Am 19.04.2011 15:41, schrieb Sergiu Dumitriu:
On 04/19/2011 09:09 AM, Andreas Hahn wrote:
Am 19.04.2011 00:08, schrieb Sergiu Dumitriu:
On 04/18/2011 03:52 PM, [email protected] wrote:
Hey,
thanks it works now. At least I do not get any errors at the moment. Could you explain to me what the difference is? (HTTP vs. ajp) When proxying? ajp is a direct connection between httpd and tomcat, which basically lets Tomcat take care of a request on behalf of httpd. The request details are kept in place, including the hostname, the port, whether the connection is secure (SSL/https) or not...
HTTP proxying means that httpd clones the initial client request as a new http request to the Tomcat server. Thus, unless specifically configured to try and pass some details from the initial request, Tomcat will just see this as an ordinary request done to "localhost" on port 8080.
ProxyPass, ProxyPassReverse and ProxyPreserveHost try to keep the same hostname (www.my-domain.de instead of localhost), but don't keep the fact that HTTPS was used. Thanks for the backgrounder !
The problem with the blog is that unlike most parts of the wiki, absolute URLs are used when creating a new category. This means that a full URL, trying to match the requested URL, is computed. Now, since Tomcat/XWiki doesn't really know the original URL, the generated URL will be most likely wrong, pointing to the wrong host and port. Couldn't that / shouldn't that be changed so that its possible to reliably operate XWiki with all of its features behind a httpd ? It does work, if the configuration is correct.
I see. I just checked with my installation that Blog works ok. If I understand correctly that has to do with the fact that I'm just using simple HTTP proxying (and not HTTPS, in this case ajp should be used).
I think a lot of people are doing that (including me). Never tried blogging though .... Are other parts of XWiki also affected by this 'wrong' computation of the url ?