If you want to convert the GRIB2 to NetCDF, you should know the limitation with both of formants.
One of the NetCDF limitations is its special projections. which mean your data should be just in Mercator, Gaussian,Lambert or Polar projection. otherwise that's impossible to convert your data from GRIB2 to NetCDF.
So I ask you to define you data exactly, then I hope I'll be able to help you.
If you have NCAR command language tools (NCL) installed then use `ncl_convert2nc`. On Linux machines (e.g Debian) the package that contains it is called `ncl_ncarg`.
The usage is:
>> ncl_convert2nc
GRIB2 support has been added in version 4.3.0
You can also look at Climate Data Operators (cdo) or python IRIS package: http://stackoverflow.com/questions/15432587/converting-netcdf-to-grib2
GRIB file can be read into Python, using module "pygrib". For example, as in my script "Merep-rohk.py", where I read and plot ECMWF sea level pressure data.
Data fields can be written into netcdf, using module "netCDF4". Here I attached a script "MSGtoNC.py", which I use to write netcdf from METEOSAT files. To make regular grid in the projection, what I need, I also use an interpolation.
some times, ncl_convert2nc is not work well because of its definition of data type. you can write a ncl scripts and convert it to nc by yourself, it is only need read and write, no additinal work.