problems encountered when using bond/swap commend

Hello professor Steve, Axel and all other lammps users,
I encountered some problems when I try to include bond/swap command into my script. I am using that command to help me relax my long polymer chains (described by bead-spring model). The parameters I used for the polymers are exactly the same as the parameters used in Kremer's 1990 paper.

Here are my problems:
First, sometimes the MD simulation stops, and the error message shows bad FENE bonds.

Second, after the relaxation, I tried to use my post processing code to reconstruct my polymer chians. I found some of my atoms have three bonds, and sometimes, there are more chain ends comparing to the configurations before the MD simulation.

Would you please help me fix these two problems?

Thank you very much

Walter