/etc in CVS

[Ulrich Spörlein]

On Thu, 22.04.2010 at 12:10:50 -0400, Mike Meyer wrote:
> On Thu, 22 Apr 2010 16:41:27 +0200
> Ulrich Spörlein <uqs at spoerlein.net> wrote:
> 
> > On Thu, 22.04.2010 at 12:18:21 +0200, Dag-Erling Smørgrav wrote:
> > > Sergey Babkin <babkin at verizon.net> writes:
> > > > I wonder if a version control system, like SVN, could be used to keep
> > > > track of all the changes in /etc. (Or maybe it already is and I'm
> > > > simply out of date).
> > > 
> > > arch is commonly used for things like this.
> > 
> > I have a .hg directory sitting in / for every machine I usually take
> > care of. hgignore is of course set to *, so only explicitly added files
> > are tracked.
> > 
> > FWIW, I would *strongly* suggest you do *not* use SVN, but a system
> > capable of offline usage, because when the shit hits the fan, you can't
> > rely on a working network.
> 
> My take is the exact opposite: when the shit *really* hits the fan,
> you can't rely on data from the local disk, so you want your VCS data
> to be stored somewhere else.
> 
> Ideally, the VCS has *everything* on the remote server, including data
> about what's checked out where, so your worst case is to reinstall the
> base OS and then just checkout everything.

I take it you have never worked with a distributed SCM then. As there
are checksums that can assure integrity and, more importantly,
distributed SCMs can of course get data over the network from a central
storage.

Also, ZFS and geli provide further data integrity mechanisms for the
truly paranoid.

Regards,
Uli