Stackdio

Fixes a critical bug where deleting a single host in a stack would delete the whole stack

This is the sixth bugfix release for the 0.7 line.

Bugfixes
- Fixed an issue where blueprints that had labels with keys contained in the title would show up multiple times in a `q` filter
- Fixed an issue where volume tagging would fail when changing labels during launch
- The `/api/stacks/<id>/volumes/` endpoint no longer returns 500 responses
- Core provisioning now supports non-upstart OSs
- Core provisioning now supports RHEL 6
- Fixed an issue where termination would retry on failure, but still continue to fail

This is the fifth bugfix release for the 0.7 line. It also includes a couple of new features.

New Features
- Blueprints can now have labels like stacks could previously
- You can now search for stacks and blueprints based on their labels

Bugfixes
- Fixed a bug where stacks that had attached volumes could not be terminated and re-launched
- Fixed a bug where the `bootstrap_args` were incorrectly set when creating cloud imaged
- Fixed a bug where the `git` portion of ssh formula URIs was being stripped off, causing import errors

This is the fourth bugfix release for the 0.7 line.

Fixes:
- Launching & terminating with salt-cloud is now much more stable
- You can now launch a stack from the blueprints page
- Added new AWS instance types
- Fixed an issue with changing stack labels before volumes get created
- All formulas that have a formula version on a stack get copied into the stack's env
- Timing logging is much more accurate
- Update how bootstrap_args get set to allow for use of the stable type

This is the third bugfix release for the 0.7 line.

Fixes:
- Several issues around creating volumes on a stack
- Private formulas now get copied to the stack directory every orchestration, not only if the dir is missing
- A couple issues where the UI would show a server error instead of bad request
- You can now import ssh formulas

This is the second bugfix release for the 0.7 line.

Fixes:
- Fixed an issue that was introduced in 0.7.1 where the timing logging would cause an exception in certain situations
- `/api` endpoints will now _always_ return JSON. (500 responses previously returned html)