Closed lizziel closed 3 years ago
Interesting. That logging.yaml
looks like the one we are using in GEOS with an extra chunk:
CAP.EXTDATA:
handlers: [mpi_shared]
level: WARNING
root_level: WARNING
propagate: 0
I just tried it with GEOS and it works for me. Now, GEOS is still using pFlogger v1.4.5 because moving to the GFE Namespace (PFLOGGER::pflogger
in CMake) is a bit more involved for us, so you are triggering one of the new return code statuses we don't have.
@lizziel can you try changing the propagate
value to false
:
propagate: false
I'm not the expert @tclune is, but propagate
in pFlogger seems to want a logical, and maybe 0 isn't strictly acceptable as a logical.
That fixed it. Thank you!
I am testing GCHP with library updates, specifically pFlogger v1.5.0 with MAPL 2.6.4, yaFyaml v0.5.0, gFTL-shared v1.2.0, and GMAO_Shared v1.3.8. I am running into an error in pFlogger that causes the program to fail very early in the run. My temporary workaround is to comment out setting
cap_options%logging_config
in main (GCHPctm.F90
). After verifying all results look okay after the library updates I am turning my attention back to this pFlogger issue.The logger yaml file I am using is here. It was originally provided by @tclune for use with MAPL v2.5.0. I am not sure which pFlogger version it went with.
This is the stdout message:
/n/home08/elundgren/testruns/features/update_gmao_libraries/gchp_TransportTracer_gfortran83/CodeDir/src/pFlogger/src/Config.F90 931
This is the stderr message:
Here is a link to the code where the error is caught. It is getting 'propagate' from the config object.
Do you know what the issue is? I suspect it is something in
logging.yaml
but am not yet familiar with configuring pFlogger.