Apparently some systems define MSG_NOSIGNAL but do not necessarily
check it (or maybe binaries are built somewhere and used on older
versions). There were reports of very recent FreeBSD setups causing
SIGPIPEs, while older ones catch the signal. Recent FreeBSD manpages
indeed define MSG_NOSIGNAL.
So let's now unconditionnaly catch the signal. It's useless not to do
it for the rare cases where it's not needed (linux 2.4 and below).
We are seeing both real servers repeatedly going on- and off-line with
a period of tens of seconds. Packet tracing, stracing, and adding
debug code to HAProxy itself has revealed that the real servers are
always responding correctly, but HAProxy is sometimes receiving only
part of the response.
It appears that the real servers are sending the test page as three
separate packets. HAProxy receives the contents of one, two, or three
packets, apparently randomly. Naturally, the health check only
succeeds when all three packets' data are seen by HAProxy. If HAProxy
and the real servers are modified to use a plain HTML page for the
health check, the response is in the form of a single packet and the
checks do not fail.
(...)
I've added buffer and length variables to struct server, and allocated
space with the rest of the server initialisation.
(...)
It seems to be working fine in my tests, and handles check responses
that are bigger than the buffer.
Marcello Gorlani reported that at least on FreeBSD, a long hostname
was reported with garbage on the stats page. POSIX does not make it
mandatory for gethostname() to NULL-terminate the string in case of
truncation, and at least FreeBSD appears not to do it. So let's
force null-termination to keep safe.
The bounce realign function was algorithmically good but as expected
it was not cache-friendly. Using it with large requests caused so many
cache thrashing that the function itself could drain 70% of the total
CPU time for only 0.5% of the calls !
Revert back to a standard memcpy() using a specially allocated swap
buffer. We're now back to 2M req/s on pipelined requests.
Thich patch fixes cfgparser not to leak memory on each
default server statement and adds several missing free
calls in deinit():
- free(l->name)
- free(l->counters)
- free(p->desc);
- free(p->fwdfor_hdr_name);
None of them are critical, hopefully.
Released version 1.4-rc1 with the following main changes :
- [MEDIUM] add a maintenance mode to servers
- [MINOR] http-auth: last fix was wrong
- [CONTRIB] add base64rev-gen.c that was used to generate the base64rev table.
- [MINOR] Base64 decode
- [MINOR] generic auth support with groups and encrypted passwords
- [MINOR] add ACL_TEST_F_NULL_MATCH
- [MINOR] http-request: allow/deny/auth support for frontend/backend/listen
- [MINOR] acl: add http_auth and http_auth_group
- [MAJOR] use the new auth framework for http stats
- [DOC] add info about userlists, http-request and http_auth/http_auth_group acls
- [STATS] make it possible to change a CLI connection timeout
- [BUG] patterns: copy-paste typo in type conversion arguments
- [MINOR] pattern: make the converter more flexible by supporting void* and int args
- [MINOR] standard: str2mask: string to netmask converter
- [MINOR] pattern: add support for argument parsers for converters
- [MINOR] pattern: add the "ipmask()" converting function
- [MINOR] config: off-by-one in "stick-table" after list of converters
- [CLEANUP] acl, patterns: make use of my_strndup() instead of malloc+memcpy
- [BUG] restore accidentely removed line in last patch !
- [MINOR] checks: make the HTTP check code add the CRLF itself
- [MINOR] checks: add the server's status in the checks
- [BUILD] halog: make without arch-specific optimizations
- [BUG] halog: fix segfault in case of empty log in PCT mode (cherry picked from commit fe362fe476)
- [MINOR] http: disable keep-alive when process is going down
- [MINOR] acl: add build_acl_cond() to make it easier to add ACLs in config
- [CLEANUP] config: use build_acl_cond() instead of parse_acl_cond()
- [CLEANUP] config: use warnif_cond_requires_resp() to check for bad ACLs
- [MINOR] prepare req_*/rsp_* to receive a condition
- [CLEANUP] config: specify correct const char types to warnif_* functions
- [MEDIUM] config: factor out the parsing of 20 req*/rsp* keywords
- [MEDIUM] http: make the request filter loop check for optional conditions
- [MEDIUM] http: add support for conditional request filter execution
- [DOC] add some build info about the AIX platform (cherry picked from commit e41914c77e)
- [MEDIUM] http: add support for conditional request header addition
- [MEDIUM] http: add support for conditional response header rewriting
- [DOC] add some missing ACLs about response header matching
- [MEDIUM] http: add support for proxy authentication
- [MINOR] http-auth: make the 'unless' keyword work as expected
- [CLEANUP] config: use build_acl_cond() to simplify http-request ACL parsing
- [MEDIUM] add support for anonymous ACLs
- [MEDIUM] http: switch to tunnel mode after status 101 responses
- [MEDIUM] http: stricter processing of the CONNECT method
- [BUG] config: reset check request to avoid double free when switching to ssl/sql
- [MINOR] config: fix too large ssl-hello-check message.
- [BUG] fix error response in case of server error
Support the new syntax (http-request allow/deny/auth) in
http stats.
Now it is possible to use the same syntax is the same like in
the frontend/backend http-request access control:
acl src_nagios src 192.168.66.66
acl stats_auth_ok http_auth(L1)
stats http-request allow if src_nagios
stats http-request allow if stats_auth_ok
stats http-request auth realm LB
The old syntax is still supported, but now it is emulated
via private acls and an aditional userlist.
Add generic authentication & authorization support.
Groups are implemented as bitmaps so the count is limited to
sizeof(int)*8 == 32.
Encrypted passwords are supported with libcrypt and crypt(3), so it is
possible to use any method supported by your system. For example modern
Linux/glibc instalations support MD5/SHA-256/SHA-512 and of course classic,
DES-based encryption.
Cyril Bont found that when an error is detected in one config file, it
is also reported in all other ones, which is wrong. The fix obviously
consists in checking the return code from readcfgfile() and not the
accumulator.
Cameron Simpson reported an annoying case where haproxy simply reports
"Error(s) found in configuration file" when the file is not found or
not readable.
Fortunately the parsing function still returns -1 in case of open
error, so we're able to detect the issue from the caller and report
the corresponding errno message.
Some rarely information are stored in fdtab, making it larger for no
reason (source port ranges, remote address, ...). Such information
lie there because the checks can't find them anywhere else. The goal
will be to move these information to the stream interface once the
checks make use of it.
For now, we move them to an fdinfo array. This simple change might
have improved the cache hit ratio a little bit because a 0.5% of
performance increase has measured.
During troubleshooting, it's often useful to get the list of supported
pollers but until now it was required to have a working configuration
first. Since the pollers are known before main() is called, let's list
them with the build options.
This patch implements "description" (proxy and global) and "node" (global)
options, removes "node-name" and adds "show-node" & "show-desc" options
for "stats". It also changes the way the header lines (with proxy name) and
the statistics are displayed, so stats no longer look so clumsy with very
long names.
Instead of "node-name" it is possible to use show-node/show-desc with
an optional parameter that overrides a default node/description.
backend cust-0045
# report specific values for this customer
stats show-node Europe
stats show-desc Master node for Europe, Asia, Africa
Commit 27a674efb8 introduced the ability
to configure buffer sizes. Unfortunately, the pool was created before
the conf was read, so that is was always set to the default size.
In order to fix that, we delay the call to init_buffer(), which is not
a problem since nothing uses it during the initialization.
The new tune.bufsize and tune.maxrewrite global directives allow one to
change the buffer size and the maxrewrite size. Right now, setting bufsize
too low will block stats sockets which will not be able to write at all.
An error checking must be added to buffer_write_chunk() so that if it
cannot write its message to an empty buffer, it causes the caller to abort.
Creating a frontend for the global stats socket will help merge
unix sockets management with the other socket management. Since
frontends are huge structs, we only allocate it if required.
This Linux-specific option was never really used in production and
has since been superseded by new splicing options brought by recent
Linux kernels.
It caused several particular cases in the code because the kernel
would take care of the session without haproxy being able to do
anything on it, which became hard to handle in the new architecture.
Let's simply get rid of it now that there is a replacement available.
Do not exit early at the first error found while checking configuration
validity. This particularly helps spotting multiple wrong tracked server
names at once.
Try not to immediately exit on non-fatal errors while parsing the
global section, so that the user has a chance to get most of the
errors at once, which is quite convenient especially during config
checks with the -c argument. Some other errors such as unresolved
server names also don't make the parser exit too early.
We now support up to 10 distinct configuration files. They are
all loaded in the order defined by -f <file1> -f <file2> ...
This can be useful in order to store global, private, public,
etc... configurations in distinct files.
This is a first step towards support of multiple configuration files.
Now readcfgfile() only reads a file in memory and performs very minimal
parsing. The checks are performed afterwards.
When a new process fails to grab some ports, it sends a signal to
the old process in order to release them. Then it tries to bind
again. If it still fails (eg: one of the ports is bound to a
completely different process), it must send the continue signal
to the old process so that this one re-binds to the ports. This
is correctly done, but the newly bound ports are not released
first, which sometimes causes the old process to remain running
with no port bound. The fix simply consists in unbinding all
ports before sending the signal to the old process.
It is recommended to have -D in init scripts, but -D also implies
quiet mode, which hides warning messages, and both options are now
completely unrelated. Remove the implication to get warnings with
-D.
The small list of signals currently handled by haproxy were processed
as soon as they were received. This has caused trouble with calls to
pool_gc2() occuring in the middle of libc's memory management functions
seldom causing deadlocks preventing the old process from leaving.
Now these signals use the new async signal framework and are called
asynchronously, when there is no risk of recursion. This ensures more
reliable operation, especially for sensible processing such as memory
management.
The byte counters have long been 64-bit to avoid overflows. But with
several sites nowadays, we see session counters wrap around every 10-days
or so. So it was the moment to switch counters to 64-bit, including
error and warning counters which can theorically rise as fast as session
counters even if in practice there is very low risk.
The performance impact should not be noticeable since those counters are
only updated once per session. The stats output have been carefully checked
for proper types on both 32- and 64-bit platforms.
If we get very large data at once, it's almost certain that it's
worthless trying to read again, because we got everything we could
get.
Doing this has made all -EAGAIN disappear from splice reads. The
threshold has been put in the global tunable structures so that if
we one day want to make it accessible from user config, it will be
easy to do so.
On overloaded systems, it sometimes happens that hundreds or thousands
of incoming connections are queued in the system's backlog, and all get
dequeued at once. The problem is that when haproxy processes them and
does not apply any limit, this can take some time and the internal date
does not progress, resulting in wrong timer measures for all sessions.
The most common effect of this is that all of these sessions report a
large request time (around several hundreds of ms) which is in fact
caused by the time spent accepting other connections. This might happen
on shared systems when the machine swaps.
For this reason, we finally apply a reasonable limit even in mono-process
mode. Accepting 100 connections at once is fast enough for extreme cases
and will not cause that much of a trouble when the system is saturated.
The "bind-process" keyword lets the admin select which instances may
run on which process (in multi-process mode). It makes it easier to
more evenly distribute the load across multiple processes by avoiding
having too many listen to the same IP:ports.
Setting "nosplice" in the global section will disable the use of TCP
splicing (both tcpsplice and linux 2.6 splice). The same will be
achieved using the "-dS" parameter on the command line.
The global tuning options right now only concern the polling mechanisms,
and they are not in the global struct itself. It's not very practical to
add other options so let's move them to the global struct and remove
types/polling.h which was not used for anything else.
global.maxconn/4 seems to be a good hint for global.maxpipes when that
one must be guessed. If the limit is reached, it's still possible to
set it manually in the configuration.
Using pipe pools makes pipe management a lot easier. It also allows to
remove quite a bunch of #ifdefs in areas which depended on the presence
or not of support for kernel splicing.
The buffer now holds a pointer to a pipe structure which is always NULL
except if there are still data in the pipe. When it needs to use that
pipe, it dynamically allocates it from the pipe pool. When the data is
consumed, the pipe is immediately released.
That way, there is no need anymore to care about pipe closure upon
session termination, nor about pipe creation when trying to use
splice().
Another immediate advantage of this method is that it considerably
reduces the number of pipes needed to use splice(). Tests have shown
that even with 0.2 pipe per connection, almost all sessions can use
splice(), because the same pipe may be used by several consecutive
calls to splice().
If splicing is enabled in a backend, we need to guess how many
pipes will be needed. We used to rely on fullconn, but this leads
to non-working splicing when fullconn is not specified. So we now
fallback to global.maxconn.
This code provides support for linux 2.6 kernel splicing. This feature
appeared in kernel 2.6.25, but initial implementations were awkward and
buggy. A kernel >= 2.6.29-rc1 is recommended, as well as some optimization
patches.
Using pipes, this code is able to pass network data directly between
sockets. The pipes are a bit annoying to manage (fd creation, release,
...) but finally work quite well.
Preliminary tests show that on high bandwidths, there's a substantial
gain (approx +50%, only +20% with kernel workarounds for corruption
bugs). With 2000 concurrent connections, with Myricom NICs, haproxy
now more easily achieves 4.5 Gbps for 1 process and 6 Gbps for two
processes buffers. 8-9 Gbps are easily reached with smaller numbers
of connections.
We also try to splice out immediately after a splice in by making
profit from the new ability for a data producer to notify the
consumer that data are available. Doing this ensures that the
data are immediately transferred between sockets without latency,
and without having to re-poll. Performance on small packets has
considerably increased due to this method.
Earlier kernels return only one TCP segment at a time in non-blocking
splice-in mode, while newer return as many segments as may fit in the
pipe. To work around this limitation without hurting more recent kernels,
we try to collect as much data as possible, but we stop when we believe
we have read 16 segments, then we forward everything at once. It also
ensures that even upon shutdown or EAGAIN the data will be forwarded.
Some tricks were necessary because the splice() syscall does not make
a difference between missing data and a pipe full, it always returns
EAGAIN. The trick consists in stop polling in case of EAGAIN and a non
empty pipe.
The receiver waits for the buffer to be empty before using the pipe.
This is in order to avoid confusion between buffer data and pipe data.
The BF_EMPTY flag now covers the pipe too.
Right now the code is disabled by default. It needs to be built with
CONFIG_HAP_LINUX_SPLICE, and the instances intented to use splice()
must have "option splice-response" (or option splice-request) enabled.
It is probably desirable to keep a pool of pre-allocated pipes to
avoid having to create them for every session. This will be worked
on later.
Preliminary tests show very good results, even with the kernel
workaround causing one memcpy(). At 3000 connections, performance
has moved from 3.2 Gbps to 4.7 Gbps.
