search

NOAA-EMC / RDASApp

Regional DAS
GNU Lesser General Public License v2.1
1 stars 13 forks source link

A problem when running the MPAS-JEDI case with the recent variable name changes #226

Closed HuiLiu-NOAA closed 4 days ago

HuiLiu-NOAA commented 4 days ago

I cloned the latest RDAS with the recent variable name changes. I also updated the geovals.yaml and obsop_name_map.yaml accordingly. When I ran the 3D-enVar case at 2022052700 case and got an error:

0: OOPS_TRACE[0] Variational Assimilation Type=3D-Var 0: OOPS_TRACE[0] CostFactory::create found cost function type 0: OOPS_TRACE[0] Geometry::Geometry starting 0: OOPS_TRACE[0] ========= Geometry::Geometry step 1 ========== 0: ==> create geom 0: Abort(437918234) on node 0 (rank 0 in comm 0): application called MPI_Abort(MPI_COMM_WORLD, 437918234) - process 0

Any idea about this error or anyone have successfully run the case?

The full error log is in /scratch1/NCEPDEV/fv3-cam/Hui.Liu/testmpas/log.ttt

SamuelDegelia-NOAA commented 4 days ago

Hui, I tried copying your run directory and running mpas-jedi with ttt.yaml and found the following error: 

ERROR: No stream matching static exists in call to MPAS_stream_mgr_set_property().
ERROR: ********************************************************************************
ERROR:  Error: Stream 'static' was not defined in the Registry.xml file as
ERROR:         an immutable stream. Immutable streams may only be defined in the Registry.xml
ERROR:         file for a core.
ERROR: ********************************************************************************
CRITICAL ERROR: xml stream parser failed: ./streams.atmosphere_15km
Logging complete.  Closing file at 2024/11/18 16:32:24

It looks like the immutable stream static changed names to invariant in a recent update to MPAS in RDASApp. So your streams.atmosphere_15km needs to be modified for that purpose. After doing that, I got an error in VEGPARM.TBL which also has been updated with a new version of MPAS. Replacing that with the latest version in RDASApp gets me further and I now get an error related to the SfcCorrected obs operator.

I think it might be easier to run your tests using the new MPAS-JEDI case in RDASApp instead of the older case you are trying to run with. You can try running in RDASApp/expr/mpas_2024052700, where all of the input files should already be updated for the latest versions of MPAS and MPAS-JEDI.

HuiLiu-NOAA commented 4 days ago

Thanks @SamuelDegelia-NOAA for the help! I will try the new setup.

HuiLiu-NOAA commented 3 days ago

hi, Sam,

I noticed that the 2nd ombg should be removed from the rrfs_mpasjedi_2024052700_Ens3Dvar.yaml in the RDASApp/expr/mpas_2024052700/

hui

On Mon, Nov 18, 2024 at 9:50 AM Samuel Degelia @.***> wrote:

Hui, I tried copying your run directory and running mpas-jedi with ttt.yaml and found the following error:

ERROR: No stream matching static exists in call to MPAS_stream_mgr_set_property(). ERROR: **** ERROR: Error: Stream 'static' was not defined in the Registry.xml file as ERROR: an immutable stream. Immutable streams may only be defined in the Registry.xml ERROR: file for a core. ERROR: **** CRITICAL ERROR: xml stream parser failed: ./streams.atmosphere_15km Logging complete. Closing file at 2024/11/18 16:32:24

It looks like the immutable stream static changed names to invariant in a recent update to MPAS in RDASApp. So your streams.atmosphere_15km needs to be modified for that purpose. After doing that, I got an error in VEGPARM.TBL which also has been updated with a new version of MPAS. Replacing that with the latest version in RDASApp gets me further and I now get an error related to the SfcCorrected obs operator.

I think it might be easier to run your tests using the new MPAS-JEDI case in RDASApp instead of the older case you are trying to run with. You can try running in RDASApp/expr/mpas_2024052700, where all of the input files should already be updated for the latest versions of MPAS and MPAS-JEDI.

— Reply to this email directly, view it on GitHub https://github.com/NOAA-EMC/RDASApp/issues/226#issuecomment-2483582545, or unsubscribe https://github.com/notifications/unsubscribe-auth/AOF252C4HX3QTDCZOKTNERD2BILF7AVCNFSM6AAAAABR77AMHGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDIOBTGU4DENJUGU . You are receiving this because you authored the thread.Message ID: @.***>

SamuelDegelia-NOAA commented 3 days ago

@HuiLiu-NOAA Thank you for catching this. I just tested it and it luckily it looks like the extra ombg does not change the analysis or the o-b data output to the hofx files. But it should still be removed. I'll add a quick hotfix to remove this.

guoqing-noaa commented 3 days ago

@HuiLiu-NOAA Thank you for catching this. I just tested it and it luckily it looks like the extra ombg does not change the analysis or the o-b data output to the hofx files. But it should still be removed. I'll add a quick hotfix to remove this.

Yes, the YAML function will just overwrite the first ombg diagnostic dictionary with the second specification internally.