Problem with combining coordination numbers as one collective variable

[wxwth]

Dear CP2K developers,
I'm trying to use the &COMBINE_COLVAR section in CP2K (version 2022.1) to combine two coordination numbers as one collective variable (CV), and output the Lagrange multipliers of this CV. However, I encounter problems in both the AIMD simulation and the Lagrange multipliers. I update an input file as an example to explain these problems. The input file is for an AIMD simulation of a simple system containing 64 water molecules. The difference between the O-H coordination numbers of the O atoms in two water molecules is chosen as the CV (by using the &COMBINE_COLVAR section). It is reasonable to set the difference to be zero, i.e., these two O atoms have same O-H coordination numbers.
However, after each MD step, CP2K output several warnings with same text:

More seriously, all outputted Lagrange multipliers (both Shake and Rattle) are equal to zero:

I am wondering if there are any problems in my input file, or in CP2K? Thanks in advance!
Sincerely,
Xu

64H2O_CV-CN.inp.txt