Skip to content

Latest commit

 

History

History
87 lines (53 loc) · 4.53 KB

Enterprise.md

File metadata and controls

87 lines (53 loc) · 4.53 KB
Raw

Enterprise Deployment Guide

In an enterprise environment, physical servers or long-lasting virtual machines are more likely to be used for a build pipeline (than Docker containers, which we used in our main deployment example). This demo can be set up on separated machines as well, instead of Docker containers, to model an enterprise deployment pipeline.

The Servers

As in our Docker container configuration, the enterprise pipeline environment consists of:

  • A Build Server running Jenkins
  • A Security Server running OpenSCAP and nmap
  • A Compliance Server running GovReady-Q

The Build Server

We leave it to you to install Jenkins on the Build Server. We have found it convenient to start Jenkins using Docker, e.g. with the following command:

docker run --name jenkins --rm -u root \
  -v jenkins-data:/var/jenkins_home \
  -v /var/run/docker.sock:/var/run/docker.sock \
  jenkinsci/blueocean

However, you can install Jenkins in the manner most familiar to you.

The Security Server

In our demonstration, the Security Server is a CentOS 7 machine running a small daemon that performs security scans on other machines at the request of other machines.

On your security server, install Python and nmap, and place security-server/scan-http-daemon.py onto the server. Edit the script to modify the ssh and nmap commands as appropriate to scan the target application running on the Build Server.

The following pre-modified files should also help you get started. They remove some of the configuration we used specifically for the Docker container constellation the demo creates.

Jenkinsfile.separated
security-server/scan-http-daemon.py.separated
security-server/keys/config.separated

Described below is how to use them in your servers. First, we'll create a public/private key pair, because you do not want to use the insecure key pair included in this repo.

Create A Key Pair For Access To The Ephemeral Target App Server

The Target App Server will still be created and run by Jenkins as a Docker container, within the machine that runs Jenkins. A public key will be installed on the Target App Server; you will install the private key on your Security and Monitoring Server.

Run ssh-keygen in a temporary directory on your computer:

ssh-keygen -t ecdsa -b 521 -N '' -f target_ecdsa

You will end up with two files, target_ecdsa and target_ecdsa.pub.

Configure Your Security and Monitoring Server

Copy the target_ecdsa file from the previous step and security-server/keys/config.separated from this repo into the ~/.ssh directory on your Security and Monitoring Server. Rename the config file to config and replace build.example.net in it with the hostname of your Build Server.

Copy security-server/scan-http-daemon.py.separatedfrom this repo to your Security and Monitoring Server. Call it security-server/scan-http-daemon.py and make sure it is executable. Use some method of running and keeping it running. Running it in a screen session is an easy solution while you're prototyping. You may wish to use something more permanent that starts it during server startup, and re-runs it if it exits unexpectedly. (Add some more examples here.)

The nmap port scan in the separated version removes the -p- and -sT flags; they work with the Docker containers and Docker network, but the more conservative default options might be better on your network. Go ahead and tweak as appropriate for your network.

Configure Jenkins

  • Click "Configure".
  • Click "Pipeline".
  • For "Definition", select "Pipeline script" (instead of "Pipeline script from SCM").
  • Copy and paste the content of Jenkinsfile.separated into the text box.
  • Replace the two instances of security.example.netwith the hostname of your Security and Monitoring Server.
  • When setting up the credentials, use the target_ecdsa.pub file you created above, instead of the file included in the repo.

Breaking the Build

To demonstrate a compliance failure, log into the Jenkins server and run:

firewall-cmd --zone=public --add-port=21/tcp --permanent
firewall-cmd --reload

to open port 21 in the software firewall. The open port will fail a business logic rule in the compliance stage of the Jenkinsfile.

To have the build succeed again, close the port:

firewall-cmd --zone=public --remove-port=21/tcp --permanent
firewall-cmd --reload

To check whether the port is open:

firewall-cmd --list-all