Use SELFE Masked results

[rsignell-usgs]

It was reported here https://github.com/nctoolbox/nctoolbox/issues/50 that we should have used SELFE results that were masked.

Output stuff on old TESTBEDAPPS server:

[root@testbedapps 2D_varied_manning_windstress_with_wave]# pwd
/data/ftp/upload/Inundation/vims/selfe_tropical/runs/Ike/2D_varied_manning_windstress_with_wave
[root@testbedapps 2D_varied_manning_windstress_with_wave]# du -h
11G     ./inputs/sflux
11G     ./inputs
3.2G    ./config_20121212/masked
673M    ./config_20121212/inputs
2.6G    ./config_20121212/outputs
6.4G    ./config_20121212
7.3M    ./2D_wrong_name
16G     ./zz_config_20121212/outputs
16G     ./zz_config_20121212
15G     ./new_10min/whole_domain
15G     ./new_10min
15G     ./config_20120929/whole_domain
15G     ./config_20120929
9.4M    ./old/new
17M     ./old
19G     ./whole_domain/10min
26G     ./whole_domain

Output stuff on new COMT server:

[testbed@testbed2 Hurricane_Ike_2D_final_run_with_waves]$ du -h
2.6G    ./Output
673M    ./Input

So it appears that the old ./config_20121212/outputs got moved to the new .\Output as both are 2.6GB.

@drf5n, should have instead used ./config_20121212/masked ?

And is this true for all the other SELFE runs as well?

[drf5n]

Looking at it, there are only two masked directories:

/data/ftp/upload/Inundation/vims/selfe_tropical/runs/Ike/2D_varied_manning_windstress_with_wave/config_20121212/masked and 

/data/ftp/upload/Inundation/vims/selfe_tropical/runs/Ike/2D_varied_manning_windstress/config_20121212/masked

I remember that we updated the output processing for newer runs, but for older runs I did the wet-dry masking with NCO.

The 00_dir.ncml file in each parent of them refers to the .../masked/ files, so I think they were the files I should have tagged/flagged/symlinked for copying. I can't remember how the migration process was supposed to work.

[rsignell-usgs]

@drf5n, thanks for taking a look.

Good idea looking at the old 00_dir.html file -- I'm not sure why the files it pointed to didn't get transferred, but I'm also not sure of the process. If it was a manual process, I'm sure a few things could have been missed. I'll move those two masked selfe datasets over and link them so that they appear instead of the non-masked datasets in the catalog.

@brianmckenna, can you please remind us of the process/logic that was used in the transfer? That might help us figure out which datasets still need to be transferred and/or swapped out.

[drf5n]

I think it had manual parts on my end, choosing and copying things to some staging area, that I certainly could have confused myself with.

[rsignell-usgs]

Okay, I renamed the existing comt output as ./Output_orig, copied the data from testbedapps ./config_20121212/masked to comt ./Output, and after copying over the coords.nc file from the old ./Output_orig into the new ./Output, the selfe results are now masked:

[drf5n]

Thanks Rich.