This can probably be folded into the information about what is running on which server so that it is easier to keep current. That assumes it is current right now - which may not be the case.
To avoid port collisions, developers should not start/stop any services on these two machines. All development should happen on personal machines.
For testing individual commits to master, we will use a machine set up by Chris Colvard. The Continuous Integration server will automatically deploy to this machine so it should always be up-to-date with the last commit to master.
At the end of each sprint cycle, a manual build can be started on the Continuous Integration server which will build, run tests, build documentation, tag in the git repo, release in JIRA (if possible), and deploy to pawpaw. This instance should be stable and only updated at the end of each cycle. This test server is located at http://pawpaw.dlib.indiana.edu.