Hi Luca,
1) there is no error in error_log during that time. Maybe we need to
change the log level?
2) In this example it's a post request to upload a file. It's the same
request, sometimes it's working without any errors (HTTP 200) and
sometimes an error occurs (HTTP 400 or HTTP 500).
3) Below you find the proxy config:
<VirtualHost x.x.x.x:80>
ServerName example.com
DocumentRoot /var/www/docs/empty
RedirectMatch .*$ https://example.com
</VirtualHost>
<VirtualHost x.x.x.x:443>
ServerName example.com
DocumentRoot /var/www/docs/empty
SSLEngine on
SSLCipherSuite
EECDH:EDH:!RC4!DES!3DES!DES-CBC3-SHA!EDH-RSA-DES-CBC3-SHA!ECDHE-RSA-DES-CBC3-SHA
Header always set Strict-Transport-Security
"max-age=31556926"
SSLCertificateFile "/etc/apache2/ssl/certs/example.crt"
SSLCertificateKeyFile "/etc/apache2/ssl/keys/example.key"
SSLCertificateChainFile "/etc/apache2/ssl/certs/example-ca.crt"
ProxyPreserveHost On
RewriteEngine On
RewriteCond %{REQUEST_METHOD} ^(HEAD|GET) [NC]
RewriteCond %{HTTP:Range} ([0-9]*-[0-9]*)(\s*,\s*[0-9]*-[0-9]*)+ [OR]
RewriteCond %{HTTP:Request-Range}
([0-9]*-[0-9]*)(\s*,\s*[0-9]*-[0-9]*)+
RewriteCond %{REQUEST_METHOD} ^OPTIONS
RewriteRule .* - [F]
RewriteCond %{HTTP:Connection} Upgrade [NC]
RewriteCond %{HTTP:Upgrade} websocket [NC]
RewriteRule /(.*) ws://y.y.y.y:8080/$1 [P,L]
BrowserMatch ".*MSIE.*" \
nokeepalive ssl-unclean-shutdown \
downgrade-1.0 force-response-1.0
ProxyRequests Off
ProxyPass /jmx-console !
ProxyPass /web-console !
AddOutputFilterByType DEFLATE text/html text/xml text/css
text/javascript application/x-javascript application/javascript
ProxyPass / http://y.y.y.y:8080/
ProxyPassReverse / http://y.y.y.y:8080/
</VirtualHost>
4) In case of HTTP 400/HTTP 500 the request is not reaching the access.log
of the EAP/Tomact.
Thanks & best regards
Marco
Von: Luca Toscano <toscano.l...@gmail.com>
An: users@httpd.apache.org
Datum: 31.08.2017 13:44
Betreff: Re: [users@httpd] Problems with Http11NioProtocol and
proxy server
Hi Marco,
2017-08-29 11:07 GMT+02:00 <bu...@dakosy.de>:
Hi,
we've build a web application with JSF 2.1 and RichFaces 4.5.13.Final
running on JBoss EAP 6.4.12. We're also using a Apache HTTP server 2.4.7
as a HTTPS/WSS proxy to access the application for customers.
After we've changed the EAP http connector to
"org.apache.coyote.http11.Http11NioProtocol" (for using websockets/push),
ajax requests are often answered by HTTP 400 (Bad Request) or HTTP 500
(Bad Gateway).
We're switched back to the EAP http connector "HTTP/1.1" and everything is
working again as expected.
If we're accessing the EAP directly (not via proxy) there is also no
problem. It looks like that there is a problem using Http11NioProtocol
with Apache HTTP server as a proxy.
Is there anyone who run into the same problem or has an idea how we can
solove it?
We'd need to know a couple of things to help:
1) Is there any specific error in the httpd error_log that gets logged
when the 400/500s are returned?
2) What kind of requests are you issuing to httpd when you get a 400 and a
500? Are they different?
3) Can you share your configuration? (at least the one for the proxy)
4) Do you find anything useful in the Tomcat logs when httpd returns the
errors?
Luca
