Some notes moved from various documents - please help to enhance ...
To see how your project website is handled:
'ssh people.apache.org; cd /www; ls -l'
These are the sites for each top-level project. These are the generated documents that are served as static files (except for some special CGIs and .htaccess rules).
It is up to your projects how the website is generated and what software is used. The basic requirements for site management are that only committers should be able to modify the site and that notifications of all site changes should be sent to the relevant project mailing lists.
Typically these requirements are satisfied by keeping the site sources
(and often the generated site) in the Subversion repository. If your
project does this, you can update your website by generating the site
and committing the results to svn. Then login to people.apache.org and
run: 'umask 002; svn update /www/$tlp.apache.org'
(or add a cronjob to do that).
These sites are rsync'ed into production on ajax every hour (starting at 11 minutes past). Deletes happen once per day.
You can view your changes before they go into production by using a trick with your web browser. Temporarily set the preferences to use 140.211.11.10:80 as your HTTP proxy. Beware: that IP address and crontab might change at some time - if so then please send a patch for this doc.
Note that sites are owned by your project group, so be sure that each committer has properly configured their UNIX account and especially their umask (0002 is advised), otherwise the file permissions will be incorrect and so other committers in your group will be prevented. Whenever you create files on the server, ensure that the files all belong to the relevant group, that the files are read/write (664) and that the directories are read/write (775) by the group.
There is an effort to coordinate improved site management. Any committer can join the site-dev list to help.
The websites are served from directories under /www/ on
people.apache.org. But you usually do not edit any content in those
directories. Each website is an anonymous checkout of a source module
on the source repository (svn.apache.org) - for example, the
www.apache.org site is an anonymous checkout of
infrastructure/site/trunk/docs. The exception to this rule
is the software distributions, which are not kept in version control.
Then to update websites, you should commit to the appropriate SVN repository
('svn info' to find out which one). Then do a svn update in the
appropriate directory under /www/ to bring the content on
www.apache.org up to date.
Yes, the central config file allows you to use .htaccess
files in your website directories to control configuration.
Of course, reading and parsing an .htaccess file on each
request can slow down the server, so you should consider requesting
an adjustment in the central config file for permanent configuration
changes.
Each day's error and access logs are kept in
www.apache.org:/x1/logs/www/. They are shuffled off each night
to a directory under www.apache.org:/x1/logarchive/www/ and are
then periodically archived to other media.
No formal log analysis is performed, but you are free to grab the logs and do whatever analysis you would like.