Apologies in advance for driving this threat off-topic and bike shedding...
On Wed, Jul 16, 2014 at 12:51 AM, Rodger R. Getz <rgetz@xxxxxxxx> wrote:
> Obviously, you need to implement a cron job that tars or bzips up critical
> system files like that and maintain a rotating archive
That sounds suboptimal. Regular backups are important (though I would
opt for a dedicated tool instead of writing a custom script), but this
particular case calls out for at least the use of version control, and
preferably a configuration management tool (e.g. Puppet, Chef,
CFengine, etc). This makes much scaling the number of admins and/or
server incredibly less painful. In addition to recording *what*
changed, you also get a record of *why* changes were made. Also, it
becomes trivial to have identically configured servers for redundancy
or load purposes. I could go on at length, but I'll spare the list.
(Full disclosure: I'm guilty of not taking my own advice on my home
machines. They're backed up, but I abandoned my effort to use Ansible
after a few hours of getting it set up because I got distracted by
some other shiny object.)
--
Ben Cotton
