problem

While developing TRANSLATE-1654 it happened that a task import was hanging.

Some debugging did not show the problem. In the debug process the Zf_worker was cleaned (a lot of done and defunct workers were contained). After that the import was running again.

Assumption

The worker table also contained defunct termtagger import worker. I assume that the combination of the fact that we seldom have defunct termtagger workers in real live and the complex slot based wake up procedure of termtagging workers lead to that seldom problem.

This must be evaluated and probably fixed.