Am 25.04.2016 um 11:33 schrieb Reindl Harald:

Am 17.04.2016 um 01:26 schrieb Leif Hedstrom:
On Apr 16, 2016, at 4:56 PM, Reindl Harald <[email protected]
<mailto:[email protected]>> wrote:

Am 17.04.2016 um 00:52 schrieb Leif Hedstrom:
On Apr 16, 2016, at 4:44 PM, Reindl Harald <[email protected]
<mailto:[email protected]>> wrote:

Am 17.04.2016 um 00:38 schrieb Leif Hedstrom:
so what gives you ab -c 5 -n 100https://www.thelounge.net/on
<http://www.thelounge.net/on> your client?

It fails too from that CentOS7 box to your box.

[harry@srv-rhsoft:~]$ ab -c 5 -n
100https://docs.trafficserver.apache.org/
<http://docs.trafficserver.apache.org/>
This is ApacheBench, Version 2.3 <$Revision: 1706008 $>
Copyright 1996 Adam Twiss, Zeus Technology
Ltd,http://www.zeustech.net/
Licensed to The Apache Software Foundation,http://www.apache.org/

Benchmarkingdocs.trafficserver.apache.org
<http://benchmarkingdocs.trafficserver.apache.org>
<http://docs.trafficserver.apache.org/>(be patient)...^C

Server Software:        ATS/6.2.0
Server Hostname: docs.trafficserver.apache.org
<http://docs.trafficserver.apache.org/>
Server Port:            443
SSL/TLS Protocol:       TLSv1.2,ECDHE-RSA-AES128-GCM-SHA256,2048,128

So it seems your ab works against this CentOS7 box running ATS
v6.2.0 ?

yes

What version of OpenSSL did you build ATS with? I am running
docs.trafficserver.a.o with OpenSSL v1.0.2g if I recall (latest
stable
release)

seems not matter that much since i have this issue for years now and
the httpd servers are built in the same environments with the same
libraries and don't have that issue

Wonder if it could be one of those -f compiler flags? I’m attaching
my config.nice that I run on docs.trafficserver, this compiles with
ASAN though, so you likely want to remove that at least (if you are
willing to try).

i will give it a try ASAP, however the whole web and mail stack is
built with that flags (based on the flags below which are %{optflags}
and only ATS has the specific problem

Yeah, it seems odd that it’d break like that because of compiler flags.
But I honestly have no other ideas as to why it breaks on your system,
and not mine :-/. Can anyone else confirm or deny this breakage on their
installs?

just a notice again before i try to build with other flags
_____________________________________________

https://www.ssllabs.com/ssltest/

docs.trafficserver.apache.org:
SSL 2 handshake compatibility     Yes

www.thelounge.net:
SSL 2 handshake compatibility     No
_____________________________________________

CONFIG proxy.config.ssl.SSLv2 INT 0
CONFIG proxy.config.ssl.SSLv3 INT 0
CONFIG proxy.config.ssl.TLSv1 INT 1
CONFIG proxy.config.ssl.TLSv1_1 INT 1
CONFIG proxy.config.ssl.TLSv1_2 INT 1
CONFIG proxy.config.ssl.client.certification_level INT 0
CONFIG proxy.config.ssl.server.multicert.filename STRING
ssl_multicert.config
CONFIG proxy.config.ssl.server.cert.path STRING /etc/trafficserver/ssl/
CONFIG proxy.config.ssl.server.private_key.path STRING
/etc/trafficserver/ssl/
CONFIG proxy.config.ssl.CA.cert.path STRING /etc/trafficserver/ssl/
CONFIG proxy.config.ssl.server.cipher_suite STRING
ECDHE-RSA-AES128-GCM-SHA256:ECDHE-RSA-AES256-GCM-SHA384:DHE-RSA-AES128-GCM-SHA256:DHE-RSA-AES256-GCM-SHA384:ECDHE-RSA-AES128-SHA256:ECDHE-RSA-AES256-SHA384:ECD$

CONFIG proxy.config.ssl.server.honor_cipher_order INT 1
CONFIG proxy.config.ssl.server.dhparams_file STRING
/etc/trafficserver/ssl/dhparams.pem

no idea - same with the build below - with -fsanitize=address it becomes fat and unstable (no client connection at all)

#export CFLAGS="-g0 -fstack-protector -fsanitize=address -fno-omit-frame-pointer" #export CXXFLAGS="-g0 -fstack-protector -fsanitize=address -fno-omit-frame-pointer"
#export CPPFLAGS="-D_FORTIFY_SOURCE=2"
#export LDFLAGS="-Wl,-z,relro -Wl,-z,now"

export CFLAGS="-m64 -O2 -g0 -pipe -fstack-protector -fno-strict-aliasing -fno-omit-frame-pointer -D_FORTIFY_SOURCE=2"
export CXXFLAGS="$CFLAGS"
export CPPFLAGS="-D_FORTIFY_SOURCE=2"
export LDFLAGS="-Wl,-z,relro -Wl,-z,now"

Attachment: signature.asc
Description: OpenPGP digital signature

Reply via email to