Water-Ge simulation instability

Dear lammps user,
I’m having the following issue, i built my system with avogadro (Ge slab) and packmol to pack the box with water. Minimization seems to work but as soon as i try to thermalize i see bonds stretching and particles moving crazily.

This is the output, which seems to me almost meaningful:

Setting up cg style minimization …
Unit style : metal
Current step : 0
Per MPI rank memory allocation (min/avg/max) = 134.7 | 134.7 | 134.7 Mbytes
Step Temp KinEng Volume Lx Ly Lz Press E_pair E_vdwl E_mol TotEng E_long E_tail
0 0 0 1000000 50 50 400 159739.63 17460.166 17454.541 0.0091634717 17460.175 -18979.158 -2.1619983
99 0 0 1000000 50 50 400 269.92826 -8246.4694 -6336.2896 4.1266117 -8242.3428 -19071.408 -2.1619983
Loop time of 42.0237 on 1 procs for 99 steps with 16900 atoms

99.7% CPU use with 1 MPI tasks x no OpenMP threads

Minimization stats:
Stopping criterion = energy tolerance
Energy initial, next-to-last, final =
17460.1753297 -8241.55412973 -8242.34278974
Force two-norm initial, final = 9262.94 28.7378
Force max component initial, final = 702.425 4.08492
Final line search alpha, max atom move = 0.011189 0.045706
Iterations, force evaluations = 99 170

MPI task timing breakdown:
Section | min time | avg time | max time |%varavg| %total

Screenshot from 2019-01-15 10-42-18.png

Screenshot from 2019-01-15 10-41-28.png

Screenshot from 2019-01-15 10-41-05.png

Dear lammps user,
I'm having the following issue, i built my system with avogadro (Ge slab) and packmol to pack the box with water. Minimization seems to work but as soon as i try to thermalize i see bonds stretching and particles moving crazily.

there is not enough information here to make any particular assessment
as to what is going wrong. are you certain, your bond topology is
correct and your atom ordering is consistent with the requirements for
TIP4P styles? the latter can be checked by switching to a regular
3-site potential like TIP3P or SPC/E or similar.

axel.