[TriLUG] Keeping track of system changes

[Tom Bryan]

On Wednesday 08 September 2004 10:38 pm, Rick DeNatale wrote:
> > On Wed, 8 Sep 2004 21:46:08 -0400, Rick DeNatale
> >
> > <rick.denatale at gmail.com> wrote:
> > > So do you use a working directory under /home?  You don't check stuff
> > > out into /etc... or do you?

No.  I only just started using CVS with my upgrade from RedHat 7.3 to 9.0.  
And I'm only just now looking at upgrading my systems again.  Since it's just 
for me at home, we'll see how energetic I get.  

> > CVS operates with the concept of a repository.  You could have your
> > repository on /home and then use /etc as your working copy.
>
> Yes, I know that. But, CVS is really suited for version control for
> the source code in software development projects. It's not clear to me
> how to map the configuration files which make up part of the /etc
> tree, and in some cases are scattered around in other places with the
> CVS concept of modules, and how CVS seems to me to want to control the
> total contents of the working directory. 

The CVS Book I keep mentioning at http://cvsbook.red-bean.com/ gives a couple 
of examples.  Both use make or rdist or something to distribute the files to 
the appropriate machines.  It looks like they're really thinking about do it 
yourself ways for multiple admins to maintain multiple systems in the same 
way that developers maintain code: revision history of past changes, ability 
to diff between revisions, simple roll back procedures for any change, and 
some sort of pre/post-commit triggers to verify some correctness.

Like I said, for me, it was just difficult to track what I had to change after 
every install.  With RH 9.0, I just copied the pristine copies of the files 
to a separate working directory and imported them.  Then I made all of my 
changes and committed them.  I'm hoping to be able to use the "vendor branch" 
stuff in CVS to merge my changes in with any changes RedHat/Fedora makes to 
those files in the future.  

---Tom