Some of the scenarios have a GraphInfoDir setting which points to nonexistent directories, causing workflow failures.
There are also a number of different directories being pointed to by various scenarios, which we may want to change to be uniform?
Another thing, I noted that in the reference tests you mentioned about the graph file missing. This is because is using /glade/work/liuz/pandac/prepare_mpas/mpas_static_30km as the path for the GraphInfoDir variable, but this folder does not exist anymore. This is also used in scenarios/VerifyingGFSAnalysesFromRDA.yaml. I would suggest to remove the path from the scenario yaml files and use the default path for GraphInfoDir specified in data/Model.py.
We are using graphs from /glade/work/duda/static_moved_to_campaign (in some scenario yamls) and /glade/campaign/mmm/parc/liuz/pandac_common/static_from_duda (default path), however, my guess is that we could use just /glade/campaign/mmm/parc/liuz/pandac_common/static_from_duda.
To Reproduce
Run the ForecastFromGFSAnalysesMPT test scenario.
Some of the scenarios have a GraphInfoDir setting which points to nonexistent directories, causing workflow failures. There are also a number of different directories being pointed to by various scenarios, which we may want to change to be uniform?
As noted by @ibanos90 in PR https://github.com/NCAR/MPAS-Workflow/pull/306
To Reproduce Run the
ForecastFromGFSAnalysesMPT
test scenario.