when they remove some annoyance or when they make
334 backporting easier. By nature, a cleanup is always minor.
335
336 - REORG code reorganization. Some blocks may be moved to other places,
337 some important checks might be swapped, etc... These changes
338 always present a risk of regression. For this reason, they should
339 never be mixed with any bug fix nor functional change. Code is
340 only moved as-is. Indicating the risk of breakage is highly
341 recommended.
342
343 - BUILD updates or fixes for build issues. Changes to makefiles also fall
344 into this category. The risk of breakage should be indicated if
345 known. It is also appreciated to indicate what platforms and/or
346 configurations were tested after the change.
347
348 - OPTIM some code was optimised. Sometimes if the regression risk is very
349 low and the gains significant, such patches may be merged in the
350 stable branch. Depending on the amount of code changed or replaced
351 and the level of trust the author has in the change, the risk of
352 regression should be indicated.
353
354 - RELEASE release of a new version (development or stable).
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200355
Willy Tarreau9a639a12011-09-10 22:48:36 +0200356 - LICENSE licensing updates (may impact distro packagers).
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200357
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200358
Willy Tarreau9a639a12011-09-10 22:48:36 +0200359When the patch cannot be categorized, it's best not to put any tag. This is
360commonly the case for new features, which development versions are mostly made
361of.
362
363Additionally, the importance of the patch should be indicated when known. A
364single upper-case word is preferred, among :
365
366 - MINOR minor change, very low risk of impact. It is often the case for
367 code additions that don't touch live code. For a bug, it generally
368 indicates an annoyance, nothing more.
369
370 - MEDIUM medium risk, may cause unexpected regressions of low importance or
371 which may quickly be discovered. For a bug, it generally indicates
372 something odd which requires changing the configuration in an
373 undesired way to work around the issue.
374
375 - MAJOR major risk of hidden regression. This happens when I rearrange
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200376 large parts of code, when I play with timeouts, with variable
377 initializations, etc... We should only exceptionally find such
Willy Tarreau9a639a12011-09-10 22:48:36 +0200378 patches in stable branches. For a bug, it indicates severe
379 reliability issues for which workarounds are identified with or
380 without performance impacts.
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200381
Willy Tarreau9a639a12011-09-10 22:48:36 +0200382 - CRITICAL medium-term reliability or security is at risk and workarounds,
383 if they exist, might not always be acceptable. An upgrade is
384 absolutely required. A maintenance release may be emitted even if
385 only one of these bugs are fixed. Note that this tag is only used
386 with bugs. Such patches must indicate what is the first version
387 affected, and if known, the commit ID which introduced the issue.
388
389If this criterion doesn't apply, it's best not to put it. For instance, most
390doc updates and most examples or test files are just added or updated without
391any need to qualify a level of importance.
392
393The area the patch applies to is quite important, because some areas are known
394to be similar in older versions, suggesting a backport might be desirable, and
395conversely, some areas are known to be specific to one version. When the tag is
396used alone, uppercase is preferred for readability, otherwise lowercase is fine
397too. The following tags are suggested but not limitative :
398
399 - doc documentation updates or fixes. No code is affected, no need to
400 upgrade. These patches can also be sent right after a new feature,
401 to document it.
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200402
Willy Tarreau9a639a12011-09-10 22:48:36 +0200403 - examples example files. Be careful, sometimes these files are packaged.
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200404
Willy Tarreau9a639a12011-09-10 22:48:36 +0200405 - tests regression test files. No code is affected, no need to upgrade.
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200406
Willy Tarreau9a639a12011-09-10 22:48:36 +0200407 - init initialization code, arguments parsing, etc...
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200408
Willy Tarreau9a639a12011-09-10 22:48:36 +0200409 - config configuration parser, mostly used when adding new config keywords
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200410
Willy Tarreau9a639a12011-09-10 22:48:36 +0200411 - http the HTTP engine
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200412
Willy Tarreau9a639a12011-09-10 22:48:36 +0200413 - stats the stats reporting engine as well as the stats socket CLI
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200414
Willy Tarreau9a639a12011-09-10 22:48:36 +0200415 - checks the health checks engine (eg: when adding new checks)
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200416
Willy Tarreau9a639a12011-09-10 22:48:36 +0200417 - acl the ACL processing core or some ACLs from other areas
418
419 - peers the peer synchronization engine
420
421 - listeners everything related to incoming connection settings
422
423 - frontend everything related to incoming connection processing
424
425 - backend everything related to LB algorithms and server farm
426
427 - session session processing and flags (very sensible, be careful)
428
429 - server server connection management, queueing
430
431 - proxy proxy maintenance (start/stop)
432
433 - log log management
434
435 - poll any of the pollers
436
437 - halog the halog sub-component in the contrib directory
438
439 - contrib any addition to the contrib directory
440
441Other names may be invented when more precise indications are meaningful, for
442instance : "cookie" which indicates cookie processing in the HTTP core. Last,
443indicating the name of the affected file is also a good way to quickly spot
444changes. Many commits were already tagged with "stream_sock" or "cfgparse" for
445instance.
446
447It is desired that AT LEAST one of the 3 criteria tags is reported in the patch
448subject. Ideally, we would have the 3 most often. The two first criteria should
449be present before a first colon (':'). If both are present, then they should be
450delimited with a slash ('/'). The 3rd criterion (area) should appear next, also
451followed by a colon. Thus, all of the following messages are valid :
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200452
453Examples of messages :
Willy Tarreau9a639a12011-09-10 22:48:36 +0200454 - DOC: document options forwardfor to logasap
455 - DOC/MAJOR: reorganize the whole document and change indenting
456 - BUG: stats: connection reset counters must be plain ascii, not HTML
457 - BUG/MINOR: stats: connection reset counters must be plain ascii, not HTML
458 - MEDIUM: checks: support multi-packet health check responses
459 - RELEASE: Released version 1.4.2
460 - BUILD: stats: stdint is not present on solaris
461 - OPTIM/MINOR: halog: make fgets parse more bytes by blocks
462 - REORG/MEDIUM: move syscall redefinition to specific places
463
464Please do not use square brackets anymore around the tags, because they give me
465more work when merging patches. By default I'm asking Git to keep them but this
466causes trouble when patches are prefixed with the [PATCH] tag because in order
467not to store it, I have to hand-edit the patches. So as of now, I will ask Git
468to remove whatever is located between square brackets, which implies that any
469subject formatted the old way will have its tag stripped out.
470
471In fact, one of the only square bracket tags that still makes sense is '[RFC]'
472at the beginning of the subject, when you're asking for someone to review your
473change before getting it merged. If the patch is OK to be merged, then I can
474merge it as-is and the '[RFC]' tag will automatically be removed. If you don't
475want it to be merged at all, you can simply state it in the message, or use an
476alternate '[WIP]' tag ("work in progress").
477
478The tags are not rigid, follow your intuition first, anyway I reserve the right
479to change them when merging the patch. It may happen that a same patch has a
480different tag in two distinct branches. The reason is that a bug in one branch
481may just be a cleanup in the other one because the code cannot be triggered.
482
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200483
484For a more efficient interaction between the mainline code and your code, I can
485only strongly encourage you to try the Git version control system :
486
487 http://git-scm.com/
488
489It's very fast, lightweight and lets you undo/redo your work as often as you
490want, without making your mistakes visible to the rest of the world. It will
491definitely help you contribute quality code and take other people's feedback
492in consideration. In order to clone the HAProxy Git repository :
493
494 $ git clone http://git.1wt.eu/git/haproxy-1.4.git (stable 1.4)
495 $ git clone http://git.1wt.eu/git/haproxy.git/ (development)
496
Willy Tarreau663148c2012-12-12 00:38:22 +0100497The site above is slow, a faster mirror is maintained up to date here :
498
499 $ git clone http://master.formilux.org/git/people/willy/haproxy.git/
500
Willy Tarreaub1a34b62010-05-09 22:37:12 +0200501If you decide to use Git for your developments, then your commit messages will
502have the subject line in the format described above, then the whole description
503of your work (mainly why you did it) will be in the body. You can directly send
504your commits to the mailing list, the format is convenient to read and process.
505
willy tarreau78345332005-12-18 01:33:16 +0100506-- end