blob: bf9e82b18a921ece9672e613f4794f3158beb8d8 [file] [log] [blame]
----------------------
HAProxy how-to
----------------------
version 1.6-dev
willy tarreau
2014/06/19
1) How to build it
------------------
First, please note that this version is a development version, so in general if
you are not used to build from sources or if you don't have the time to track
very frequent updates, it is recommended that instead you switch to the stable
version (1.5) or follow the packaged updates provided by your software vendor
or Linux distribution. Most of them are taking this task seriously and are
doing a good job. If for any reason you'd prefer a different version than the
one packaged for your system, or to get some commercial support, other choices
are available at :
http://www.haproxy.com/
To build haproxy, you will need :
- GNU make. Neither Solaris nor OpenBSD's make work with the GNU Makefile.
If you get many syntax errors when running "make", you may want to retry
with "gmake" which is the name commonly used for GNU make on BSD systems.
- GCC between 2.95 and 4.8. Others may work, but not tested.
- GNU ld
Also, you might want to build with libpcre support, which will provide a very
efficient regex implementation and will also fix some badness on Solaris' one.
To build haproxy, you have to choose your target OS amongst the following ones
and assign it to the TARGET variable :
- linux22 for Linux 2.2
- linux24 for Linux 2.4 and above (default)
- linux24e for Linux 2.4 with support for a working epoll (> 0.21)
- linux26 for Linux 2.6 and above
- linux2628 for Linux 2.6.28, 3.x, and above (enables splice and tproxy)
- solaris for Solaris 8 or 10 (others untested)
- freebsd for FreeBSD 5 to 10 (others untested)
- osx for Mac OS/X
- openbsd for OpenBSD 3.1 to 5.2 (others untested)
- aix51 for AIX 5.1
- aix52 for AIX 5.2
- cygwin for Cygwin
- generic for any other OS or version.
- custom to manually adjust every setting
You may also choose your CPU to benefit from some optimizations. This is
particularly important on UltraSparc machines. For this, you can assign
one of the following choices to the CPU variable :
- i686 for intel PentiumPro, Pentium 2 and above, AMD Athlon
- i586 for intel Pentium, AMD K6, VIA C3.
- ultrasparc : Sun UltraSparc I/II/III/IV processor
- native : use the build machine's specific processor optimizations
- generic : any other processor or no specific optimization. (default)
Alternatively, you may just set the CPU_CFLAGS value to the optimal GCC options
for your platform.
You may want to build specific target binaries which do not match your native
compiler's target. This is particularly true on 64-bit systems when you want
to build a 32-bit binary. Use the ARCH variable for this purpose. Right now
it only knows about a few x86 variants (i386,i486,i586,i686,x86_64), two
generic ones (32,64) and sets -m32/-m64 as well as -march=<arch> accordingly.
If your system supports PCRE (Perl Compatible Regular Expressions), then you
really should build with libpcre which is between 2 and 10 times faster than
other libc implementations. Regex are used for header processing (deletion,
rewriting, allow, deny). The only inconvenient of libpcre is that it is not
yet widely spread, so if you build for other systems, you might get into
trouble if they don't have the dynamic library. In this situation, you should
statically link libpcre into haproxy so that it will not be necessary to
install it on target systems. Available build options for PCRE are :
- USE_PCRE=1 to use libpcre, in whatever form is available on your system
(shared or static)
- USE_STATIC_PCRE=1 to use a static version of libpcre even if the dynamic
one is available. This will enhance portability.
- with no option, use your OS libc's standard regex implementation (default).
Warning! group references on Solaris seem broken. Use static-pcre whenever
possible.
Recent systems can resolve IPv6 host names using getaddrinfo(). This primitive
is not present in all libcs and does not work in all of them either. Support in
glibc was broken before 2.3. Some embedded libs may not properly work either,
thus, support is disabled by default, meaning that some host names which only
resolve as IPv6 addresses will not resolve and configs might emit an error
during parsing. If you know that your OS libc has reliable support for
getaddrinfo(), you can add USE_GETADDRINFO=1 on the make command line to enable
it. This is the recommended option for most Linux distro packagers since it's
working fine on all recent mainstream distros. It is automatically enabled on
Solaris 8 and above, as it's known to work.
It is possible to add native support for SSL using the GNU makefile, by passing
"USE_OPENSSL=1" on the make command line. The libssl and libcrypto will
automatically be linked with haproxy. Some systems also require libz, so if the
build fails due to missing symbols such as deflateInit(), then try again with
"ADDLIB=-lz".
To link OpenSSL statically against haproxy, build OpenSSL with the no-shared
keyword and install it to a local directory, so your system is not affected :
$ export STATICLIBSSL=/tmp/staticlibssl
$ ./config --prefix=$STATICLIBSSL no-shared
$ make && make install_sw
When building haproxy, pass that path via SSL_INC and SSL_LIB to make and
include additional libs with ADDLIB if needed (in this case for example libdl):
$ make TARGET=linux26 USE_OPENSSL=1 SSL_INC=$STATICLIBSSL/include SSL_LIB=$STATICLIBSSL/lib ADDLIB=-ldl
It is also possible to include native support for ZLIB to benefit from HTTP
compression. For this, pass "USE_ZLIB=1" on the "make" command line and ensure
that zlib is present on the system.
By default, the DEBUG variable is set to '-g' to enable debug symbols. It is
not wise to disable it on uncommon systems, because it's often the only way to
get a complete core when you need one. Otherwise, you can set DEBUG to '-s' to
strip the binary.
For example, I use this to build for Solaris 8 :
$ make TARGET=solaris CPU=ultrasparc USE_STATIC_PCRE=1
And I build it this way on OpenBSD or FreeBSD :
$ gmake TARGET=freebsd USE_PCRE=1 USE_OPENSSL=1 USE_ZLIB=1
And on a classic Linux with SSL and ZLIB support (eg: Red Hat 5.x) :
$ make TARGET=linux26 CPU=native USE_PCRE=1 USE_OPENSSL=1 USE_ZLIB=1
And on a recent Linux >= 2.6.28 with SSL and ZLIB support :
$ make TARGET=linux2628 CPU=native USE_PCRE=1 USE_OPENSSL=1 USE_ZLIB=1
In order to build a 32-bit binary on an x86_64 Linux system with SSL support
without support for compression but when OpenSSL requires ZLIB anyway :
$ make TARGET=linux26 ARCH=i386 USE_OPENSSL=1 ADDLIB=-lz
The SSL stack supports session cache synchronization between all running
processes. This involves some atomic operations and synchronization operations
which come in multiple flavors depending on the system and architecture :
Atomic operations :
- internal assembler versions for x86/x86_64 architectures
- gcc builtins for other architectures. Some architectures might not
be fully supported or might require a more recent version of gcc.
If your architecture is not supported, you willy have to either use
pthread if supported, or to disable the shared cache.
- pthread (posix threads). Pthreads are very common but inter-process
support is not that common, and some older operating systems did not
report an error when enabling multi-process mode, so they used to
silently fail, possibly causing crashes. Linux's implementation is
fine. OpenBSD doesn't support them and doesn't build. FreeBSD 9 builds
and reports an error at runtime, while certain older versions might
silently fail. Pthreads are enabled using USE_PTHREAD_PSHARED=1.
Synchronization operations :
- internal spinlock : this mode is OS-independant, light but will not
scale well to many processes. However, accesses to the session cache
are rare enough that this mode could certainly always be used. This
is the default mode.
- Futexes, which are Linux-specific highly scalable light weight mutexes
implemented in user-space with some limited assistance from the kernel.
This is the default on Linux 2.6 and above and is enabled by passing
USE_FUTEX=1
- pthread (posix threads). See above.
If none of these mechanisms is supported by your platform, you may need to
build with USE_PRIVATE_CACHE=1 to totally disable SSL cache sharing. Then
it is better not to run SSL on multiple processes.
If you need to pass other defines, includes, libraries, etc... then please
check the Makefile to see which ones will be available in your case, and
use the USE_* variables in the Makefile.
AIX 5.3 is known to work with the generic target. However, for the binary to
also run on 5.2 or earlier, you need to build with DEFINE="-D_MSGQSUPPORT",
otherwise __fd_select() will be used while not being present in the libc, but
this is easily addressed using the "aix52" target. If you get build errors
because of strange symbols or section mismatches, simply remove -g from
DEBUG_CFLAGS.
You can easily define your own target with the GNU Makefile. Unknown targets
are processed with no default option except USE_POLL=default. So you can very
well use that property to define your own set of options. USE_POLL can even be
disabled by setting USE_POLL="". For example :
$ gmake TARGET=tiny USE_POLL="" TARGET_CFLAGS=-fomit-frame-pointer
2) How to install it
--------------------
To install haproxy, you can either copy the single resulting binary to the
place you want, or run :
$ sudo make install
If you're packaging it for another system, you can specify its root directory
in the usual DESTDIR variable.
3) How to set it up
-------------------
There is some documentation in the doc/ directory :
- architecture.txt : this is the architecture manual. It is quite old and
does not tell about the nice new features, but it's still a good starting
point when you know what you want but don't know how to do it.
- configuration.txt : this is the configuration manual. It recalls a few
essential HTTP basic concepts, and details all the configuration file
syntax (keywords, units). It also describes the log and stats format. It
is normally always up to date. If you see that something is missing from
it, please report it as this is a bug. Please note that this file is
huge and that it's generally more convenient to review Cyril Bonté's
HTML translation online here :
http://cbonte.github.io/haproxy-dconv/configuration-1.5.html
- haproxy-en.txt / haproxy-fr.txt : these are the old outdated docs. You
should never need them. If you do, then please report what you didn't
find in the other ones.
- gpl.txt / lgpl.txt : the copy of the licenses covering the software. See
the 'LICENSE' file at the top for more information.
- the rest is mainly for developers.
There are also a number of nice configuration examples in the "examples"
directory as well as on several sites and articles on the net which are linked
to from the haproxy web site.
4) How to report a bug
----------------------
It is possible that from time to time you'll find a bug. A bug is a case where
what you see is not what is documented. Otherwise it can be a misdesign. If you
find that something is stupidly design, please discuss it on the list (see the
"how to contribute" section below). If you feel like you're proceeding right
and haproxy doesn't obey, then first ask yourself if it is possible that nobody
before you has even encountered this issue. If it's unlikely, the you probably
have an issue in your setup. Just in case of doubt, please consult the mailing
list archives :
http://marc.info/?l=haproxy
Otherwise, please try to gather the maximum amount of information to help
reproduce the issue and send that to the mailing list :
haproxy@formilux.org
Please include your configuration and logs. You can mask your IP addresses and
passwords, we don't need them. But it's essential that you post your config if
you want people to guess what is happening.
Also, keep in mind that haproxy is designed to NEVER CRASH. If you see it die
without any reason, then it definitely is a critical bug that must be reported
and urgently fixed. It has happened a couple of times in the past, essentially
on development versions running on new architectures. If you think your setup
is fairly common, then it is possible that the issue is totally unrelated.
Anyway, if that happens, feel free to contact me directly, as I will give you
instructions on how to collect a usable core file, and will probably ask for
other captures that you'll not want to share with the list.
5) How to contribute
--------------------
It is possible that you'll want to add a specific feature to satisfy your needs
or one of your customers'. Contributions are welcome, however I'm often very
picky about changes. I will generally reject patches that change massive parts
of the code, or that touch the core parts without any good reason if those
changes have not been discussed first.
The proper place to discuss your changes is the HAProxy Mailing List. There are
enough skilled readers to catch hazardous mistakes and to suggest improvements.
I trust a number of them enough to merge a patch if they say it's OK, so using
the list is the fastest way to get your code reviewed and merged. You can
subscribe to it by sending an empty e-mail at the following address :
haproxy+subscribe@formilux.org
If you have an idea about something to implement, *please* discuss it on the
list first. It has already happened several times that two persons did the same
thing simultaneously. This is a waste of time for both of them. It's also very
common to see some changes rejected because they're done in a way that will
conflict with future evolutions, or that does not leave a good feeling. It's
always unpleasant for the person who did the work, and it is unpleasant for me
too because I value people's time and efforts. That would not happen if these
were discussed first. There is no problem posting work in progress to the list,
it happens quite often in fact. Also, don't waste your time with the doc when
submitting patches for review, only add the doc with the patch you consider
ready to merge.
Another important point concerns code portability. Haproxy requires gcc as the
C compiler, and may or may not work with other compilers. However it's known
to build using gcc 2.95 or any later version. As such, it is important to keep
in mind that certain facilities offered by recent versions must not be used in
the code :
- declarations mixed in the code (requires gcc >= 3.x)
- GCC builtins without checking for their availability based on version and
architecture ;
- assembly code without any alternate portable form for other platforms
- use of stdbool.h, "bool", "false", "true" : simply use "int", "0", "1"
- in general, anything which requires C99 (such as declaring variables in
"for" statements)
Since most of these restrictions are just a matter of coding style, it is
normally not a problem to comply.
If your work is very confidential and you can't publicly discuss it, you can
also mail me directly about it, but your mail may be waiting several days in
the queue before you get a response.
If you'd like a feature to be added but you think you don't have the skills to
implement it yourself, you should follow these steps :
1. discuss the feature on the mailing list. It is possible that someone
else has already implemented it, or that someone will tell you how to
proceed without it, or even why not to do it. It is also possible that
in fact it's quite easy to implement and people will guide you through
the process. That way you'll finally have YOUR patch merged, providing
the feature YOU need.
2. if you really can't code it yourself after discussing it, then you may
consider contacting someone to do the job for you. Some people on the
list might sometimes be OK with trying to do it.
Note to contributors: it's very handy when patches comes with a properly
formated subject. There are 3 criteria of particular importance in any patch :
- its nature (is it a fix for a bug, a new feature, an optimization, ...)
- its importance, which generally reflects the risk of merging/not merging it
- what area it applies to (eg: http, stats, startup, config, doc, ...)
It's important to make these 3 criteria easy to spot in the patch's subject,
because it's the first (and sometimes the only) thing which is read when
reviewing patches to find which ones need to be backported to older versions.
Specifically, bugs must be clearly easy to spot so that they're never missed.
Any patch fixing a bug must have the "BUG" tag in its subject. Most common
patch types include :
- BUG fix for a bug. The severity of the bug should also be indicated
when known. Similarly, if a backport is needed to older versions,
it should be indicated on the last line of the commit message. If
the bug has been identified as a regression brought by a specific
patch or version, this indication will be appreciated too. New
maintenance releases are generally emitted when a few of these
patches are merged.
- CLEANUP code cleanup, silence of warnings, etc... theorically no impact.
These patches will rarely be seen in stable branches, though they
may appear when they remove some annoyance or when they make
backporting easier. By nature, a cleanup is always minor.
- REORG code reorganization. Some blocks may be moved to other places,
some important checks might be swapped, etc... These changes
always present a risk of regression. For this reason, they should
never be mixed with any bug fix nor functional change. Code is
only moved as-is. Indicating the risk of breakage is highly
recommended.
- BUILD updates or fixes for build issues. Changes to makefiles also fall
into this category. The risk of breakage should be indicated if
known. It is also appreciated to indicate what platforms and/or
configurations were tested after the change.
- OPTIM some code was optimised. Sometimes if the regression risk is very
low and the gains significant, such patches may be merged in the
stable branch. Depending on the amount of code changed or replaced
and the level of trust the author has in the change, the risk of
regression should be indicated.
- RELEASE release of a new version (development or stable).
- LICENSE licensing updates (may impact distro packagers).
When the patch cannot be categorized, it's best not to put any tag. This is
commonly the case for new features, which development versions are mostly made
of.
Additionally, the importance of the patch should be indicated when known. A
single upper-case word is preferred, among :
- MINOR minor change, very low risk of impact. It is often the case for
code additions that don't touch live code. For a bug, it generally
indicates an annoyance, nothing more.
- MEDIUM medium risk, may cause unexpected regressions of low importance or
which may quickly be discovered. For a bug, it generally indicates
something odd which requires changing the configuration in an
undesired way to work around the issue.
- MAJOR major risk of hidden regression. This happens when I rearrange
large parts of code, when I play with timeouts, with variable
initializations, etc... We should only exceptionally find such
patches in stable branches. For a bug, it indicates severe
reliability issues for which workarounds are identified with or
without performance impacts.
- CRITICAL medium-term reliability or security is at risk and workarounds,
if they exist, might not always be acceptable. An upgrade is
absolutely required. A maintenance release may be emitted even if
only one of these bugs are fixed. Note that this tag is only used
with bugs. Such patches must indicate what is the first version
affected, and if known, the commit ID which introduced the issue.
If this criterion doesn't apply, it's best not to put it. For instance, most
doc updates and most examples or test files are just added or updated without
any need to qualify a level of importance.
The area the patch applies to is quite important, because some areas are known
to be similar in older versions, suggesting a backport might be desirable, and
conversely, some areas are known to be specific to one version. When the tag is
used alone, uppercase is preferred for readability, otherwise lowercase is fine
too. The following tags are suggested but not limitative :
- doc documentation updates or fixes. No code is affected, no need to
upgrade. These patches can also be sent right after a new feature,
to document it.
- examples example files. Be careful, sometimes these files are packaged.
- tests regression test files. No code is affected, no need to upgrade.
- init initialization code, arguments parsing, etc...
- config configuration parser, mostly used when adding new config keywords
- http the HTTP engine
- stats the stats reporting engine as well as the stats socket CLI
- checks the health checks engine (eg: when adding new checks)
- acl the ACL processing core or some ACLs from other areas
- peers the peer synchronization engine
- listeners everything related to incoming connection settings
- frontend everything related to incoming connection processing
- backend everything related to LB algorithms and server farm
- session session processing and flags (very sensible, be careful)
- server server connection management, queueing
- proxy proxy maintenance (start/stop)
- log log management
- poll any of the pollers
- halog the halog sub-component in the contrib directory
- contrib any addition to the contrib directory
Other names may be invented when more precise indications are meaningful, for
instance : "cookie" which indicates cookie processing in the HTTP core. Last,
indicating the name of the affected file is also a good way to quickly spot
changes. Many commits were already tagged with "stream_sock" or "cfgparse" for
instance.
It is desired that AT LEAST one of the 3 criteria tags is reported in the patch
subject. Ideally, we would have the 3 most often. The two first criteria should
be present before a first colon (':'). If both are present, then they should be
delimited with a slash ('/'). The 3rd criterion (area) should appear next, also
followed by a colon. Thus, all of the following messages are valid :
Examples of messages :
- DOC: document options forwardfor to logasap
- DOC/MAJOR: reorganize the whole document and change indenting
- BUG: stats: connection reset counters must be plain ascii, not HTML
- BUG/MINOR: stats: connection reset counters must be plain ascii, not HTML
- MEDIUM: checks: support multi-packet health check responses
- RELEASE: Released version 1.4.2
- BUILD: stats: stdint is not present on solaris
- OPTIM/MINOR: halog: make fgets parse more bytes by blocks
- REORG/MEDIUM: move syscall redefinition to specific places
Please do not use square brackets anymore around the tags, because they give me
more work when merging patches. By default I'm asking Git to keep them but this
causes trouble when patches are prefixed with the [PATCH] tag because in order
not to store it, I have to hand-edit the patches. So as of now, I will ask Git
to remove whatever is located between square brackets, which implies that any
subject formatted the old way will have its tag stripped out.
In fact, one of the only square bracket tags that still makes sense is '[RFC]'
at the beginning of the subject, when you're asking for someone to review your
change before getting it merged. If the patch is OK to be merged, then I can
merge it as-is and the '[RFC]' tag will automatically be removed. If you don't
want it to be merged at all, you can simply state it in the message, or use an
alternate '[WIP]' tag ("work in progress").
The tags are not rigid, follow your intuition first, anyway I reserve the right
to change them when merging the patch. It may happen that a same patch has a
different tag in two distinct branches. The reason is that a bug in one branch
may just be a cleanup in the other one because the code cannot be triggered.
For a more efficient interaction between the mainline code and your code, I can
only strongly encourage you to try the Git version control system :
http://git-scm.com/
It's very fast, lightweight and lets you undo/redo your work as often as you
want, without making your mistakes visible to the rest of the world. It will
definitely help you contribute quality code and take other people's feedback
in consideration. In order to clone the HAProxy Git repository :
$ git clone http://git.haproxy.org/git/haproxy-1.5.git (stable 1.5)
$ git clone http://git.haproxy.org/git/haproxy.git/ (development)
If you decide to use Git for your developments, then your commit messages will
have the subject line in the format described above, then the whole description
of your work (mainly why you did it) will be in the body. You can directly send
your commits to the mailing list, the format is convenient to read and process.
-- end