Patch for issue #895

Created by: marcus-oscarsson

The problem described in #895 is due to a retry mechanism in the client that re-sends a request if there are no response after a certain amount of time. The call to execute_entry is in the case of #895 blocking so a retry is made after n minutes, effectively calling the entry twice (or even more depending on the duration of the call)

A parameter _async to make the call non-blocking was added, the task is executed in a greenlet if _async is passed True. The default behavior is to still that execute_entry is blocking, so all logic that need a non blocking behavior need to set _async to True