[develop] Fix issue 555 (restore default behavior of EXPT_BASEDIR option) #562
Conversation
of old annoying debug print statement.
mkavulich
added
the
ci-jet-intel-WE
mkavulich
requested review from
gsketefian,
JeffBeck-NOAA,
RatkoVasic-NOAA,
BenjaminBlake-NOAA,
ywangwof,
chan-hoo,
panll,
christinaholtNOAA,
christopherwharrop-noaa,
danielabdi-noaa,
mark-a-potts,
jkbk2004,
willmayfield,
dmwright526,
gspetro-NOAA,
natalie-perlin,
EdwardSnyder-NOAA and
MichaelLueken
as code owners
venitahagerty
removed
the
ci-jet-intel-WE
|
Machine: jet |
mkavulich
changed the title
MichaelLueken
added
the
run_we2e_coverage_tests
There was a problem hiding this comment.
@mkavulich All of the WE2E tests are failing in Jenkins. Attempting to manually run the WE2E tests on Jet was encountering issues as well. I'm seeing:
GLOBAL_VAR_DEFNS_FP = 'var_defns.sh'
rather than:
GLOBAL_VAR_DEFNS_FP = '/mnt/lfs4/HFIP/hfv3gfs/Michael.Lueken/exp_dirs/grid_RRFS_CONUS_13km_ics_FV3GFS_lbcs_FV3GFS_suite_GFS_v15p2/var_defns.sh'
while generating the experiment.
The following files are present in ush/, but not in the expt_dirs - ush/data_table, ush/fd_nems.yaml, ush/field_table, ush/fix_am, ush/input.nml, ush/nems.configure, ush/suite_FV3_GFS_v15p2.xml, and ush/var_defns.sh. Thus, I'm only seeing FV3LAM_wflow.xml, config.yaml, crontab.bak.2023-01-27_18:46:39, launch_FV3LAM_wflow.sh, log.generate_FV3LAM_wflow, and log.launch_FV3LAM_wflow in the experiment directory.
Can you think of what might be happening to keep var_defns.sh from being brought in?
|
@MichaelLueken I'm not sure what happened here, I must have run my tests on the incorrect code, because I am also seeing these same failures now. Clearly the code doesn't work as-is, I'll need to investigate what's going on. |
…mental tests now *actually* pass.
mkavulich
added
ci-hera-intel-WE
venitahagerty
removed
ci-jet-intel-WE
|
Machine: hera
|
|
Machine: jet |
|
@MichaelLueken I have applied the correct fix for this issue. I discussed with @venitahagerty offline and it looks like the Hera test is failing due to running out of disk space in the test location; I ran the Hera fundamental tests successfully on my own. |
|
@mkavulich resubmission of the Jenkins tests showed that they all passed, with the exception of Orion. The |
There was a problem hiding this comment.
@mkavulich Following the modification you pushed on Friday (and a resubmission of the Orion tests this morning), all Jenkins tests are now passing successfully. Reapproving these changes.
DESCRIPTION OF CHANGES:
This PR restores the previous behavior of the variable
EXPT_BASEDIR(Item 2 is the behavior that was previously broken), which has the following effect on the experiment directoryEXPTDIR:EXPT_BASEDIRis not set or set to a null value, the default value (${HOMEdir}/../expt_dirs) will be usedEXPT_BASEDIRis set to a relative path (i.e. the first character is not/), the user-specified path will be appended to the default value${HOMEdir}/../expt_dirs(for example if the user specifiesEXPT_BASEDIR=some/relative/pathin their config.yaml, it will be updated toEXPT_BASEDIR=${HOMEdir}/../expt_dirs/some/relative/pathin the workflowEXPT_BASEDIRis set to an absolute path, that path will be used as enteredAfter the above logic is applied,
EXPTDIRwill be created by joining the pathsEXPT_BASEDIRandEXPT_SUBDIRas usual.Type of change
TESTS CONDUCTED:
Ran fundamental test suite on Hera, confirmed that the old, expected behavior for the expt_basedir argument has been restored. Ran python unit tests, only UFS_plot_domains.py failed which is a known pre-existing error.
DEPENDENCIES:
None
DOCUMENTATION:
While this behavior had been documented in comments and usage doc string for
run_WE2E_tests.sh, it had not been included inconfigure_defaults.yamlor the Users Guide; I updated the documentation in those locations.ISSUE:
Fixes issue mentioned in #555
CHECKLIST