This commit introduces a sample fetch, `le2dec`, to convert
little-endian binary input samples into their decimal representations.
The function converts the input into a string containing unsigned
integer numbers, with each number derived from a specified number of
input bytes. The numbers are separated using a user-defined separator.
This new sample is achieved by adding a parametrized sample_conv_2dec
function, unifying the logic for be2dec and le2dec converters.
Co-authored-by: Christian Norbert Menges <christian.norbert.menges@sap.com>
[wt: tracked as GH issue #2915]
Signed-off-by: Willy Tarreau <w@1wt.eu>
The aim of this patch is to identify the QUIC traces between the QUIC frontend
and backend parts. Two haproxy instances are created. The c(1|2) http clients
connect to ha1 with TCP frontends and QUIC backends. ha2 embeds two QUIC listeners
with s1 as TCP backend. When the traces are activated, they are dumped to stderr.
Hopefully, they are prefixed by the haproxy instance name (h1 or h2). This is very
useful to identify the QUIC instances.
"HAVE_TCP_MD5SIG" feature is now registered if TCP MD5 signature is
supported. This will help the feature detection in the reg-test script
dedicated to this feature.
First simple VTC file for QUIC reg tests. Two listeners are configured, one without
Retry enabled and the other without. Two clients simply tries to connect to these
listeners to make an basic H3 request.
This patch adds the support for the RFC2385 (Protection of BGP Sessions via
the + TCP MD5 Signature Option) for the listeners and the servers. The
feature is only available on Linux. Keywords are not exposed otherwise.
By setting "tcp-md5sig <password>" option on a bind line, TCP segments of
all connections instantiated from the listening socket will be signed with a
16-byte MD5 digest. The same option can be set on a server line to protect
outgoing connections to the corresponding server.
The primary use case for this option is to allow BGP to protect itself
against the introduction of spoofed TCP segments into the connection
stream. But it can be useful for any very long-lived TCP connections.
A reg-test was added and it will be executed only on linux. All other
targets are excluded.
Add a fourth character to the second column of the "typed output format"
to indicate whether the value results from a volatile or persistent metric
('V' or 'P' characters respectively). A persistent metric means the value
could possibily be preserved across reloads by leveraging a shared memory
between multiple co-processes. Such metrics are identified as "shared" in
the code (since they are possibly shared between multiple co-processes)
Some reg-tests were updated to take that change into account, also, some
outputs in the configuration manual were updated to reflect current
behavior.
When threads are enabled and running on a machine with multiple CCX
or multiple nodes, thread groups are now enabled since 3.3-dev2, causing
load-balancing algorithms to randomly fail due to incoming connections
spreading over multiple groups and using different load balancing indexes.
Let's just force "thread-groups 1" into all configs when threads are
enabled to avoid this.
Using certificates in the jwt_verify converter allows to make use of the
CLI certificate updates, which is still impossible with public keys (the
legacy behavior).
The jwt_verify can now take public certificates as second parameter,
either with actual certificate path (no previously mentioned) or from a
predefined crt-store or from a variable.
As mentioned in [1], the "dispatch" directive from haproxy 1.0 has long
outlived its original purpose and still suffers from a number of technical
limitations (no checks, no SSL, no idle connes etc) and still hinders some
internal evolutions. It's now time to mark it as deprecated, and to remove
it in 3.5 [2]. It was already recommended against in the documentation but
remained popular in raw TCP environments for being shorter to write.
The directive will now cause a warning to be emitted, suggesting an
alternate method involving "server". The warning can be shut using
"expose-deprecated-directives". The rare configs from 1.0 where
"dispatch" is combined with sticky servers using cookies will just
need to set these servers's weights to zero to prevent them from
being selected by the load balancing algorithm. All of this is
explained in the doc with examples.
Two reg tests were using this method, one purposely for this directive,
which now has expose-deprecated-directives, and another one to test the
behavior of idle connections, which was updated to use "server" and
extended to test both "http-reuse never" and "http-reuse always".
[1] https://github.com/orgs/haproxy/discussions/2921
[2] https://github.com/haproxy/wiki/wiki/Breaking-changes
With commit e93f3ea3f8 ("MEDIUM: proxy: deprecate the "transparent" and
"option transparent" directives") this one no longer works as the config
either has to be adjusted to use server 0.0.0.0 or to enable the deprecated
feature. The test used to validate a technical limitation ("transparent"
not supporting shared connections), indicated as being comparable to
"http-reuse never". Let's now duplicate the test for "http-reuse never"
and "http-reuse always" and validate both behaviors.
Take this opportunity to fix a few problems in this config:
- use "nbthread 1": depending on the thread where the connection
arrives, the connection may or may not be reused
- add explicit URLs to the clients so that they can be recognized
in the logs
- add comments to make it clearer what to expect for each test
This patch removes completely the support for the program section, the
parsing of the section as well as the internals in the mworker does not
support it anymore.
The program section was considered dysfonctional and not fully
compatible with the "mworker V3" model. Users that want to run an
external program must use their init system.
The documentation is cleaned up in another patch.
Introduced in:
25bcdb1d9 BUG/MAJOR: h1: Be stricter on request target validation during message parsing
see also:
fbbbc33df REGTESTS: Do not use REQUIRE_VERSION for HAProxy 2.5+
VTest2 (https://github.com/vtest/VTest2) was released and is a remplacement
for VTest. VTest was archived. So let's use the new version now.
If this commit is backported, the 2 following commits must also be
backported:
* 2808e3577 ("REGTESTS: Explicitly allow failing shell commands in some scripts")
* 82c291124 ("REGTESTS: Make the script testing conditional set-var compatible with Vtest2")
Vtest2, that should replaced Vtest in few months, will reject any failing
commands in shell blocks. However, some scripts are executing some commands,
expecting an error to be able to parse the error output. So, now use "set
+e" in those scripts to explicitly state failing commads are expected.
It is just used for non-final commands. At the end, the shell block must
still report a success.
VTest2 will replaced VTest in few months. There is not so much change
expected. One of them is that a User-Agent header is added by default in all
requests, except if an custom one is already set or if "-nouseragent" option
is used. To still be compatible with VTest, it is not possible to use the
option to avoid the header addition. So, a custom user-agent is added in the
last test of "sample_fetches/cond_set_var.vtc" to be sure it will pass with
Vtest and Vtest2. It is mandatory because the request length is tested.
Several users already reported that it would be nice to support
strict-sni in ssl-default-bind-options. However, in order to support
it, we also need an option to disable it.
This patch moves the setting of the option from the strict_sni field
to a flag in the ssl_options field so that it can be inherited from
the default bind options, and adds a new "no-strict-sni" directive to
allow to disable it on a specific "bind" line.
The test file "del_ssl_crt-list.vtc" which already tests both options
was updated to make use of the default option and the no- variant to
confirm everything continues to work.
A test run on a dual-socket EPYC 9845 (2x160 cores) showed that we'll
be facing new limits during the lifetime of 3.2 with our current 16
groups and 256 threads max:
$ cat test.cfg
global
cpu-policy perforamnce
$ ./haproxy -dc -c -f test.cfg
...
Thread CPU Bindings:
Tgrp/Thr Tid CPU set
1/1-32 1-32 32: 0-15,320-335
2/1-32 33-64 32: 16-31,336-351
3/1-32 65-96 32: 32-47,352-367
4/1-32 97-128 32: 48-63,368-383
5/1-32 129-160 32: 64-79,384-399
6/1-32 161-192 32: 80-95,400-415
7/1-32 193-224 32: 96-111,416-431
8/1-32 225-256 32: 112-127,432-447
Raising the default limit to 1024 threads and 32 groups is sufficient
to buy us enough margin for a long time (hopefully, please don't laugh,
you, reader from the future):
$ ./haproxy -dc -c -f test.cfg
...
Thread CPU Bindings:
Tgrp/Thr Tid CPU set
1/1-32 1-32 32: 0-15,320-335
2/1-32 33-64 32: 16-31,336-351
3/1-32 65-96 32: 32-47,352-367
4/1-32 97-128 32: 48-63,368-383
5/1-32 129-160 32: 64-79,384-399
6/1-32 161-192 32: 80-95,400-415
7/1-32 193-224 32: 96-111,416-431
8/1-32 225-256 32: 112-127,432-447
9/1-32 257-288 32: 128-143,448-463
10/1-32 289-320 32: 144-159,464-479
11/1-32 321-352 32: 160-175,480-495
12/1-32 353-384 32: 176-191,496-511
13/1-32 385-416 32: 192-207,512-527
14/1-32 417-448 32: 208-223,528-543
15/1-32 449-480 32: 224-239,544-559
16/1-32 481-512 32: 240-255,560-575
17/1-32 513-544 32: 256-271,576-591
18/1-32 545-576 32: 272-287,592-607
19/1-32 577-608 32: 288-303,608-623
20/1-32 609-640 32: 304-319,624-639
We can change this default now because it has no functional effect
without any configured cpu-policy, so this will only be an opt-in
and it's better to do it now than to have an effect during the
maintenance phase. A tiny effect is a doubling of the number of
pool buckets and stick-table shards internally, which means that
aside slightly reducing contention in these areas, a dump of tables
can enumerate keys in a different order (hence the adjustment in the
vtc).
The only really visible effect is a slightly higher static memory
consumption (29->35 MB on a small config), but that difference
remains even with 50k servers so that's pretty much acceptable.
Thanks to Erwan Velu for the quick tests and the insights!
Recently, work on connection reuses reveals an issue when mixed with
transparent proxy and set-dst. This patch rewrites the related regtests
to be able to catch this now fixed bug.
Note that it is the first regtest which relies on bc_reused recently
introduced sample fetches. This fetch could be reuse in other related
connection reuse regtests to simplify them.
This test brought by commit 8ed1e91efd ("MEDIUM: lb-chash: add directive
hash-preserve-affinity") seems to have hit a limitation of what can be
expressed in vtc, as it would be desirable to have one server response
release two clients at once but the various attempts using barriers
have failed so far. The test seems to work fine locally but still fails
almost 100% of the time on the CI, so it remains timing dependent in
some ways. Tests have been done with nbthread 1, pool-idle-shared off,
http-reuse never (since always fails locally) etc but to no avail. Let's
just mark it broken in case we later figure another way to fix it. It's
still usable locally most of the time, though.
When using hash-based load balancing, requests are always assigned to
the server corresponding to the hash bucket for the balancing key,
without taking maxconn or maxqueue into account, unlike in other load
balancing methods like 'first'. This adds a new backend directive that
can be used to take maxconn and possibly maxqueue in that context. This
can be used when hashing is desired to achieve cache locality, but
sending requests to a different server is preferable to queuing for a
long time or failing requests when the initial server is saturated.
By default, affinity is preserved as was the case previously. When
'hash-preserve-affinity' is set to 'maxqueue', servers are considered
successively in the order of the hash ring until a server that does not
have a full queue is found.
When 'maxconn' is set on a server, queueing cannot be disabled, as
'maxqueue=0' means unlimited. To support picking a different server
when a server is at 'maxconn' irrespective of the queue,
'hash-preserve-affinity' can be set to 'maxconn'.
This is the introduction of "minsize-req" and "minsize-res".
These two options allow you to set the minimum payload size required for
compression to be applied.
This helps save CPU on both server and client sides when the payload does
not need to be compressed.
It was proved in GH #2875 that the regtest was broken, at least for the
server-side abnsz, as the connect() was not performed using the proper
family, which results in kernel refusing to perform the call, while the
reg-test actually succeeds.
Indeed, in the test we used vtest client to connect to haproxy, which
then routed the request to another haproxy instance listening on an
abnsz socket, and this last haproxy was the one to answer the http
request.
As we only used "rxresp" in vtest client, the test succeeded with empty
responses, which was the case due to the server connection failing on the
first haproxy process.
This reverts commit 0b9a75e878.
Thanks to the previous fixes ("REGTESTS: Fix truncated.vtc to send 0-CRLF" and
"BUG/MINOR: mux-h2: Properly handle full or truncated HTX messages on shut"),
this script can be reenabled for FreeBSD.
We never succeed to make the truncated.vtc reg-test work constantly on
the Cirrus FreeBSD CI.
Let's exclude it from the FreeBSD tests so the CI don't break randomly.
A better name was found for the option implemented in ec74438
("MINOR: hlua: add option to preserve bool type from smp to lua")
Indeed, "tune.lua.preserve-smp-bool {on | off}" wasn't explicit enough
nor did it encourage the adoption of the new "fixed" behavior (vs
historical behavior which is now considered as a bug).
Thus it becomes "tune.lua.bool-sample-conversion { normal | pre-3.1-bug }"
which actively encourage users to switch the new behavior after having
patched in-use Lua script if needed. From a technical point of view,
the logic remains the same, as the option currently defaults to
"pre-3.1-bug" to prevent script breakage, and a warning is emitted if
the option isn't set explicily and Lua is used.
Documentation and regtests were updated.
Must be backported in 3.1 with ec74438 and f2838f5 ("REGTESTS: fix
lua-based regtests using tune.lua.smp-preserve-bool")
Because of the previous commit, configs making use of lua script without
setting "tune.lua.smp-preserve-bool" explicitly now raise a warning.
However, since 6f746af91 ("REGTESTS: use -dW by default on every
reg-tests"), regtests are not allowed to raise warnings anymore.
Because of this the CI now fails for every tests that relies on Lua.
To fix this, let's explicitly set the "tune.lua.smp-preserve-bool" for
all tests involving Lua. Here we set the value to "on" because we know
it is safe to do so, and this way it will be future-proof.
If ec7443827 ("MINOR: hlua: add option to preserve bool type from smp to
lua") is backported, then this patch must be backported with it (if it
is not trivial to backport, then simply follow this rule: grep for
"lua-load" in reg-tests directory, then for each match, make sure to set
the tune.smp-preserve-bool tunable in the global section.
User tried to update a PEM, generated automatically. Part of this PEM has LF
line endings, and another part (CA certificate), added by some API, has CRLF
line endings. This has revealed a bug in cli_snd_buf(), see more
details in issue GitHUB #2818. So, let's add an example of such PEM in our
SSL regtest.
Regtest ocsp_auto_update.vtc used to fail here on FreeBSD because the
base64 utility was not installed by default. Once installed it would
still fail because the utility doesn't support -w to wrap lines. Since
the regtest already relies on openssl base64 for a few commands, let's
just rely on it for the other ones. The only limitation is that openssl
freezes on lines longer than 1024 bytes, and doesn't seem to process more
than 255 chars at once, which might be the reason for using base64 -w 1000
in the first place (the script was probably tested like this). Instead
sed is efficient at wrapping long lines and does the job pretty well.
The output was fixed at 72 chars so that the output is also readable on
a terminal for debugging.
The regtest4be_1srv_smtpchk_httpchk_layer47errors.vtc fails on musl
because it reports "Network unreachable" for -EUNREACH while the
check matches "Network is unreachable" as on other OSes. Let's just
replace " is" with ".*". It now works on both glibc and musl.
-Ws on VTest is not working correctly for an unknown reason, the polling
of the NOTIFY_SOCKET seems to timeout, and VTest never receives the
READY message.
This patch disables the reg-tests using -Ws on OS X.
The -W mode implemented in VTest is not reliable anymore, because VTest
waits for the pidfile to be created. But with the new master-worker
mode, this file is created long before haproxy is ready. This can lead
to the test being started too soon, and failing from time to time.
The -Ws option allows to wait for haproxy to deliver a message to VTest
once it is ready.
Every reg-test now runs without any warning, so let's acivate -dW by
default so the new ones will inheritate the option.
This patch reverts 9d511b3c ("REGTESTS: enable -dW on almost all tests
to fail on warnings") and adds -dW in the default HAPROXY_ARGS of
scripts/run-regtests.sh instead.
As master parses now expose-deprecated-directives option, let's emit warning
about deprecated 'progam' section only in case, if this option wasn't set in
the 'global' section. This allows to people, who don't prefer to remove the
'program' section immediately to continue to start the process in zero-warning
mode.
Adjust the warning message accordingly and mcli_start_progs.vtc test. As
expose-deprecated-directives option is a 'global' section keyword, this section
must always precede any 'program' section, if users still continue to keep
'program' section.
This doesn't need to be backported, as related to the latest changes in
the master-worker architecture.
Now that warnings were almost all removed, let's enable zero-warning
via -dW. All tests were adjusted, but two:
- mcli/mcli_start_progs.vtc:
the programs section currently cannot be silenced
- stats/stats-file.vtc:
the warning comes from the stats file itself on comment lines.
All other ones are now OK.
No less than 30 tests were missing timeouts, preventing them from being
started with zero-warning. Since they were not supposed to trigger, they
have been set to 30s so as never to trigger, and now they do not produce
any warning anymore.
This reg-test uses req.len in an HTTP backend. It does work but emits a
warning suggesting that this is ignored, so most likely its days are
counted now. Let's just use req.hdrs,length instead.
The following rules are triggering warnings about content-type being
ignored:
http-request return content-type "text/plain" if { path /def-4 }
http-request return content-type "text/plain" file /dev/null hdr "x-custom-hdr" "%[url]" if { path /empty-file }
Annoyingly, the content-type is mandatory when the file is not empty,
that might be something to revisit in the future to relax at least one
of the rules so that the config doesn't strictly require to know the
file contents upfront.
Two tests were using "timeout {client,server} 1" to forcefully trigger
them, but a forthcoming patch will emit a warning for such small unitless
values, so let's be explicit about the unit.
The regtest "h1or2_to_h1c" contains both an allow and a deny at the end,
likely to help catch rare bugs. But this triggers a warning that we can
silence by placing a condition on the penultimate rule.
