This setting gets propogated to the launch configuration and is currently
causing unexpected behaviour.

Sometimes we get an instance id back from the creation call but then
get an exception saying the ID was not found when we do a get_all_instances.

Set params on BlockDeviceType instead of BlockDeviceMapping.

Use SSD backed EBS instead of standard.
Delete the volume upon abbey termination.

At 30 minutes we still hit it sometimes when building from a bare ubuntu box.

Defaulting it to hammer for now.

The instance gets tagged with the repo and hash that was deployed.

Instead, we take any extra_vars that have a suffix of '_version' and
make sure those exist as tags on the AMI. All other vars do not get
saved anywhere.

Turn on instance tagging with abbey.

Also use the proper return code when abbey has issues.