search

ioos / comt

IOOS Coastal and Ocean Modeling Testbed. See:
https://github.com/ioos/comt/wiki
MIT License
1 stars 7 forks source link

Problem with variable "h" getting dropped in opendap #38

Closed rsignell-usgs closed 9 years ago

rsignell-usgs commented 9 years ago

This is bizarre, but the variable "h" is getting dropped when we access the netcdf file /data/comt_2/cb_hypoxia/VIMS_CHesROMS-ECB/5year/Output/fengjgr2015.nc via OPeNDAP at http://comt.sura.org/thredds/dodsC/comt_2_full/cb_hypoxia/VIMS_CHesROMS-ECB/5year/Output/fengjgr2015.nc.html with thredds 4.3.16. I discovered this because our NcML file, which adds SGRID info for "h" was giving errors because it was acting like the "h" variable didn't exist yet, so couldn't be modified.

ncdump on the netcdf file has no problems:

cd /data/comt_2/cb_hypoxia/VIMS_CHesROMS-ECB/5year/Output
ncdump -h fengjgr2015.nc | grep bathy
                h:long_name = "bathymetry at RHO-points" ;
rsignell-usgs commented 9 years ago

I tried upgrading the TDS from 4.3.16 to 4.3.23 (and installed ncSOS v1.1 again) and now the NcML at http://comt.sura.org/thredds/dodsC/comt_2_full/cb_hypoxia/VIMS_CHesROMS-ECB/5year/00_dir.ncml.html seems to be working okay.

Note: we don't really care because we access the NcML, but the OPeNDAP URL from datasetScan still doesn't show the variable 'h': http://comt.sura.org/thredds/dodsC/comt_2_full/cb_hypoxia/VIMS_CHesROMS-ECB/5year/Output/fengjgr2015.nc.html

rsignell-usgs commented 9 years ago

the SOS v1.1 service on the new TDS 4.3.23 also seems to be working: http://comt.sura.org/thredds/sos/comt_1_archive_full/inundation_tropical/observations/tropical/netcdf/Ike/NOAA/8723970_Vaca_Key_Ike_WL.nc

rsignell-usgs commented 9 years ago

@brianmckenna , please let me know if there is anything else I should try running to make sure that TDS 4.3.23 is working okay.