NEMO5 - Xios3 does not produce the right time_counter in grid_U/V files

Versions
1

Hi NEMO users,

We are now using Nemo5 with XIos3 for our regional configuration. The model runs well and produces the right output but the produced NetCDF files have a weird behavior for the grid_U/V files. It seems that the time_counter values are filled with NetCDF missing values. See this Python code for example :

>>> import xarray as xr
>>> import glob
>>> files = glob.glob("out/BS_1m**grid_U**.nc")
>>> ds=xr.open_dataset(files[0], decode_times=False)
>>> ds["time_counter"]
<xarray.DataArray 'time_counter' (time_counter: 12)> Size: 96B
array([9.96921e+36, 9.96921e+36, 9.96921e+36, 9.96921e+36, 9.96921e+36,
       9.96921e+36, 9.96921e+36, 9.96921e+36, 9.96921e+36, 9.96921e+36,
       9.96921e+36, 9.96921e+36])
Coordinates:
    time_instant   (time_counter) float64 96B ...
  * time_counter   (time_counter) float64 96B 9.969e+36 9.969e+36 ... 9.969e+36
    time_centered  (time_counter) float64 96B ...
Attributes:
    axis:           T
    standard_name:  time
    long_name:      Time axis
    calendar:       gregorian
    units:          seconds since 1900-01-01 00:00:00
    time_origin:    1900-01-01 00:00:00
    bounds:         time_counter_bounds
>>> files = glob.glob("out/BS_1m**grid_T**.nc")
>>> ds=xr.open_dataset(files[0], decode_times=False)
>>> ds["time_counter"]
<xarray.DataArray 'time_counter' (time_counter: 12)> Size: 96B
array([1.705406e+09, 1.707955e+09, 1.710504e+09, 1.713139e+09, 1.715774e+09,
       1.718410e+09, 1.721045e+09, 1.723723e+09, 1.726358e+09, 1.728994e+09,
       1.731629e+09, 1.734264e+09])
Coordinates:
    time_centered  (time_counter) float64 96B ...
  * time_counter   (time_counter) float64 96B 1.705e+09 1.708e+09 ... 1.734e+09
    time_instant   (time_counter) float64 96B ...
Attributes:
    axis:           T
    standard_name:  time
    long_name:      Time axis
    calendar:       gregorian
    units:          seconds since 1900-01-01 00:00:00
    time_origin:    1900-01-01 00:00:00
    bounds:         time_counter_bounds

Have anyone encoutered this issue ?

Kind regards,

Guillaume