Fixes using DevOps Tools

[dokuhebi]

After today's face-to-face discussion, I wanted to open an issue to track the need to port the fixes over to DevOps tools like puppet, chef, and others.

[dokuhebi]

My plan is to take the existing bash scripts in "stig-fix-el6/cat1" and move them into "stig-fix-el6/cat1/bash". Then the various scripts for different tools will be in respective tool directories. (Unless people want the directory structure to be "stig-fix-el6/fixes/bash/cat1" and "stig-fix-el6/fixes/puppet/cat1"

[fcaviggia]

The biggest difference between platforms (Amazon Web Services, Puppet, and Bare Metal [Current]) will be the configuration especially the sudoers, pam files, and sshd configurations) are different enough to at least maintain 3 different versions of the script. I was originally looking at forking the project into stig-fix-el6-aws for AWS, stig-fix-el6-puppet, and stig-fix-el6 for bare metal. What do you think? Is there any other input on the idea? I'm trying to go with the will of the community. I'm going to try and get some input from some other Red Hatters as well, specifically Jason Callaway who has taken the scripts and made them work on AWS.

[dokuhebi]

My concern is that as we move beyond puppet into more DevOps tools, there will be too many forks. I would suggest reorganizing the file structure to move the cat1, cat2, cat3, ca4, and manual directories into separate tool platforms (i.e. bash, puppet, chef, ansible). The "config" directory can stay where it is, since the files will be used as inputs across all the platforms.

[ruckc]

It may be better if someone could build a simple DSL describing the stigs,
which could then generate the relevant tool configurations. For example,
if you standardized usage of chown/chmod and "templates" for various files,
it could then be parsed and could automatically generate the relevant tool
specific code. Otherwise you would be spending time trying to make each
tool have identical outcomes, which incurs a potentially higher testing
burden.

On Fri, Aug 1, 2014 at 3:56 PM, Tom Albrecht [email protected]
wrote:

My concern is that as we move beyond puppet into more DevOps tools, there
will be too many forks. I would suggest reorganizing the file structure to
move the cat1, cat2, cat3, ca4, and manual directories into separate tool
platforms (i.e. bash, puppet, chef, ansible). The "config" directory can
stay where it is, since the files will be used as inputs across all the
platforms.

—
Reply to this email directly or view it on GitHub
#12 (comment)
.

[chris-rock]

I propose you have a look at the Hardening Framework. It does the implementations for puppet, chef and ansible. The framework uses the same validation tests to ensure all implementations behave the same. Have a look at our implementation: https://github.com/hardening-io Would be great if you could provide feedback.

[fcaviggia]

Awesome, thanks for the link. I'll pass that project around Red Hat to review.

[chris-rock]

Amazing, let me know if you need anything.