Reported by user:
Hello Wind Ninja team,
I am using windninja CLI on Ubuntu 20.04. I have set up wind ninja on a server to automatically run an hourly point initialization of station points collected and stored on disk for each of the 4 counties in Hawaii.
So far it has been running fine on our largest domain Hawaii county (BI) however when I run any other domain / dem, such as Kauai (KA) windninja fails quickly without any message except a 139 code. I am baffled because the cmd windninja call is identical, except dem, input files and paths. I even went as far as to recreate and align the dem for both BI and KA in the same exact manner with the same resolution, and crs. I also checked and compared formatting and data of the station input files for both. Even after this only BI allowed a windninja run to proceed, and KA failed with a 139 exit code which I assume is C++ indicating a segmentation fault?
Below are the windninja commands for the BI and KA respectively. Additionally I also attached a zipped folder with the directories and files in the locations mentioned in the windninja cmd (without dirs: /home/wn1/). This includes station input files, and dem used in this Apr 8th 2025 23:00 hourly run. Just to reiterate, using this cmd and respective input data (included) BI ran fine but KA fails with a 139 exit code. Except for the BI, this happens for all other domains/dem for any date-time I tried. Is there a bug I need to know about or something missing or need to fix? Any ideas would be a great help. Thanks again for all your hard work, this is a great program and project.
BI windninja cmd (works) :
/home/wn1/wn_build/src/cli/WindNinja_cli --num_threads=1 --elevation_file=/home/wn1/data/dem/BI_utm_z4.tif --initialization_method=pointInitialization --time_zone=UTC --match_points=true --fetch_type=stid --wx_station_filename=/home/wn1/data/input/wn_input/20250407_2300/BI/BI_stations.csv --write_wx_station_kml=false --output_wind_height=10.0 --units_output_wind_height=m --vegetation=grass --mesh_resolution=250 --units_mesh_resolution=m --ascii_out_resolution=250 --units_ascii_out_resolution=m --write_goog_output=false --write_shapefile_output=false --write_ascii_output=true --write_farsite_atm=false --output_path=/home/wn1/data/output/wn_out/20250407_2300/BI --momentum_flag=false --start_year=2025 --start_month=04 --start_day=07 --start_hour=23 --start_minute=00 --stop_year=2025 --stop_month=04 --stop_day=08 --stop_hour=00 --stop_minute=00 --number_time_steps=1
KA windninja cmd (fails with 139 exit code):
/home/wn1/wn_build/src/cli/WindNinja_cli --num_threads=1 --elevation_file=/home/wn1/data/dem/KA_utm_z4.tif --initialization_method=pointInitialization --time_zone=UTC --match_points=true --fetch_type=stid --wx_station_filename=/home/wn1/data/input/wn_input/20250407_2300/KA/KA_stations.csv --write_wx_station_kml=false --output_wind_height=10.0 --units_output_wind_height=m --vegetation=grass --mesh_resolution=250 --units_mesh_resolution=m --ascii_out_resolution=250 --units_ascii_out_resolution=m --write_goog_output=false --write_shapefile_output=false --write_ascii_output=true --write_farsite_atm=false --output_path=/home/wn1/data/output/wn_out/20250407_2300/KA --momentum_flag=false --start_year=2025 --start_month=04 --start_day=07 --start_hour=23 --start_minute=00 --stop_year=2025 --stop_month=04 --stop_day=08 --stop_hour=00 --stop_minute=00 --number_time_steps=1
Answer:
The issue is that you have the stop time set to 04/08/2025, but you only have data for KA at 04/07/2025 2300Z. If you change your start and stop times to the single time step you have (04/07/2025 2300Z) the simulation runs. The station files for your other DEM have at least one observation at 04/08/2025 0000z which is why that one runs. We need to add a better check on the start/stop times compared to the station times and properly handle and report this error.
See data attached to reproduce the bug.
data.zip
Reported by user:
Answer:
See data attached to reproduce the bug.
data.zip