CESM link error during build

[Gabriel.Hes]

Thank you for this clarification.
The first option would be surprising because I have loaded the hdf5 module in the confg_machines.xml file:
<modules compiler="intel">
<command name="load">PrgEnv-intel</command>
<command name="load">craype</command>
<command name="load">cray-mpich</command>
<command name="load">cray-netcdf-hdf5parallel</command>
<command name="load">cray-parallel-netcdf</command>
<command name="load">daint-mc</command>
<command name="load">perftools-base</command>
<command name="load">perftools-preload</command>
</modules>


The file that is trying to be opened is the following,
/scratch/snx3000/ghes/cesm_inputdata/lnd/clm2/surfdata_map/surfdata_360x720cru_
16pfts_Irrig_CMIP6_simyr2000_c210624_2D.nc
It is a netCDF-4 format.

 

[jedwards]

Try converting that file using nccopy

nccopy -k cdf5
/scratch/snx3000/ghes/cesm_inputdata/lnd/clm2/surfdata_map/surfdata_360x720cru_
16pfts_Irrig_CMIP6_simyr2000_c210624_2D.nc
/scratch/snx3000/ghes/cesm_inputdata/lnd/clm2/surfdata_map/surfdata_360x720cru_
16pfts_Irrig_CMIP6_simyr2000_c210624_2D_cdf5.nc

Then add this new file to the model using user_nl_clm

 

[Gabriel.Hes]

I tried to run the model again with the cdf5.nc file. I still have the same error: Abort with message NetCDF: Attempt to use feature that was not turned on when netCDF was built. But looking at the end of the lnd.log file, I believe that it arises because of another nc file:

Reading in urban input data from fsurdat file ...
(GETFIL): attempting to find local file
surfdata_360x720cru_16pfts_Irrig_CMIP6_simyr2000_c210624_2D_cdf5.nc
(GETFIL): using
/scratch/snx3000/ghes/cesm_inputdata/lnd/clm2/surfdata_map/surfdata_360x720cru_
16pfts_Irrig_CMIP6_simyr2000_c210624_2D_cdf5.nc
UrbanInput
/scratch/snx3000/ghes/cesm_inputdata/lnd/clm2/surfdata_map/surfdata_360x720cru_
16pfts_Irrig_CMIP6_simyr2000_c210624_2D_cdf5.nc
Sucessfully read urban input data

Attempting to read PFT physiological data .....
(GETFIL): attempting to find local file ctsm51_params_newz0.c211112.nc
(GETFIL): using
/scratch/snx3000/ghes/cesm_inputdata/lnd/clm2/paramdata/ctsm51_params_newz0.c21
1112.nc

Should I do the same operation (change the file to cdf5) for
/scratch/snx3000/ghes/cesm_inputdata/lnd/clm2/paramdata/ctsm51_params_newz0.c21
1112 ?

 

[jedwards]

Yes. And any other files which you have created locally using hdf5 format.

 

[Gabriel.Hes]

Thank you. This has solved the netcdf error. Now I think that the code runs more or less correctly. I just have a surprising .bin file and no .nc file for the atmospheric output. Do you know what this means?

ghes@daint103:/scratch/snx3000/ghes/builds_cesm2.2.0/I2000Clm51Sp_hcru_hcru_dain t_intel_60_32_SET1/run> ls
atm_modelio.nml
CASEROOT
cesm.exe+30203-2504s
cpl_modelio.nml
datm_in
datm.streams.txt.CLMGSWP3v1.Precip
datm.streams.txt.CLMGSWP3v1.Solar
datm.streams.txt.CLMGSWP3v1.TPQW
datm.streams.txt.presaero.clim_2000
datm.streams.txt.topo.observed
drv_flds_in
drv_in
esp_modelio.nml
finidat_interp_dest.nc
glc_modelio.nml
I2000Clm51Sp_hcru_hcru_daint_intel_60_32_SET1.clm2.h0.1998-01.nc
I2000Clm51Sp_hcru_hcru_daint_intel_60_32_SET1.clm2.r.1998-02-02-00000.nc
I2000Clm51Sp_hcru_hcru_daint_intel_60_32_SET1.clm2.rh0.1998-02-02-00000.nc
I2000Clm51Sp_hcru_hcru_daint_intel_60_32_SET1.cpl.r.1998-02-02-00000.nc
I2000Clm51Sp_hcru_hcru_daint_intel_60_32_SET1.datm.rs1.1998-02-02-00000.bin
I2000Clm51Sp_hcru_hcru_daint_intel_60_32_SET1.mosart.h0.1998-01.nc
I2000Clm51Sp_hcru_hcru_daint_intel_60_32_SET1.mosart.r.1998-02-02-00000.nc
I2000Clm51Sp_hcru_hcru_daint_intel_60_32_SET1.mosart.rh0.1998-02-02-00000.nc
iac_modelio.nml
ice_modelio.nml
lnd_in
lnd_modelio.nml
memory.3.86400.log
mosart_in
ocn_modelio.nml
rof_modelio.nml
rpointer.atm
rpointer.drv
rpointer.lnd
rpointer.rof
seq_maps.rc
timing
wav_modelio.nml

 

[Gabriel.Hes]

Dear Jedwards and Bill Sacks,
Thank you so much for helping me figuring the different issues I faced. Now the CLM5 version is running. Please ignore my last message (I don't have atmospheric outputs because I am forcing clm with atmospheric data). I have two final questions:
Do you know what is the reason why the nc4 format is not suitable and the cdf5 is ?
I plan to force CLM with RCP atmospheric projections. Do you know if this has already been done and if there are compsets made for this purpose?

Kind regards,
Gabriel

 

[sacks]

Hi Gabriel,

Please post your second question (regarding forcing CLM with RCP projections) as a new thread here: CTSM, CLM, MOSART, RTM

For the question about netcdf4 format: Jim may be able to answer this better than I can, but my understanding is that certain options need to be enabled (or at least not disabled) when building the netCDF library on your machine in order for it to support the newer nc4 format; this format is based on hdf5, so one of the requirements is building with support for hdf5 and linking against an hdf5 library. cdf5, in contrast, is (somewhat confusingly) based on the older netcdf3 format, but just allows much larger variables than "classic" netcdf3. So a library can support the cdf5 format without having support for netcdf4 / hdf5. For more information about the different netcdf formats, see NetCDF: FAQ .