Dear Manuel, thank you very much for your helpful reply!
In the meantime, you could try to use "-tc 0" which uses the non-parallel code path instead of the parallel one (-tc 1 would call the parallel code path with one thread).
The run with -tc 0 completed without errors. Further investigations from my bio-informatics colleague found out that our input files maybe were not too well-suited for the task.
As our main goal was to test that the installation was successful and everything is fine for the end users, we tried running RazerS on the ecoli data from your paper. It completed without errors. The content of the resulting sam file still needs to be investigated by my colleague.
For now, we assume, that everything is fine. Again, thank you for all your helpful hints. Best, Nomi