Water H epsilon must be 0.0 for pair style lj/cut/tip4p/long

Dear all,
Hi,
An error occurred during my calculation: Water H epsilon must be 0.0 for pair style lj / cut / tip4p / long. I checked my in file and found that I changed epsilon in pair_coeff to 0, and the error still exists. Please help me check what went wrong.Errors and in files are below,thank you for your help.
sincere guanguan

ERROR: Water H epsilon must be 0.0 for pair style lj/cut/tip4p/long (…/pair_lj_cut_tip4p_long.cpp:499)
Last command: minimize 1.0e-12 1.0e-12 1000 1000
in file:

dimension 3
boundary p p p
units real
atom_style full
read_data angle8000.data
neighbor 2 bin
neigh_modify delay 0 every 1

pair_style lj/cut/tip4p/long 1 2 1 1 0.1546 8.5 #otype htype btype atype qdist cutoff (cutoff2)
pair_coeff 1 1 0.0000 0.0000
pair_coeff 1 2 0.0000 0.0000
pair_coeff 2 2 0.1852 3.1589
bond_style harmonic
bond_coeff 1 450 0.9527
angle_style harmonic
angle_coeff 1 55 104.52
kspace_style pppm/tip4p 1.0e-4
kspace_modify gewald 0.001

minimize 1.0e-12 1.0e-12 1000 1000
print “minimize is OK”

fix 1 all shake 0.0001 20 10 b 1 a 2
variable t equal 300.0
velocity all create ${t} 12345 rot yes mom yes dist gaussian
thermo 100
thermo_style custom step cpu etotal ke pe evdwl ecoul elong temp press vol density

dump 1 all atom 100 dump.lammpstrj
fix 2 all npt temp $t t (100*dt) iso 1 1 800
timestep 1
run 20000
write_restart restart1.file

just look at the pair style command in your input. you put the important hint right next to the offending input as a comment.
you set otype to 1 and htype to 2. …and for type 2 the epsilon is NOT zero as LAMMPS is correctly telling you.
axel.