gmao-cda/cycle-oletkf

controlling scripts for doing cycling DA with GEOS

cycle-oletkf's Issues

remove cylc dir-dependent path in flow

replace os.rename with shutil.move

https://www.codespeedy.com/difference-between-os-rename-and-shutil-move-in-python/

move recenter log to log dir

nx,ny calculation seems redundant

run_letkf.py

separate discover_env

remove module load section from flow.cylc

set `found_binary` = True/False

get_exec_rst

`num_oserver_nodes` needs to be user assigned

combine `Over-Specified CPU resources`

set_exp_run_params

test 0d25 flow

replace c-shell create_history_collection

additional output from HISTORY.rc to letkf_exp_dir

After each member generates 2d, 3d (e.g., 2d.0001, ...., 2d.000N)
we only save 2d.mean = avg(2d.X) for each cycle instead of saving every member

save both mean sprd

2d: hourly output
3d: decide later

delete step reconcile_recenter_mem.py

This step is needed because the current prep_geos_fcst gets all atmosphere rstfiles from bgkd dir. Later, we need to change the prep_geos_fcst to let it get atmosphere rstfiles from anal

run_fcst.py

• use gcm_setup run using gcm_run_lib.py at 5 deg
• use clone_exp.csh to clone the expdir and run
• use prepared can to run fcst under projects
• use prepared can to run fcst under arbitury location
• create the prep_fcst.py to automize this part
  What are really needed are prep_fcst.py that prepare the dir to launch the fcst

parallize Copy Restarts to Scratch Directory

Copy Restarts to Scratch Directory

use sequential copy in py, but should parallize here

write a function to generate `linkbcs` directly

0.25deg restarts tarball has restarts with no suffix ".nc4"

cause failure of prep_restart_files_GEOSgcm.py

create flow.cylc

under cpld_odas

Test Openwater Restart for Number of tiles correctness

mpirun is not allowed at the login nodes

gcm_run_lib gives no error even if mpirun does not start

4-member 0.25 3D LETKF test case

@sanAkel

1. copy the whole directory /discover/nobackup/projects/gmao/scda_iesa/shared/0d25_4mem_oletkf
2. Inside this directory, modify the L6-15 in the script test_LETKF_run.bsh
3. on a cas or sky node, run bash test_LETKF_run.bsh

Let me know if it works. Currently I cannot get a even 10-minute slot to test it

progress

This issue is for monitoring my progress of re-writing gcm_run.j in python

• set_env_vars
• set_exp_vars
• create_exp_subdirs
• set_exp_run_params
• prepare_fixed_files
• create history_collection
• link_bcs
• get_exec_rst
• Multiple iteration
• Iter: split water
• Iter: Rename checkpoints
• Iter: Move HISTORY files to holding dir
• Iter: Run PPS
• Iter: Update Iteration Counter
• Re-Submit job

create_init_ensemble_from_restarts.py

run_obsop

Update cylc option

1. How does cylc determine if a slurm job run succesfully or not?
2. Add retry option for different jobs

enable post to generate files required by HISTORY.rc

prep_fcst geos_gcm.x not included

let `use_ioserver` be logical

three restart files missing

rst, type= catchcn_internal_rst cannot open catchcn_internal_rst' (No such file or directory) rst, type= catchcn_internal_rst cannot open catchcn_internal_rst' (No such file or directory)
rst, type= gocartdata_internal_rst cannot open `gocartdata_internal_rst' (No such file or directory)

Split exp_template out as a submodule

replace grep with re.search

test new env