Re: [users@httpd] Urgent !!! - Rewrite issue
Hello Spliff, Well actually to explain the complete scenario, RewriteEngine On RewriteCond %{HTTPS} off RewriteRule (.*) https://dnsname%{REQUEST_URI} The above rewirte is set to redirect all the http requests to https request. It is working fine when I give the http://localhostname it is rewriting to https://dnsname but when I type https://loacalhostname it is redirecting to https:// localhostname I want it to be redirecting to https://dnsname as the ssl certificate CN : is the dnsname. On Sun, Mar 18, 2012 at 10:00 AM, SpliFF spl...@warriorhut.org wrote: On 18/03/12 15:02, aparna Puram wrote: Hello Daniel/Mathijs, There is only one virtual host. server name is :selecvcp1.sw.ericsson.se Dns name for that server is : select.ericsson.se both will redirect to the same apache when I use http:selectvcp1.sw.ericsson.se, I m being redirected to https://select.ericsson.se. However, When i type https://selectvcp1.sw.ericsson.se, I m still being in https://selectvcp1.sw.ericsson.se I m not sure y. Is there any other way around? If you expect https to redirect then your condition: RewriteCond %{SERVER_PORT} !^443$ ... is preventing that. What is the purpose of that condition if not to prevent https from redirecting?
Re: [users@httpd] Urgent !!! - Rewrite issue
I think you're saying you want: RewriteEngine On RewriteCond %{HTTP_HOST} ^localhostname$ RewriteRule (.*) https://dnsname%{REQUEST_URI} As long as you are checking for https or port 443 in the rewrite condition you're not redirecting anywhere on HTTPS requests. If you want HTTPS to redirect then your condition is blocking that. The rules above send ALL requests to the dnsname using https regardless of the scheme actually requested. That seems to be the behaviour you're asking for. This all assumes you are using this rule in your port 80 and port 433 virtualhosts. On 18/03/12 17:45, aparna Puram wrote: Hello Spliff, Well actually to explain the complete scenario, RewriteEngine On RewriteCond %{HTTPS} off RewriteRule (.*) https://dnsname%{REQUEST_URI} The above rewirte is set to redirect all the http requests to https request. It is working fine when I give the http://localhostname it is rewriting to https://dnsname but when I type https://loacalhostname it is redirecting to https://localhostname I want it to be redirecting to https://dnsname as the ssl certificate CN : is the dnsname. On Sun, Mar 18, 2012 at 10:00 AM, SpliFF spl...@warriorhut.org mailto:spl...@warriorhut.org wrote: On 18/03/12 15:02, aparna Puram wrote: Hello Daniel/Mathijs, There is only one virtual host. server name is :selecvcp1.sw.ericsson.se http://selecvcp1.sw.ericsson.se Dns name for that server is : select.ericsson.se http://select.ericsson.se both will redirect to the same apache when I use http:selectvcp1.sw.ericsson.se http://selectvcp1.sw.ericsson.se, I m being redirected to https://select.ericsson.se. However, When i type https://selectvcp1.sw.ericsson.se, I m still being in https://selectvcp1.sw.ericsson.se I m not sure y. Is there any other way around? If you expect https to redirect then your condition: RewriteCond %{SERVER_PORT} !^443$ ... is preventing that. What is the purpose of that condition if not to prevent https from redirecting? - To unsubscribe, e-mail: users-unsubscr...@httpd.apache.org For additional commands, e-mail: users-h...@httpd.apache.org
Re: [users@httpd] Urgent !!! - Rewrite issue
Hello Spliff, The rewirtes that you have specified is performing the same. https://localhost is not being redirected to https://dnsname However, http://localhost is redirected to https://dns name Since the ssl ceritificate is installed only for the dns name, When I am accessing https://localhost, It says ssl is installed for differncet common name, Do you still want to access the site. Please let me know if there is any fix to this. On Sun, Mar 18, 2012 at 12:45 PM, SpliFF spl...@warriorhut.org wrote: I think you're saying you want: RewriteEngine On RewriteCond %{HTTP_HOST} ^localhostname$ RewriteRule (.*) https://dnsname%{REQUEST_URI} As long as you are checking for https or port 443 in the rewrite condition you're not redirecting anywhere on HTTPS requests. If you want HTTPS to redirect then your condition is blocking that. The rules above send ALL requests to the dnsname using https regardless of the scheme actually requested. That seems to be the behaviour you're asking for. This all assumes you are using this rule in your port 80 and port 433 virtualhosts. On 18/03/12 17:45, aparna Puram wrote: Hello Spliff, Well actually to explain the complete scenario, RewriteEngine On RewriteCond %{HTTPS} off RewriteRule (.*) https://dnsname%{REQUEST_URI} The above rewirte is set to redirect all the http requests to https request. It is working fine when I give the http://localhostname it is rewriting to https://dnsname but when I type https://loacalhostname it is redirecting to https:// localhostname I want it to be redirecting to https://dnsname as the ssl certificate CN : is the dnsname. On Sun, Mar 18, 2012 at 10:00 AM, SpliFF spl...@warriorhut.org mailto: spl...@warriorhut.org wrote: On 18/03/12 15:02, aparna Puram wrote: Hello Daniel/Mathijs, There is only one virtual host. server name is :selecvcp1.sw.ericsson.se http://selecvcp1.sw.ericsson.se Dns name for that server is : select.ericsson.se http://select.ericsson.se both will redirect to the same apache when I use http:selectvcp1.sw.ericsson.se http://selectvcp1.sw.ericsson.se, I m being redirected to https://select.ericsson.se. However, When i type https://selectvcp1.sw.ericsson.se, I m still being in https://selectvcp1.sw.ericsson.se I m not sure y. Is there any other way around? If you expect https to redirect then your condition: RewriteCond %{SERVER_PORT} !^443$ ... is preventing that. What is the purpose of that condition if not to prevent https from redirecting?
[users@httpd] httpd crashes with segmentation fault: Incompatible libpng version in application and library
Hi, I've compiled httpd 2.2.21 and php 5.3.10, the version of libpng I used is 1.5.7, and I installed all the three packages under a custom path under /opt. The OS is CentOS 6.2, which contains an official distribution of libpng, the version is 1.2.46. When I start the server and visit programs hosted on it, the error.log keeps being appended with the following messages: *libpng warning: Application was compiled with png.h from libpng-1.5.7 libpng warning: Application is running with png.c from libpng-1.2.46 gd-png: fatal libpng error: Incompatible libpng version in application and library* I know the reason is that the version running with is not the one compiled with. But setting the environment variable LD_LIBRARY_PATH won't solve the problem. Can anybody tell me how to fix this ? I know removing the old version of libpng or recompile the server using the old version may be ok, but there are some other potential problems, so I just want to run my application server with the new version of libpng compiled by my self. Thanks
[users@httpd] Re: site loads correctly using lan ip, but not from other computers on same lan
Hi Serge, thanks for the reply! Have you read into AllowCONNECT I've tried adding it for ports and . does your URL change when you visit the pages. Since I'd expect proxypass and proxypassreverse to refer to the same url I've changed the setting for ProxyPassReverse to match ProxyPass so now my vhost setting for the webapp looks like this with AllowConnect added and ProxyPassReverse edited: VirtualHost *: ServerAdmin ad...@example.com ServerName www.example.com ServerAlias example.com # Indexes + Directory Root. DirectoryIndex index.php DocumentRoot /var/www/html/example.com/wordpress ProxyRequests Off AllowConnect Proxy * Order deny,allow Allow from all /Proxy ProxyPass / http://localhost:/ ProxyPassReverse / http://localhost: http://www.example.com:/ #ProxyPassReverseCookieDomain localhost: www.example.com:/ ProxyPassReverseCookiePath / / /VirtualHost Have you read the logs and could you post any relevant entries After the changes, the behavior is the same, and both the access and error logs don't show any entries for attempts to access 192.168.1.5: from other computers on the network, only for the attempts made from 192.168.1.5 itself. Any other ideas? -Daniel Lee
Re: [users@httpd] [RHEL6.2] SSL handshake failure
Here is what I got when I put the loglevel to debug in httpd.conf === [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_kernel.c(1866): OpenSSL: Handshake: start [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_kernel.c(1874): OpenSSL: Loop: before/accept initialization [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1897): OpenSSL: read 11/11 bytes from BIO#7fa4600011a0 [mem: 7fa460006ac0] (BIO dump follows) [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1830): +-+ [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1869): | : 16 03 00 00 2d 01 00 00-29 03-...). | [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1873): | 0011 - SPACES/NULS [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1875): +-+ [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1897): OpenSSL: read 39/39 bytes from BIO#7fa4600011a0 [mem: 7fa460006acb] (BIO dump follows) [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1830): +-+ [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1869): | : 4f 66 66 ec 02 5d 92 3d-4d db ee c7 10 f5 d5 43 Off..].=M..C | [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1869): | 0010: 3e 16 87 86 7b c9 a0 88-db 60 5a c8 f1 46 10 8f ...{`Z..F.. | [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1869): | 0020: 00 00 02 00 04 01.. | [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1873): | 0039 - SPACES/NULS [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1875): +-+ [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_kernel.c(1884): OpenSSL: Write: SSLv3 read client hello C [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_kernel.c(1903): OpenSSL: Exit: error in SSLv3 read client hello C [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_kernel.c(1903): OpenSSL: Exit: error in SSLv3 read client hello C [Mon Mar 19 06:51:12 2012] [info] [client 10.2.1.2] SSL library error 1 in handshake (server www.example.com:443) [Mon Mar 19 06:51:12 2012] [info] SSL Library Error: 336109761 error:1408A0C1:SSL routines:SSL3_GET_CLIENT_HELLO:no shared cipher Too restrictive SSLCipherSuite or using DSA server certificate? [Mon Mar 19 06:51:12 2012] [info] [client 10.2.1.2] Connection closed to child 2 with abortive shutdown (server www.example.com:443) == quite strange, openssl s_client command can pass the SSL handshake while this java application cannot. openssl version is 0.9.8u Welcome any inputs! Thanks, -Aubrey On Fri, Mar 16, 2012 at 1:50 AM, Mark Montague m...@catseye.org wrote: On March 15, 2012 13:31 , Aubrey Li aubrey...@gmail.com wrote: Thanks for your reply. here is the output of httpd -V. [...] -D HTTPD_ROOT=/export/bench/benchmarks/apache2 -D SUEXEC_BIN=/export/bench/benchmarks/apache2/bin/suexec -D DEFAULT_PIDLOG=logs/httpd.pid -D DEFAULT_SCOREBOARD=logs/apache_runtime_status -D DEFAULT_ERRORLOG=logs/error_log -D AP_TYPES_CONFIG_FILE=conf/mime.types -D SERVER_CONFIG_FILE=conf/httpd.conf I built httpd-2.2.22 on a RHEL6.2 system with SSL enabled. Then I made a client to create a connection to httpd but received a handshake failure report. [...] When I connect the client to the server(RHEL6.2), there is no access_log, no err_log, nothing added in /var/log/messages, it's very weird. So you are saying that you have a file at /export/bench/benchmarks/apache2/conf/httpd.conf that contains all of the correct directives to configure SSL, logging, and appropriate virtual hosts? And you are saying that no logs are appearing at /export/bench/benchmarks/apache2/logs/error_log nor at the location that you specify in your ErrorLog directive in /export/bench/benchmarks/apache2/conf/httpd.conf ? In this case, what user are you starting httpd as? What are the values for the User and Group directives in /export/bench/benchmarks/apache2/conf/httpd.conf ? Do that user and group have write access to the place you are telling this version of httpd to write its error logs? Is this system running any Mandatory Access Control system such as SELinux, AppArmor, Tomoyo, or grsecurity that could be interferring with what this version of httpd is trying to do or where it is trying to do it? If so, then check the log files for the Mandatory Access Control system that you are running to find out what the problem is. Hopefully other people on this list will have additional, and better, suggestions of things to check. -- Mark Montague m...@catseye.org - To unsubscribe, e-mail: users-unsubscr...@httpd.apache.org
Re: [users@httpd] site loads correctly using lan ip, but not from other computers on same lan
Drop the cookie thing and replace localhost with 192.168.1.5 in the proxy lines On Mar 18, 2012 1:12 PM, Serge Fonville serge.fonvi...@gmail.com wrote: Also, does your URL change when you visit the pages. Since I'd expect proxypass and proxypassreverse to refer to the same url Kind regards/met vriendelijke groet, Serge Fonville http://www.sergefonville.nl Convince Google!! They need to add GAL support on Android (star to agree) http://code.google.com/p/android/issues/detail?id=4602 2012/3/18 Serge Fonville serge.fonvi...@gmail.com: Hi, Two things: Have you read into AllowCONNECT Have you read the logs and could you post any relevant entries HTH Kind regards/met vriendelijke groet, Serge Fonville http://www.sergefonville.nl Convince Google!! They need to add GAL support on Android (star to agree) http://code.google.com/p/android/issues/detail?id=4602 2012/3/18 Joon Lee jundo...@gmail.com: Hi, I've got a configuration using ProxyPass for a webapp running on localhost to be accessible from example.com: and a wordpress blog on example.com:80. the httpd.conf vhost setting like this: VirtualHost *: ServerAdmin ad...@example.com ServerName www.example.com ServerAlias example.com # Indexes + Directory Root. DirectoryIndex index.php DocumentRoot /var/www/html/example.com/wordpress ProxyRequests Off Proxy * Order deny,allow Allow from all /Proxy ProxyPass / http://localhost:/ ProxyPassReverse / http://www.example.com:/ #ProxyPassReverseCookieDomain localhost: www.example.com:/ ProxyPassReverseCookiePath / / /VirtualHost VirtualHost *:80 ServerAdmin ad...@example.com ServerName www.example.com ServerAlias example.com # Indexes + Directory Root. DirectoryIndex index.php DocumentRoot /var/www/html/example.com/wordpress /VirtualHost the computer (192.168.1.5) running apache itself is trying to serve wordpress from the standard example.com and proxy a localhost app on port (example.com:) to the internet. the wordpress site works from 192.168.1.5 from anywhere on the network (on port 80), but the localhost app proxy on 192.168.1.5: works on 192.168.1.5 itself but not on other computers within the network. why would it only work on the originating server and not its peers? Als - To unsubscribe, e-mail: users-unsubscr...@httpd.apache.org For additional commands, e-mail: users-h...@httpd.apache.org
Re: [users@httpd] [RHEL6.2] SSL handshake failure
[Mon Mar 19 06:51:12 2012] [info] SSL Library Error: 336109761 error:1408A0C1:SSL routines:SSL3_GET_CLIENT_ HELLO:no shared cipher Too restrictive SSLCipherSuite or using DSA server certificate? Check the SSLCipherSuite directive in your SSL host as the error says it might be too restrictive. Try adding more options. On Mar 19, 2012 2:00 AM, Aubrey Li aubrey...@gmail.com wrote: Here is what I got when I put the loglevel to debug in httpd.conf === [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_kernel.c(1866): OpenSSL: Handshake: start [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_kernel.c(1874): OpenSSL: Loop: before/accept initialization [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1897): OpenSSL: read 11/11 bytes from BIO#7fa4600011a0 [mem: 7fa460006ac0] (BIO dump follows) [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1830): +-+ [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1869): | : 16 03 00 00 2d 01 00 00-29 03-...). | [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1873): | 0011 - SPACES/NULS [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1875): +-+ [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1897): OpenSSL: read 39/39 bytes from BIO#7fa4600011a0 [mem: 7fa460006acb] (BIO dump follows) [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1830): +-+ [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1869): | : 4f 66 66 ec 02 5d 92 3d-4d db ee c7 10 f5 d5 43 Off..].=M..C | [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1869): | 0010: 3e 16 87 86 7b c9 a0 88-db 60 5a c8 f1 46 10 8f ...{`Z..F.. | [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1869): | 0020: 00 00 02 00 04 01.. | [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1873): | 0039 - SPACES/NULS [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1875): +-+ [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_kernel.c(1884): OpenSSL: Write: SSLv3 read client hello C [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_kernel.c(1903): OpenSSL: Exit: error in SSLv3 read client hello C [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_kernel.c(1903): OpenSSL: Exit: error in SSLv3 read client hello C [Mon Mar 19 06:51:12 2012] [info] [client 10.2.1.2] SSL library error 1 in handshake (server www.example.com:443) [Mon Mar 19 06:51:12 2012] [info] SSL Library Error: 336109761 error:1408A0C1:SSL routines:SSL3_GET_CLIENT_HELLO:no shared cipher Too restrictive SSLCipherSuite or using DSA server certificate? [Mon Mar 19 06:51:12 2012] [info] [client 10.2.1.2] Connection closed to child 2 with abortive shutdown (server www.example.com:443) == quite strange, openssl s_client command can pass the SSL handshake while this java application cannot. openssl version is 0.9.8u Welcome any inputs! Thanks, -Aubrey On Fri, Mar 16, 2012 at 1:50 AM, Mark Montague m...@catseye.org wrote: On March 15, 2012 13:31 , Aubrey Li aubrey...@gmail.com wrote: Thanks for your reply. here is the output of httpd -V. [...] -D HTTPD_ROOT=/export/bench/benchmarks/apache2 -D SUEXEC_BIN=/export/bench/benchmarks/apache2/bin/suexec -D DEFAULT_PIDLOG=logs/httpd.pid -D DEFAULT_SCOREBOARD=logs/apache_runtime_status -D DEFAULT_ERRORLOG=logs/error_log -D AP_TYPES_CONFIG_FILE=conf/mime.types -D SERVER_CONFIG_FILE=conf/httpd.conf I built httpd-2.2.22 on a RHEL6.2 system with SSL enabled. Then I made a client to create a connection to httpd but received a handshake failure report. [...] When I connect the client to the server(RHEL6.2), there is no access_log, no err_log, nothing added in /var/log/messages, it's very weird. So you are saying that you have a file at /export/bench/benchmarks/apache2/conf/httpd.conf that contains all of the correct directives to configure SSL, logging, and appropriate virtual hosts? And you are saying that no logs are appearing at /export/bench/benchmarks/apache2/logs/error_log nor at the location that you specify in your ErrorLog directive in /export/bench/benchmarks/apache2/conf/httpd.conf ? In this case, what user are you starting httpd as? What are the values for the User and Group directives in /export/bench/benchmarks/apache2/conf/httpd.conf ? Do that user and group have write access to the place you are telling this version of httpd to write its error logs? Is this system running any Mandatory Access Control system such as SELinux, AppArmor, Tomoyo, or grsecurity that could be interferring with
Re: [users@httpd] Issue with trailing slashes after rewrite
After adapting to my folder structure it should look like RewriteCond %{HTTP_HOST} ^(.*)\.(.*)\.(.*)$ RewriteRule ^/(.*) /%1/$1 [P] Unfortunately neither your original suggestion nor my version seem to cause any redirection. On Sat, Mar 17, 2012 at 3:09 PM, Igor Cicimov icici...@gmail.com wrote: Correction RewriteCond %{HTTP_HOST} ^(.*)\.(.*)\.com$ RewriteRule ^/(.*) /%2/%1/$1 [P,L]
Re: [users@httpd] Issue with trailing slashes after rewrite
On Sun, Mar 18, 2012 at 5:56 PM, Nala Gnirut nala.gni...@gmail.com wrote: After adapting to my folder structure it should look like RewriteCond %{HTTP_HOST} ^(.*)\.(.*)\.(.*)$ RewriteRule ^/(.*) /%1/$1 [P] Unfortunately neither your original suggestion nor my version seem to cause any redirection. The P flag is explicitly used to proxy. If you want to redirect, substitute a full URL and use the R flag instead. - To unsubscribe, e-mail: users-unsubscr...@httpd.apache.org For additional commands, e-mail: users-h...@httpd.apache.org
Re: [users@httpd] Issue with trailing slashes after rewrite
On Sun, Mar 18, 2012 at 11:16 PM, Eric Covener cove...@gmail.com wrote: The P flag is explicitly used to proxy. If you want to redirect, substitute a full URL and use the R flag instead. I need to dynamically change DocumentRoot for some subdomains pointing to the same local path. This rule works (almost) as expected, but has an issue with links to subfolders without trailing slash: RewriteCond %{HTTP_HOST} ^foo\.domain\.com$ RewriteCond %{ENV:REDIRECT_STATUS} ^$ RewriteRule .* /foo%{REQUEST_URI} [QSA,L] That is http://foo.domain.com/bar/ works, while http://foo.domain.com/bar is redirected to a wrong local path (/foo/foo/bar instead of /foo/bar)
[users@httpd] Re: Apache 2.2 on Windows refuses to process more than 5 requests from one client concurrently
Hello there, A coworker of mine was able to find the solution on Google. if you add PerlInterpMax 20 to the httpd.conf, then all 20 concurrent requests will get processed simultaneously. The default value for PerlInterpMax is 5, which means that Apache will not run more than 5 Perl interpreters at the same time. I hope this answer helps someone who runs into the same problem later. Yakov Shapiro On Sun, Mar 4, 2012 at 6:19 PM, Yakov Shapiro yshapir...@gmail.com wrote: Hello, I have been trying to configure Apache 2.2 running on Windows to simultaneously handle a large number - say, a hundred - of incoming web requests from a single UNIX machine. I have already raised ThreadsPerChild on the Apache server to 128. To isolate the problem, I also wrote a short Perl script for Windows, goodnight.pl. It sleeps for 3 seconds, prints a message and exits. I wrote a second scrpit, testharness.pl, to run on UNIX. It forks off 20 processes that each send a single web request to goodnight.pl using Curl. Since my ThreadsPerChild is well above 20, and nobody else is using my Apache server, I expect it to handle all 20 requests to goodnight.pl at once. So testharness.pl should complete in 3 seconds (give or take a little time to establish the connection). If I modify testharness.pl to fork off only one process and send one web request, then it does exit in 3 seconds. However, with 20 simultaneous requests, the test harness takes 12 seconds to finish. The Apache access log (see end of this message) shows that all 20 requests arrive at the same second, but the first five take 3 seconds to process, the next five, 6 seconds, the next five, 9 seconds and the last five, 12 seconds. As far as I can tell, this slowness is not caused by goodnight.pl (which does nothing except sleep, print and exit). If I add benchmarking to goodnight.pl to log the script start time, then the logs show the last five instances of the script are started 9 seconds after the request is received by Apache. I have also ran tcpdump on UNIX to see if this slowness is caused by network issues. It does not appear to be: Apache instantly confirms that it has received the request, but then takes longer than expected (12 seconds instead of 3) to send back a response. I have also checked the load on the Windows machine, using Task Manager. The memory and CPU usage stay well below 75% through my test. Is there anything I can do (or check) to change that behavior, and make all 20 requests process at once? My access log and error log are below. My httpd.conf is attached to this e-mail. My Windows OS is Windows Server Standard, Service Pack 2. It is running Apache HTTP Server 2.2 and ActivePerl version 5.8.8, build 817. I would greatly appreciate any help with this issue. Yakov Shapiro P.S. Below is my Apache error log. To my eye it looks normal: [Sun Mar 04 17:36:06 2012] [notice] Apache/2.2.15 (Win32) mod_perl/2.0.4-dev Perl/v5.10.1 configured -- resuming normal operations [Sun Mar 04 17:36:06 2012] [notice] Server built: Mar 4 2010 11:27:46 [Sun Mar 04 17:36:06 2012] [notice] Parent: Created child process 2684 [Sun Mar 04 17:36:06 2012] [debug] mpm_winnt.c(487): Parent: Sent the scoreboard to the child [Sun Mar 04 17:36:06 2012] [info] Parent: Duplicating socket 228 and sending it to child process 2684 [Sun Mar 04 17:36:06 2012] [debug] mpm_winnt.c(605): Parent: Sent 1 listeners to child 2684 [Sun Mar 04 17:36:06 2012] [notice] Child 2684: Child process is running [Sun Mar 04 17:36:06 2012] [debug] mpm_winnt.c(408): Child 2684: Retrieved our scoreboard from the parent. [Sun Mar 04 17:36:06 2012] [debug] mpm_winnt.c(564): Child 2684: retrieved 1 listeners from parent [Sun Mar 04 17:36:06 2012] [notice] Child 2684: Acquired the start mutex. [Sun Mar 04 17:36:06 2012] [notice] Child 2684: Starting 128 worker threads. [Sun Mar 04 17:36:06 2012] [notice] Child 2684: Starting thread to listen on port 80. Here is my Apache access log. It shows the problem: the requests to the same script are taking progressively longer to process (the first few are done in 3 seconds, the last few take 12 seconds). 172.25.122.29 - - [04/Mar/2012:17:39:04 -0500] GET /perl/goodnight.pl?requestnumber=2 HTTP/1.1 200 26 3 300 172.25.122.29 - - [04/Mar/2012:17:39:04 -0500] GET /perl/goodnight.pl?requestnumber=1 HTTP/1.1 200 26 3 300 172.25.122.29 - - [04/Mar/2012:17:39:04 -0500] GET /perl/goodnight.pl?requestnumber=12 HTTP/1.1 200 26 3 300 172.25.122.29 - - [04/Mar/2012:17:39:04 -0500] GET /perl/goodnight.pl?requestnumber=10 HTTP/1.1 200 26 3 3015625 172.25.122.29 - - [04/Mar/2012:17:39:04 -0500] GET /perl/goodnight.pl?requestnumber=3 HTTP/1.1 200 26 3 3031250 172.25.122.29 - - [04/Mar/2012:17:39:04 -0500] GET /perl/goodnight.pl?requestnumber=4 HTTP/1.1 200 26 5 5984375 172.25.122.29 - - [04/Mar/2012:17:39:04 -0500] GET /perl/goodnight.pl?requestnumber=6 HTTP/1.1 200 26 6 600
Re: [users@httpd] httpd crashes with segmentation fault: Incompatible libpng version in application and library
Two things that come to my mind atm: 1. Link the existing “libpng.so” installed to the one you compiled or 2. Try using FileLoad directive in the httpd.conf to load the correct libpng.so file Igor On Sun, Mar 18, 2012 at 10:24 PM, Lenin lenin@gmail.com wrote: Hi, I've compiled httpd 2.2.21 and php 5.3.10, the version of libpng I used is 1.5.7, and I installed all the three packages under a custom path under /opt. The OS is CentOS 6.2, which contains an official distribution of libpng, the version is 1.2.46. When I start the server and visit programs hosted on it, the error.log keeps being appended with the following messages: *libpng warning: Application was compiled with png.h from libpng-1.5.7 libpng warning: Application is running with png.c from libpng-1.2.46 gd-png: fatal libpng error: Incompatible libpng version in application and library* I know the reason is that the version running with is not the one compiled with. But setting the environment variable LD_LIBRARY_PATH won't solve the problem. Can anybody tell me how to fix this ? I know removing the old version of libpng or recompile the server using the old version may be ok, but there are some other potential problems, so I just want to run my application server with the new version of libpng compiled by my self. Thanks
Re: [users@httpd] httpd crashes with segmentation fault: Incompatible libpng version in application and library
Thanks for your reply. I've tried he first method before, no luck. For the second one, I cannot find any useful information about the 'FileLoad' directive, neither from Google nor the official documentation. Are you sure you didn't misspell the directive ? 2012/3/19 Igor Cicimov icici...@gmail.com Two things that come to my mind atm: 1. Link the existing “libpng.so” installed to the one you compiled or 2. Try using FileLoad directive in the httpd.conf to load the correct libpng.so file Igor On Sun, Mar 18, 2012 at 10:24 PM, Lenin lenin@gmail.com wrote: Hi, I've compiled httpd 2.2.21 and php 5.3.10, the version of libpng I used is 1.5.7, and I installed all the three packages under a custom path under /opt. The OS is CentOS 6.2, which contains an official distribution of libpng, the version is 1.2.46. When I start the server and visit programs hosted on it, the error.log keeps being appended with the following messages: *libpng warning: Application was compiled with png.h from libpng-1.5.7 libpng warning: Application is running with png.c from libpng-1.2.46 gd-png: fatal libpng error: Incompatible libpng version in application and library* I know the reason is that the version running with is not the one compiled with. But setting the environment variable LD_LIBRARY_PATH won't solve the problem. Can anybody tell me how to fix this ? I know removing the old version of libpng or recompile the server using the old version may be ok, but there are some other potential problems, so I just want to run my application server with the new version of libpng compiled by my self. Thanks -- 初从文,三年不第。遂习武,校场发一矢,中鼓吏,逐之出。后学医,有所成,自撰一良方,服之,卒。
Re: [users@httpd] httpd crashes with segmentation fault: Incompatible libpng version in application and library
LoadFile On Mon, Mar 19, 2012 at 12:27 PM, Lenin lenin@gmail.com wrote: Thanks for your reply. I've tried he first method before, no luck. For the second one, I cannot find any useful information about the 'FileLoad' directive, neither from Google nor the official documentation. Are you sure you didn't misspell the directive ? 2012/3/19 Igor Cicimov icici...@gmail.com Two things that come to my mind atm: 1. Link the existing “libpng.so” installed to the one you compiled or 2. Try using FileLoad directive in the httpd.conf to load the correct libpng.so file Igor On Sun, Mar 18, 2012 at 10:24 PM, Lenin lenin@gmail.com wrote: Hi, I've compiled httpd 2.2.21 and php 5.3.10, the version of libpng I used is 1.5.7, and I installed all the three packages under a custom path under /opt. The OS is CentOS 6.2, which contains an official distribution of libpng, the version is 1.2.46. When I start the server and visit programs hosted on it, the error.log keeps being appended with the following messages: *libpng warning: Application was compiled with png.h from libpng-1.5.7 libpng warning: Application is running with png.c from libpng-1.2.46 gd-png: fatal libpng error: Incompatible libpng version in application and library* I know the reason is that the version running with is not the one compiled with. But setting the environment variable LD_LIBRARY_PATH won't solve the problem. Can anybody tell me how to fix this ? I know removing the old version of libpng or recompile the server using the old version may be ok, but there are some other potential problems, so I just want to run my application server with the new version of libpng compiled by my self. Thanks -- 初从文,三年不第。遂习武,校场发一矢,中鼓吏,逐之出。后学医,有所成,自撰一良方,服之,卒。
Re: [users@httpd] [RHEL6.2] SSL handshake failure
Thanks a ton, Igor! I copied SSLCiperSuite in the conf file of httpd_2.2.21, and it works! Cheers, -Aubrey #SSLCipherSuite HIGH:MEDIUM:!aNULL:!MD5 SSLCipherSuite ALL:!ADH:!EXPORT56:RC4+RSA:+HIGH:+MEDIUM:+LOW:+SSLv2:+EXP:+eNULL On Mon, Mar 19, 2012 at 5:50 AM, Igor Cicimov icici...@gmail.com wrote: [Mon Mar 19 06:51:12 2012] [info] SSL Library Error: 336109761 error:1408A0C1:SSL routines:SSL3_GET_CLIENT_ HELLO:no shared cipher Too restrictive SSLCipherSuite or using DSA server certificate? Check the SSLCipherSuite directive in your SSL host as the error says it might be too restrictive. Try adding more options. On Mar 19, 2012 2:00 AM, Aubrey Li aubrey...@gmail.com wrote: Here is what I got when I put the loglevel to debug in httpd.conf === [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_kernel.c(1866): OpenSSL: Handshake: start [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_kernel.c(1874): OpenSSL: Loop: before/accept initialization [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1897): OpenSSL: read 11/11 bytes from BIO#7fa4600011a0 [mem: 7fa460006ac0] (BIO dump follows) [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1830): +-+ [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1869): | : 16 03 00 00 2d 01 00 00-29 03 -...). | [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1873): | 0011 - SPACES/NULS [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1875): +-+ [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1897): OpenSSL: read 39/39 bytes from BIO#7fa4600011a0 [mem: 7fa460006acb] (BIO dump follows) [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1830): +-+ [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1869): | : 4f 66 66 ec 02 5d 92 3d-4d db ee c7 10 f5 d5 43 Off..].=M..C | [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1869): | 0010: 3e 16 87 86 7b c9 a0 88-db 60 5a c8 f1 46 10 8f ...{`Z..F.. | [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1869): | 0020: 00 00 02 00 04 01 .. | [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1873): | 0039 - SPACES/NULS [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_io.c(1875): +-+ [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_kernel.c(1884): OpenSSL: Write: SSLv3 read client hello C [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_kernel.c(1903): OpenSSL: Exit: error in SSLv3 read client hello C [Mon Mar 19 06:51:12 2012] [debug] ssl_engine_kernel.c(1903): OpenSSL: Exit: error in SSLv3 read client hello C [Mon Mar 19 06:51:12 2012] [info] [client 10.2.1.2] SSL library error 1 in handshake (server www.example.com:443) [Mon Mar 19 06:51:12 2012] [info] SSL Library Error: 336109761 error:1408A0C1:SSL routines:SSL3_GET_CLIENT_HELLO:no shared cipher Too restrictive SSLCipherSuite or using DSA server certificate? [Mon Mar 19 06:51:12 2012] [info] [client 10.2.1.2] Connection closed to child 2 with abortive shutdown (server www.example.com:443) == quite strange, openssl s_client command can pass the SSL handshake while this java application cannot. openssl version is 0.9.8u Welcome any inputs! Thanks, -Aubrey On Fri, Mar 16, 2012 at 1:50 AM, Mark Montague m...@catseye.org wrote: On March 15, 2012 13:31 , Aubrey Li aubrey...@gmail.com wrote: Thanks for your reply. here is the output of httpd -V. [...] -D HTTPD_ROOT=/export/bench/benchmarks/apache2 -D SUEXEC_BIN=/export/bench/benchmarks/apache2/bin/suexec -D DEFAULT_PIDLOG=logs/httpd.pid -D DEFAULT_SCOREBOARD=logs/apache_runtime_status -D DEFAULT_ERRORLOG=logs/error_log -D AP_TYPES_CONFIG_FILE=conf/mime.types -D SERVER_CONFIG_FILE=conf/httpd.conf I built httpd-2.2.22 on a RHEL6.2 system with SSL enabled. Then I made a client to create a connection to httpd but received a handshake failure report. [...] When I connect the client to the server(RHEL6.2), there is no access_log, no err_log, nothing added in /var/log/messages, it's very weird. So you are saying that you have a file at /export/bench/benchmarks/apache2/conf/httpd.conf that contains all of the correct directives to configure SSL, logging, and appropriate virtual hosts? And you are saying that no logs are appearing at /export/bench/benchmarks/apache2/logs/error_log nor at the location that you specify in your ErrorLog directive in /export/bench/benchmarks/apache2/conf/httpd.conf ? In this case, what user are you starting httpd as? What are the values for the User and Group directives in