problem

currently on termtagging and spellchecking a lot of workers are produced (segment count / 5) * slotcount.

This produces a lot of overhead and DB load in worker organization, also 5 task with 5 minutes per each import, will all be finished after 25 minutes.

solution

Change that, only one TermTagger/SpellCheck Worker per slot is created, and each worker loops over the available segments until all segments are done.