POD development: rewriting input model files to wkdir, save_nc, verbose vs. debug

[bitterbark]

A few questions that I've encountered that might be useful to other developers, especially those moving from older versions.

1. Is the framework now always reading in model data and re-writing it to the working directory?
   DEBUG: Fetch <Update with current development efforts #2.blocking_neale:zg500>: found /home/bundy/mdtf/inputdata/model/QBOi.EXP1.AMIP.001/day/QBOi.EXP1.AMIP.001.Z500.day.nc.
   ...
   Writing to $WK_DIR/blocking_neale/day/QBOi.EXP1.AMIP.001.zg500.day.nc
   DEBUG: xr.Dataset.to_netcdf on /project/amp/bundy/mdtf/MDTF_v3.0beta.OBS_WK.20210225/MDTF_QBOi.EXP1.AMIP.001_1990_1994/blocking_neale/day/QBOi.EXP1.AMIP.001.zg500.day.nc
   DEBUG: Successful preprocessor exit on <Update with current development efforts #2.blocking_neale:zg500 (=zg500) @ 1day>.

2. What if save_nc was changed to true when a pod fails?
   I know you recommend that developers use save_nc = true, but after my initial test (of all pods in this version) was too large, I turned it off. So it was quite confusing why the preprocessor exited successfully yet the file was gone.

3. Also, I had been trying to use the input file 'verbose' setting but it wasn't giving much of anything useful. When I found the command line -v I really started getting the output that I wanted. Perhaps the input file verobse setting should have the same effect as the -v?