Skip to content

Use aws ec2 for CI test#422

Closed
MinsukJi-NOAA wants to merge 8 commits into
ufs-community:developfrom
MinsukJi-NOAA:feature/ci-aws
Closed

Use aws ec2 for CI test#422
MinsukJi-NOAA wants to merge 8 commits into
ufs-community:developfrom
MinsukJi-NOAA:feature/ci-aws

Conversation

@MinsukJi-NOAA
Copy link
Copy Markdown
Contributor

Issue(s) addressed

#307

Testing

Changes in this PR are only related to CI (tests/ci directory) and Github Actions (.github/workflows directory), with only minor changes in default_vars.sh, run_test.sh, and utest.

The new CI tests were tested here: https://github.com/MinsukJi-NOAA/ufs-weather-model/actions/runs/579193301

Dependencies

No dependencies.

@MinsukJi-NOAA MinsukJi-NOAA added the Waiting for Reviews The PR is waiting for reviews from associated component PR's. label Feb 24, 2021
Comment thread tests/ci/build_status_check.py Outdated
Copy link
Copy Markdown
Collaborator

@climbfuji climbfuji left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Do you want to leverage the existing S3 bucket at some point? I have been updating the input-data-YYYYMMDD in the bucket for the last several weeks, probably almost two months.

@minsukji
Copy link
Copy Markdown

Do you want to leverage the existing S3 bucket at some point? I have been updating the input-data-YYYYMMDD in the bucket for the last several weeks, probably almost two months.

I think it's definitely something that we need to look into at some point, especially as more test cases are added to CI.

@MinsukJi-NOAA MinsukJi-NOAA deleted the feature/ci-aws branch March 10, 2021 13:30
epic-cicd-jenkins pushed a commit that referenced this pull request Apr 17, 2023
@ligiabernardet pointed out the following differences in FV3 namelist settings between the SRW and MRW apps for the FV3_GFS_v15p2 suite:
1) **effr_in**: is .true. in MRW; is currently .false. in SRW
2) **lgfdlmprad**: is .true. in MRW; is currently .false. in SRW
3) **trans_trac**: is .true. in MRW; is currently .false. in SRW

This PR changes the namelist so that the SRW and MRW use the same settings of these variables for the FV3_GFS_v15p2 suite and performs other namelist fixes.

## DESCRIPTION OF CHANGES: 

**Background note:**
In order to generate the namelist file for the forecast model (input.nml), the workflow scripts start with the base namelist file input.nml.FV3 and then make (1) suite-specific changes specified in the yaml configuration file FV3.input.yml followed by (2) experiment-specific changes specified in the workflow generation script generate_FV3LAM_wflow.sh.

* With the request above that effr_in be set to .true. for the FV3_GFS_v15p2 suite, in FV3.input.yml all suites except FV3_GFS_2017_gfdlmp_regional now set effr_in to .true..  Since effr_in is currently not specified at all in input.nml.FV3 (the base namelist), for FV3_GFS_2017_gfdlmp_regional it gets set to the default value in the weather model, which is .false..  For clarity, remove this variable from all suites in FV3.input.yml except FV3_GFS_2017_gfdlmp_regional and set it to .true. input.nml.FV3.  For FV3_GFS_2017_gfdlmp_regional, set effr_in to .false. in FV3.input.yml to recover the default .false. value it needs to have.
* With the request above that trans_trac be set to .true. for the FV3_GFS_v15p2 suite, in FV3.input.yml trans_trac is (needs to be) set to .true. for two suites (FV3_GFS_v15p2 and FV3_GFS_v16) and not specified for the others.  Since this variable is already set to .true. in input.nml.FV3, for clarity remove it completely from FV3.input.yml.
* In FV3.input.yml, change lgfdlmprad from .false. to .true. for the FV3_GFS_v15p2 suite.
* In FV3.input.yml, remove the variables ccpp_suite, stretch_fac, target_lon, and target_lat from any suites that define them because these are always overwritten (in generate_FV3LAM_wflow.sh) by values specified for the current experiment.
* WE2E tests:
  * Add two new generic suite WE2E tests for the FV3_2017_gfdlmp_regional and FV3_HRRR suites (suite_FV3_GFS_2017_gfdlmp_regional and suite_FV3_HRRR).
  * Rename the regional_001 test to make it the generic test for the FV3_GFS_2017_gfdlmp suite.

## TESTS CONDUCTED:
Ran the following WE2E tests (including the two new ones and the renamed one) on Hera:

* suite_FV3_CPT_v0
* suite_FV3_GFS_2017_gfdlmp
* suite_FV3_GFS_2017_gfdlmp_regional
* suite_FV3_GFS_v15p2
* suite_FV3_GFS_v16
* suite_FV3_GSD_SAR
* suite_FV3_GSD_v0
* suite_FV3_HRRR
* suite_FV3_RRFS_v1alpha
* suite_FV3_RRFS_v1beta
* grid_RRFS_CONUS_25km_ics_FV3GFS_lbcs_FV3GFS_suite_GFS_v16

These cover the 10 currently supported suites plus an extra test for the FV3_GFS_v16 suite (grid_RRFS_CONUS_25km_ics_FV3GFS_lbcs_FV3GFS_suite_GFS_v16; need for this is explained below).  
All except suite_FV3_CPT_v0 and suite_FV3_GFS_v16 were successful.  Both failures are pre-existing ones.  Details:

1) suite_FV3_CPT_v0 failed to do an inconsistency in the namelist (probably a variable is defined in the namelist that is not defined in the weather-model code).  Initial attempts to pinpoint this variable were unsuccessful.  Note that this is a pre-existing failure.
2) suite_FV3_GFS_v16 failed in the run_fcst step after hour 1.  Attempts to replace the namelist with one from a past successful run did not make a difference.  The start of the error message is:
```
forrtl: severe (174): SIGSEGV, segmentation fault occurred
1603 *** Error in `/scratch2/BMC/det/Gerard.Ketefian/UFS_CAM/PR_feature_Ligia_namelist_fix/ufs-srweather-app/bin/ufs_model': corrupted double-linked list: 0x000000000f916c90 ***
*** Error in `/scratch2/BMC/det/Gerard.Ketefian/UFS_CAM/PR_feature_Ligia_namelist_fix/ufs-srweather-app/bin/ufs_model': free(): corrupted unsorted chunks: 0x000000000f916460 ***
```
Note that this is a pre-existing failure.  To obtain a successful test for the FV3_GFS_v16 suite, the grid_RRFS_CONUS_25km_ics_FV3GFS_lbcs_FV3GFS_suite_GFS_v16 test was run, and it was successful.  The differences between these two tests are that suite_FV3_GFS_v16 uses
```
LBC_SPEC_INTVL_HRS="6"
DATE_FIRST_CYCL=DATE_LAST_CYCL="20190520"
EXTRN_MDL_NAME_ICS=EXTRN_MDL_NAME_LBCS="GSMGFS"
```
while grid_RRFS_CONUS_25km_ics_FV3GFS_lbcs_FV3GFS_suite_GFS_v16 uses
```
LBC_SPEC_INTVL_HRS="3"
DATE_FIRST_CYCL=DATE_LAST_CYCL="20190701"
EXTRN_MDL_NAME_ICS=EXTRN_MDL_NAME_LBCS="FV3GFS"
```
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

Waiting for Reviews The PR is waiting for reviews from associated component PR's.

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants