dav svn finally
September 21st, 2007
I took a moment yesterday and tooled around with my apache setup in order to get dav svn up. Directory listing is currently denied as I have projects in there that aren't public. The SVNListParentPath directive doesn't seem to operate properly unless I grant read to /, which then exposes the file structure of everything in the repository. The actual files can't be read as Authz is setup and actually kicks in when you try to access the files past the first level of a project ( say /mystuff/a_file ). My searches on the subject haven't been very enlightening, if anyone's dealt with a similar setup I'd appreciate any tips. This was primarily motivated by experiences with web-only proxies at some sites.
That all being said, here's a list of the trees currently:
- http://svn.hasno.info/ruby-hl7 is the hl7 tree
- http://svn.hasno.info/mephisto is my mephisto trunk and mephisto plugins
- http://svn.hasno.info/fastxml my little hpricot-like libxml library
gentoo and apache 2.2
December 3rd, 2006
After installing, I had to revise my vhost setup due to some change in the vhost handling. Apparently *.hasno.info will cause issues if you have trac.hasno.info and wiki.hasno.info defined as ServerAliases. I've switch over to using mod_rewrite for all of my *.hasno.info web routing.
apache resource usage
May 27th, 2005
A friend of mine asked me recently what setup I would suggest for seperating static content serving and dynamic content serving. The main concern being that dynamic content servicing under apache 2 is more resource intensive than static servicing (pretty obvious eh?). As far as static content delivery, there are far simpler web servers availible that can spit out data just as fast as apache for a lower resource cost (boa, lighttpd). Now you've got two web servers, one httpd for dynamic and one for static, but how are you going to make it seem like one site? mod_rewrite (lighttpd aparently supports url rewriting, so it might also do) to the rescue. If you toss apache in front of those two servers (possibly running on the same machine but sitting on different ports) then you can have that main apache process act as a traffic cop routing requests to the httpd that can service them.
That was just my first shot at the issue, if anyone else has ideas please send them on over.