Nicola Ken Barozzi <[EMAIL PROTECTED]> wrote on 26/11/2003 02:21:22 AM:
> > http://www.apache.org/dev/committers.html#web > > There are three points that are important AFAIK: > > 1 - Restoring: > Being able to easily restore the site in case it gets screwed. > Doing a CVS checkout is something that anyone here knows how to do, > so it's the main way. Isn't a backup (e.g. tarball, or site dual-posted to archives.apache.org or whatever), a lot easier than getting CVS going for getting the site back? > 2 - Security: > Sites should never be pushed on the main site but pulled from > it. This is for security reasons, so that a cron job can be made > to update the site from a staging server without any access to the > server machine. Now that we have www and cvs on the same machine > though it makes less sense, as they should be two different > machines. Huh? Sites are pushed on the main server all the time using Maven. What's the issue with that from a security perspective? All communications is done over SSH... > 3 - History: > Some have said that we should always keep history of the *generated* > sites, as even the actual results are important. I'm sure that > it will be hard to gen today's site with Maven in 4 years time, as > things will have changes, but it will still be easy to get them from > a CVS archive. Again, a backup or tarball sounds far easier for the way-back-in-time option. [snip] -- dIon Gillard, Multitask Consulting Blog: http://blogs.codehaus.org/people/dion/
