Scheduled Downtime
On Tuesday 24 October 2023 @ 5pm MT the forums will be in read only mode in preparation for the downtime. On Wednesday 25 October 2023 @ 5am MT, this website will be down for maintenance and expected to return online later in the morning.
Normal Operations
The forums are back online with normal operations. If you notice any issues or errors related to the forums, please reach out to help@ucar.edu

Regional Ensemble Spin-up Error: "Attempt to use feature that was not turned on when netCDF was built"

A

ah828

Anthony Holmes
New Member
I am running an ensemble spin-up using the ./CLM5_startup_freerun script included in the DART distribution on Cheyenne. I have subset my surface data files to a region in East Africa using the ./subset_surfdata python tool located in CTSM/tools/contrib for, updating the following attributes:

1687023341299.png

In the CLM5_startup_freerun script, I updated the setup to include the regionally subset surface data and domain files by updating the caseroot, user_nl_datm, and user_nl_clm using the following commands:

1687023860966.png
1687023889511.png
 
A

ah828

Anthony Holmes
New Member
CONTINUED FROM ABOVE:

1687025115632.png

We can confirm the namelists were updated correctly for each ensemble member by navigating to the case directory: /glade/work/anthholmes/cases/CESM/cesm2_regional_f09_f09_mg17_tst_CAM6_60.

I am running the case using CESM (cesm2.2.0) with CTSM release-cesm2.2.01

However, after running the case, I am receiving the following error:

2023-06-16 18:51:33: case.run error ERROR: RUN FAIL: Command 'mpiexec_mpt -p "%g:" -np 4320 omplace -tm open64 /glade/scratch/anthholmes/CESM/cesm2_regional_f09_f09_mg17_tst_CAM6_60/bld/cesm.exe >> cesm.log.$LID 2>&1 ' failed See log file for details: /glade/scratch/anthholmes/CESM/cesm2_regional_f09_f09_mg17_tst_CAM6_60/run/cesm.log.1366239.chadmin1.ib0.cheyenne.ucar.edu.230616-185002

Checking the CESM log, I see the following errors repeatedly

4105:MPT ERROR: Rank 4105(g:4105) is aborting with error code 1. 4105: Process ID: 59703, Host: r3i7n19, Program: /glade/scratch/anthholmes/CESM/cesm2_regional_f09_f09_mg17_tst_CAM6_60/bld/cesm.exe 4105: MPT Version: HPE MPT 2.21 11/28/19 04:21:40 4105: 4105:MPT: --------stack traceback------- 1729: pio_support::pio_die:: myrank= -1 : ERROR: ionf_mod.F90: 235 : 1729: NetCDF: Attempt to use feature that was not turned on when netCDF was built. 1765: pio_support::pio_die:: myrank= -1 : ERROR: ionf_mod.F90: 235 : 1765: NetCDF: Attempt to use feature that was not turned on when netCDF was built.

Does any one know what might be causing this issue? Could it be that the ./subset_surfdata tool is using a different version of netcdf?
 
slevis

slevis

Moderator
I think this error means that your .nc file is of a type that will not work with the model. Try:
"ncdump -k <file>.nc"
You may need to change the file to cdf5 or netcdf classic. Try one and, if it fails, try the other. If it still fails, then I'm out of ideas. To change the file, try:
"nccopy -k cdf5 <file>.nc <file>_cdf5.nc"
 
slevis

slevis

Moderator
Nevertheless, I would have expected a file coming out of the ./subset_data tool to have worked, so let me bring @erik into the conversation in case he has any insights.
 
A

ah828

Anthony Holmes
New Member
slevis said:
Nevertheless, I would have expected a file coming out of the ./subset_data tool to have worked, so let me bring @erik into the conversation in case he has any insights.
Click to expand...
I did check the format and the output from the ./subset_data tool was in netCDF-4 format whereas the original surface dataset is in "64-bit offset" format.

Code: 
(xESMF) anthholmes@cheyenne1:~/CTSM/tools/mkmapgrids> ncdump -k /glade/p/cesmdata/cseg/inputdata/lnd/clm2/surfdata_map/release-clm5.0.18/surfdata_0.9x1.25_hist_78pfts_CMIP6_simyr2000_c190214.nc
64-bit offset

So I converted the file using the following command:

Code: 
(xESMF) anthholmes@cheyenne1:~/CTSM/tools/mkmapgrids> nccopy -k nc6 /glade/scratch/anthholmes/regional/surfdata_0.9x1.25_78pfts_CMIP6_simyr2000_E.Africa.nc /glade/scratch/anthholmes/regional/surfdata_0.9x1.25_78pfts_CMIP6_simyr2000_E.Africa_64-bit.nc

However, my concern is that in writing the nc files to NetCDF-4, the ./subset_data tool would have compressed or distorted the data in some way to make it incompatible with CTSM, is that a possibility?
 
slevis

slevis

Moderator
I don't think so, but also I don't know. You may be able to check for that by carefully inspecting the data with some netcdf viewer (e.g. ncview, python, matlab, etc.).
 
A

ah828

Anthony Holmes
New Member
slevis said:
I don't think so, but also I don't know. You may be able to check for that by carefully inspecting the data with some netcdf viewer (e.g. ncview, python, matlab, etc.).
Click to expand...
Thanks for your suggestions Sam, I really appreciate your help. I will try running with the converted data and will update this thread with the results.
 
A

ah828

Anthony Holmes
New Member
slevis said:
I think this error means that your .nc file is of a type that will not work with the model. Try:
"ncdump -k <file>.nc"
You may need to change the file to cdf5 or netcdf classic. Try one and, if it fails, try the other. If it still fails, then I'm out of ideas. To change the file, try:
"nccopy -k cdf5 <file>.nc <file>_cdf5.nc"
Click to expand...
The model appears to have run without the error "NetCDF: Attempt to use feature that was not turned on when netCDF was built." after updating the surface data and domain files to match the data type of the original files ("classic" for the domain file and "nc6" for the surface data file).

I am getting other errors now related to the time domain of the landcover data and MOSART model, but I will post these issues in a new thread.
 
You must log in or register to reply here.
Top