GEN_ENS_PROD failed to produce output files for RRFS ensemble

[BinbinZhou-NOAA]

I tried to run GEN_ENS_PROD tool in METplus to produce RRFS snowfall ensemble. But it failed.
There is no any ERROR or WARNING message in the log file.
The GEN_ENS_PROD_OUTPUT_DIR defined in the conf file is also not generated
But similar run for HREF is working.
A completed simple testing case (only for one forecast hour) can be found in the directory
/lfs/h2/emc/vpppg/noscrub/binbin.zhou/met/RRFS on WCOSS2/Cactus.
The testing scripts are run_href_snow6h.sh and run_rrfs_snow6h.sh for href and rrfs, respectively

Thanks!

Binbin

NCEP/EMC

[JohnHalleyGotway]

@BinbinZhou-NOAA, thanks to @georgemccabe for hopping on a call and debugging this with me on Hera. You should make 2 important changes to your GenEnsProd_rrfs.conf file.

Comment out, delete entirely, or actually define the following line:

# LOG_LINE_FORMAT =

This is causing all the blank lines you're probably seeing in your METplus log output file. Setting the format as blank causes empty lines to be written. Commenting it out enables the warning/error messages to actually be written using the default log line formatting. And with them being printed, you'll be able to see the following error message:

11/20 21:20:33.277 metplus (command_builder.py:257) ERROR: (command_builder.py:862) Found more files than expected! Found 14 expected 10. Adjust wildcard expression in template or adjust number of expected members (N_MEMBERS). 

The RRFS configuration tells METplus to expect 10 members, but it finds 14! Just update the config to set it to 14 instead:

GEN_ENS_PROD_N_MEMBERS = 14

After making those 2 changes, I was able to get a successful run.

@georgemccabe, based on this, we could consider enhancing METplus to do some validation of the LOG_LINE_FORMAT and bail out if its not set in a sensible way.

Hopefully, this helps get you rolling @BinbinZhou-NOAA.

[BinbinZhou-NOAA]

John, Yes, after the 2 modifications, it works now. Thanks! Binbin