UFS-dev PR#118 #113
UFS-dev PR#118 #113
Conversation
* update CMEPS * update CMEPS w/ fix for error condition * turning off regional_atmaq_debug: time-out issue on cheyenne.intel *update cmeps build to implement check4nans feature using module stored in cdeps repo * initial sw fix branch * make pio rearranger=box the default for all tests * update configurations for ocnalb changes
…ommunity#1906) * update CMEPS w/ fix for error condition * turning off regional_atmaq_debug: time-out issue on cheyenne.intel *update cmeps build to implement check4nans feature using module stored in cdeps repo * make pio rearranger=box the default for all tests * update configurations for ocnalb changes * set the ocean albedo limit in config and use it to set logical flag * albdir and albdif can be set as non-std values in nems.configure will default to 0.07 and 0.06 if not set * swap signs for latent heat flux used by HYCOM * test latent flux change for hycom
remove mean prefix for fields imported to FV3atm from CICE or CMEPS rename fv3atm export fields for snow and rain as inst
* add cpld_bmark_p8_iau test * set up cpld_gfsv17 c96 iau test * add tests/cpld_control_gfsv17_iau * remove cpld_bmark_p8_iau * sync MOM input template with global-workflow * add a diag_table that works with fms fix
…community#1915) * establish mrfd * bug fixed for mraerosol * updated upp * sync up CICE/CMEPS/MOM6/Stoch
…e-opening ufs-community#1943) (ufs-community#1947) * Allow instantaneous SW and LW fluxes to be used when cpllnd=.true. (re-opening ufs-community#1943) (ufs-community#1947)
|
Expected failures for ufs-community#1894: All cpld and datm tests will need new baselines because of field name changes in the mediator restart files |
|
Expected failures for ufs-community#1915: 003 cpld_control_gfsv17_iau_intel failed in check_result |
|
Automated RT Failure Notification |
|
@mkavulich It looks like the failures are all expected according to the UFS PRs. I'll start the BL creation. |
|
Automated RT Failure Notification |
|
@mkavulich I failed to notice that ufs-community#1916 requires new input data. The new test cpld_control_gfsv17_iau_intel failed to run. Can you please stage the input data on Hera (I apparently lack permissions): cp /scratch2/NAGAPE/epic/UFS-WM_RT/NEMSfv3gfs/input-data-20221101/FV3_input_data/INPUT_L127_mx100/fv_increment*.nc /scratch1/BMC/gmtb/CCPP_regression_testing/NCAR_ufs-weather-model/input_data/20230920/FV3_input_data/INPUT_L127_mx100/ I'm going to commit the failed log. Once the input files are staged, we can create a baseline for the new test only and recommit a successful log. That way, logs will be committed showing all BL creations succeed, if not at the same time. |
|
Sounds like a good plan, I have staged the new data. I also added group write permissions in that directory for the future (I still can't figure out why my umask isn't working as expected, I think I'll open a ticket with Hera helpdesk). Should I move the new baselines that were created so far to |
Yes, please, just in case. |
|
@mkavulich Looks like we hit another snag. I ran into trouble trying to do the individual test, so I just tried to recreate all baselines. We need to do the following: Stage the following files from /scratch2/NAGAPE/epic/UFS-WM_RT/NEMSfv3gfs/input-data-20221101 to /scratch1/BMC/gmtb/CCPP_regression_testing/NCAR_ufs-weather-model/input_data/20230920: We also hit the disk quota, so we need to nuke some stuff. I would suggest nuking the new baselines that we just created (Let's just re-run the tag after we clear space). We also need to clear out /scratch1/BMC/gmtb/CCPP_regression_testing/NCAR_ufs-weather-model/run and make sure we're only keeping the last 2 sets of baselines. Once these things are done, let's add back the hera-intel-BL label. |
|
Crap, sorry about the disk quota thing. I forgot I had disabled the automatic cleanup for debugging, I will re-enable that for the automated runs when I get a chance. I cleaned out all but the last couple runs (and the latest failed baseline), should have plenty of space now for the next while. Do we not need to copy the whole directory Edit: I copied over that whole directory just in case, and started the newer baseline. |
Ya, if it doesn't exist, let's make/copy a directory there. |
|
Automated RT Failure Notification |
|
Automated RT Failure Notification |
Identical to ufs-community#1947 (no BL change)
Contains