Mediator history write issue

[yunfangsun]

Hi @uturuncoglu ,

When I tried to run the atm+sch+ww3 case, and with the run sequence including med_phases_history_write, the job can't start.

With your suggestions by modifying the PIO in the mediator configuration:

MED_model:                      cmeps
MED_petlist_bounds:             0 99
MED_omp_num_threads:            1
MED_attributes::
  ATM_model = datm
  OCN_model = schism
  WAV_model = ww3
  history_n = 1
  history_option = nsteps
  history_ymd = -999
  coupling_mode = coastal
  pio_rearranger = subset
  pio_typename = pnetcdf
  pio_stride = 4
::

This case still can not start, and if I remove med_phases_history_write, the case could successfully run, could you please help me to check the error? This example case and its log files are located at /work2/noaa/nos-surge/yunfangs/duck/duck_From_Dan_03312025/nncouple_run_final_BT1_vor_test2_1 on Hercules.

Thank you!

[anntsay]

• @mansurjisan for visibility

[anntsay]

May 12 meeting notes:
Dan: also has similar situation: the only way to check seems to be post processing from the model.. not from inside the CMEPS. if has "med_phases_restart_write" then it would crash... therefore, Ufuk suggest to comment out everything that is related to mediator output.
Mansur: suspect there are changes in UFS.

[anntsay]

workaround: turn off mediator output for now.

Next Action: wait for Ufuk to return and see if how we could work with UFS team to resolve this issue. (Ann to add Mansur to UFS biweekly meeting to stand in for Jana)