I tried to move forward by modifying the pre-processing and processing scripts, using zonal and meridional wind (uReconstructZonal and uReconstructMeridional) instead of stream_function and velocity_potential. However, when I reach step 2a_run_hdiag_var.bash, I get the following error:
OOPS_STATS Run start - Runtime: 8.39 sec, Memory: total: 11.99 Gb, per task: min = 42.09 Mb, max = 51.74 Mb
Run: Starting oops::ErrorCovarianceToolbox<MPAS>
==> create geom
==> create geom
==> create geom
==> create geom
==> create geom
==> create geom
===================================================================================
= BAD TERMINATION OF ONE OF YOUR APPLICATION PROCESSES
= PID 1331067 RUNNING AT n18
= EXIT CODE: 9
= CLEANING UP REMAINING PROCESSES
= YOU CAN IGNORE THE BELOW CLEANUP MESSAGES
===================================================================================
[proxy:1@n19] HYD_pmcd_pmip_control_cmd_cb (proxy/pmip_cb.c:487): assert (!closed) failed
[proxy:1@n19] HYDT_dmxu_poll_wait_for_event (lib/tools/demux/demux_poll.c:76): callback returned error status
[proxy:1@n19] main (proxy/pmip.c:182): demux engine error waiting for event
srun: error: n19: task 1: Exited with exit code 7
[mpiexec@n18] HYDT_bscu_wait_for_completion (lib/tools/bootstrap/utils/bscu_wait.c:109): one of the processes terminated badly; aborting
[mpiexec@n18] HYDT_bsci_wait_for_completion (lib/tools/bootstrap/src/bsci_wait.c:21): launcher returned error waiting for completion
[mpiexec@n18] HYD_pmci_wait_for_completion (mpiexec/pmiserv_pmci.c:189): launcher returned error waiting for completion
[mpiexec@n18] main (mpiexec/mpiexec.c:260): process manager error waiting for completion
Would there be any way you could help me by sending the modified scripts to generate the B Matrix for the regional mesh?
I am following the scripts provided in the tutorial (linked above) to generate a static B matrix using the develop version of MPAS-JEDI. However, after running the VBal script (1_run_vbal.bash), the unbalanced samples are not being written.
Before switching to the develop version, I was using release 3.0.0 on an older machine and did not encounter this issue.
After comparing the log files from the two environments, I found the following differences:
Info : -------------------------------------------------------------------
Info : — Write global vertical balance operators
Info : -------------------------------------------------------------------
Info : — Run vertical balance tests driver
Info : -------------------------------------------------------------------
Info : — Copy namelist support radii into C matrix, ensemble 1
Info : +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Info : +++ End of BUMP drivers
Info : +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Info : Write calibration data
Info : Creating central block: ID
Info : Read ensemble member 0
Info : Read ensemble member 1
Info : Read ensemble member 2
Info : Read ensemble member 3
Info : Read ensemble member 4
Info : Write member 0
Info : Read ensemble member 0
Info : Write member 1
Info : Read ensemble member 1
Info : Write member 2
Info : Read ensemble member 2
Info : Write member 3
Info : Read ensemble member 3
Info : Write member 4
Info : Read ensemble member 4
Info : +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Info : +++ Terminate BUMP instance 1 / 1
Info : -------------------------------------------------------------------
Develop version (unbalanced samples not written):
Info : -------------------------------------------------------------------
Info : — Write global vertical balance operators
Info : -------------------------------------------------------------------
Info : — Run vertical balance tests driver
Info : -------------------------------------------------------------------
Info : — Copy namelist support radii into C matrix
Info : +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Info : +++ End of BUMP drivers
Info : +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Info : Write calibration data
Info : Creating central block:
Info : SaberCentralBlock using multivariate strategy: deprecated
Info : Creating single central block.
Info : +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Info : +++ Terminate BUMP instance 1 / 1
Info : -------------------------------------------------------------------
Info : — NetCDF I/O report
Info : Number of opened files: 42
Info : Number of closed files: 42
Info : Registry status: 0
Info : +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Info : +++ All BUMP instances terminated
Info : +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Do you have any idea what might be causing this behavior? Why are the unbalanced samples not being written in the develop version?
While investigating the issue, I also browsed through the YAML files under test/testinput in the GitHub repository and noticed that I could not find any examples mentioning VBal. Has the VBal workflow been deprecated or replaced by another approach in recent versions?