UFS-dev PR#69 #107
UFS-dev PR#69 #107
Conversation
…3-jun-27 (ufs-community#1811) * WW3: updates
* Add new BL_DATE Update inline post with: *Clean up legacy code in inline post interface *Relocate computation of diagnostic variables *Sync changes from offline post interface *Update upp revision
* Update empty_output_bundle * Update pull_request_template.md from 1813 Co-authored-by: Brian Curtis <Brian.Curtis@noaa.gov> Co-authored-by: Dusan Jovic <Dusan.Jovic@noaa.gov>
* Update control_CubedSphereGrid_parallel test * The control_CubedSphereGrid_parallel test has been updated to run post and create native grid history files
* reverting the bug fix for ktherm=2 allows all cpld tests to pass and the single datm test using ktherm=2 (datm_cdeps_gfs) to also pass all other datm tests which use ktherm=1 fail * update CICE * change freq in for ice_diag.d global values * remove unused history* settings * update DISKNM with epic and rt log * Update bl.py for new Hera blstore
* add gridtype to gocart CAP.rc * move to GOCART 20230227 version with threading capability * udpate to netcdf/4.9.2 and add threading capability for gocart * update esmf library * update GOCART to allow no-Nitrates run
…r CMEPS (ufs-community#1807) * commit changes for sensible heat flux * add sensible heat changes * update submodules and fd for evap fields * backup to pre-evap changes to test
…s-community#1864) * modify mom6_files.cmake to reflect files being removed/added
* update FV3 submodule pointer after GT combined his physics PRs
|
Expected RT failures for #107: |
|
Automated RT Failure Notification |
|
@mkavulich I've gone through the list of reported failures and have the following list of failures that were not expected for the parent ufs/dev PR (ufs-community#1734): I'm going to go through this list and see which other included ufs-wm PRs could cause these failures. Assuming that the failures can be explained, I'll go ahead and add the BL creation tags. |
|
I've whittled the unexplained failures down. This is now the list with the UFS-WM PR that causes the failure listed afterward: I'm hypothesizing that the remaining unexplained test failures are also the result of 1794. |
|
@mkavulich OK, I think that I've solved the mystery of the final failing tests. The RT log for 1794 was only attached for WCOSS. All remaining failed tests are not conducted on WCOSS for the code as of this date. Since the tests that fail in 1794 on WCOSS are all similar to the remaining unexplained failed tests in this PR, I feel confident that the tests are supposed to fail for 1794. I'll start new BLs. |
|
@mkavulich Do you know if the scripts are active on Hera to listen for the added labels? |
|
Automated RT Failure Notification |
|
@grantfirl It looks like they are running, the cron job only runs every 5 minutes so it can take up to that long to start a script, and only one compiler is processed at a time so it can take a while before the second label is processed. It looks like the regtests haven't been updated with the new way of specifying compiler, which is why the GNU test is failing. I'll look into getting that working. |
|
Automated RT Failure Notification |
|
Okay... I've figured out the issue. Since there is no longer any way to specify one compiler or another, we will have to adjust the labels to just be platform-specific rather than platform and compiler specific. Re-adding the gnu label just now caused the original tests to fail due to the run directory being deleted, which is also not good. I need to think about the best way to prevent that going forward. |
|
So, should we manually do RTs for this or are you going to try to fix the RT scripts within the context of this PR? |
|
Automated RT Failure Notification |
|
@mkavulich OK, so the tests ran and the log is showing the BL creation as successful (See /scratch1/BMC/gmtb/CCPP_regression_testing/NCAR_ufs-weather-model/run/1584888755/20240110195011/ufs-weather-model/tests/logs/RegressionTests_hera.log). It looks like the new baselines are stored in /scratch1/BMC/gmtb/CCPP_regression_testing/NCAR_ufs-weather-model/new. So, we're supposed to manually move this to /scratch1/BMC/gmtb/CCPP_regression_testing/NCAR_ufs-weather-model/baselines/main-20240110 and update bl_date.ncar.conf before merging, correct? |
|
@mkavulich I don't have permissions to move the new baselines into the right place. Could you change permissions on the directories so that I can do this, please? Otherwise, I'm ready to start the merge process. |
There was a problem hiding this comment.
Okay the baseline has been staged (also just realized I hadn't approved until now, sorry about that). I also added a umask change that will hopefully give group write permissions for files written by the regtest; I thought I had done this already but I guess not.
|
What about permissions in /scratch1/BMC/gmtb/CCPP_regression_testing/NCAR_ufs-weather-model/baselines? For example, I still can't remove old baselines. We're supposed to keep the last 2 to save space. |
|
The umask command was added as a line in the crontab entry so it should allow group read/write permissions for all files created by the regression tests going forward. I have recursively updated permissions in baselines/main-20240110 to have group read/write just now. FYI it looks like you created baselines/main-20231031, so you'll have to update the permissions or delete that one when the time comes. |
Identical to ufs-community#1734
Includes changes from:
ufs-community#1811 (no BL changes)
ufs-community#1794 (BL changes expected): The UPP output grib2 files for global and regional FV3, such as datasets GFSPRS, NATLEV, PRSLEV due to upp upgrade.
ufs-community#1825 (no BL changes)
ufs-community#1831 (BL changes expected): control_CubedSphereGrid_parallel
ufs-community#1842 (no BL changes)
ufs-community#1745 (BL changes expected): All tests with GOCART will have gocart output files changed.
All tests will have attributes data type change in the atm history files.
ufs-community#1807 (BL changes expected): Cpld and hafs test change baselines because sign for the sensible, evaporative, latent and momentum within the mediator files are different.
ufs-community#1864 (BL changes expected): all coupled tests include DATM