The new tooling (Beta1) now allows the user to use different servers which was not possible in prior releases. In prior releases the server that the tooling used was the server referenced in the express configurations (~/.openshift/express.conf and /etc/openshift/express.conf).
There are preferences and project settings that we used in prior tooling that were referring to connections. When referring these values were only using the username (since there was no notion of a server). In the new tooling we now allow referring to a server. We should not break old settings though. To do so we forsee the following strategies:

• connection list (connection pool, that's also used by the OpenShift Explorer):
  The current key/values only use the username. Migration:
  • we keep the existing preference key/values
  • we create a new preference key
  • the connection pool is loading the old and new preferences
  • the connection pool is storing the connections that define connections to the default server to the old preferences

• Project settings:
  The current entry is using the username only. Migration:
  • we keep the existing entry
  • we add a new entry if the user uses a connection to a custom (non-default) server
  • we update the old entry if the user changes the username