Various tasks which needs to be done at some point
- Write down pinging interface, so we'll know if hosts are responding or
are mostly down (detecting hanging nodes)
- Write down support for rsync progress
- Discovery of nodes which are available for accepting distributed testing
- Test the tests rescheduling, so high-latency nodes would not take part
in that.
- make sure that C-c semantics are ok (nodes are killed properly).
There was an attempt to do so, but it's not tested and not always work.