Adapt ServiceTermination functionality to workspaces recovering

[sleshchenko]

Description

Now system manager terminate all services (including Workspace which stop all workspaces) in two cases:

• if the corresponding REST method is called;
• before stopping of a tomcat.

So, since Infrastructure may support recovering of running workspaces it is needed to improve ServiceTermination functionality. It should be possible to:

• STOP all active workspaces If infrastructure doesn't support recovering of running workspaces;
• Do NOT STOP all workspaces if infrastructure supports recovering of running workspaces, but it is needed to wait until all workspace related operations will finish (like starting, stopping);
• STOP all workspace if infrastructure supports recovering BUT there are some changes in API and all workspaces must be restarted. Admin may perform it somehow manually if it is required.

OS and version:

Diagnostics: