What it Does

This tool will generate the following on any EC11 or CS12 server for which it has access to /etc/opscode/pivotal.pem

• orgs
• users
• groups
• clients
• user-org associations

In addition, users will be populated into orgs and org admin groups at the rate specified in setup.yml. Users, clients, and groups within an org will be added more-or-less randomly to the groups created by the tool.

It will capture everything it generates into one or more yml files (currently just testdata/created-object.yml) so that this data can be retrieved and compared compared after an upgrade or migration for any discrepencies.

Usage

vagrant up
vagrant ssh
sudo -s
cd /vagrant
dpkg -i $private-chef-installer.deb
./setup.sh

Then on Enterprise Chef 11 and Chef Server 12:

./setup-ec.sh

On Open Source Chef Server 11:

./setup-osc.sh

Optional

• before running setup.sh, copy setup.yml.example to setup.yml and customize

Upgrade and Delta Capture (Work in Progress)

On the Guest

Once the above is complete, you'll need to capture the current state of data for comparison after the upgrade. To do that, run:

First, capture the current state of the server for comparison post-upgrade. To do that:

vagrant ssh
sudo -s
cd /vagrant
./pre-migration-backup.sh

After this is done (it will take a while), continue with:

sudo -s
private-chef-ctl stop
dpkg -i /vagrant/new-chef-server-core.deb
chef-server-ctl upgrade
chef-server-ctl start

Then capture the server state post-migration (again, this will take some time):

./post-migration-backup.sh

On the Host

Run the diff.rb program to capture the diff. This tool is currently aware of the 11.x -> 12.x upgrade and does some diff suppression for things that are expected but noisy. Perform the following from the repository root:

gem install easy_diff
ruby ./diff.rb  > testdata/delta.yml

The file testdata/delta.yml will contain the diff between pre- and post upgrade.

Notes

• The default setting of 10 orgs takes a while, because we keep running out of precreated orgs under EC 11. It still finishes, it just takes some retries. There is a private-chef.rb in this directory that you can copy in to /etc/opscode before initial reconfigure that will org precreation speed and depth.

TODO

Short Term

• add support for setting custom acls, at minimum on groups but ideally across the range of supported objects.
• org invite creation and validation - is this in ec backup? (yes to 2.x but what about 1.x? )
• OSC 11.2 support for data verification
• Better cookbook data generation
• Environments to OSC and EC data generation
• knife ec backup: better to have a ruby 2 install on this vm so that we can start with 2.x? Currently it's a little broken, getting ec installed.

Longer Term

• better directory structure. Better names for files...
• generate knife.rb for each created user/client. Directories for each?
• node creation and runlists
• simple cookbook generation and upload per-org
• as node client, grab the resolved runlist and make sure it's right
• diff tool - give it a gemfile, etc, etc.
• Diff handling: make diff handling smarter, with awareness of expected differences between two give versions. Currently it's more-or-less hard-coded for EC11 -> CS12 but it should handle a reasonable range of from..target versions

TODONE

• add support for creating clients per org
• add support for creating groups
• add support for group within group memebrshi
• alt2 - can we just knife ec backup and compare before/after?