search

NOAA-EMC / global-workflow

Global Superstructure/Workflow supporting the Global Forecast System (GFS)
https://global-workflow.readthedocs.io/en/latest
GNU Lesser General Public License v3.0
70 stars 162 forks source link

CICE 100 files on hera are incorrect #2595

Closed NeilBarton-NOAA closed 1 month ago

NeilBarton-NOAA commented 1 month ago

Target fix directory

cice

Type of change

Any workflow changes needed?

Fix version + additional changes

Related issues

On hera, the files at /scratch1/NCEPDEV/global/glopara/fix/cice/20231219/100

are files for MOM6 and not CICE. However, the files for cice_ver=20240416 are correct.

Please update the cice/20231219/100 files or change the cice_ver to 20240416

Pickup location

/scratch1/NCEPDEV/global/glopara/fix/cice/20240416/100

Size

12M

KateFriedman-NOAA commented 1 month ago

Please update the cice/20231219/100 files or change the cice_ver to 20240416

@NeilBarton-NOAA In this situation we should go with the latter option (change to cice_ver=20240416) since we generally avoid changing existing files and instead would incorporate the updated files in a new timestamp...so we don't impact users using the existing files. However, we were waiting to move to the new cice set once a UWM PR went in (see this g-w issue 2483 comment)...

@DeniseWorthen I see that PR was merged...do we need to go to that new hash of UWM when we move to the 20240416 cice fix set or can we move to the newer cice fix set now?

DeniseWorthen commented 1 month ago

@KateFriedman-NOAA I'm not really clear on how all this is coordinated and aligned. But, yes, if you want both the RT and the G-W to use the same CICE fix files then you should update the hash when you use the new fix files. (There were no actual code changes w/ the hash update, just pointing to the new input-data.)

As for the CICE 100 files, I had also created an issue #2480 about a month ago. I think anyone trying to use that set of files will have a failure in the g-w, since it won't actually find CICE files where it is looking for them, right? I don't know how to solve that, other than changing the cice_ver as you suggest.

KateFriedman-NOAA commented 1 month ago

@DeniseWorthen Thanks for your feedback. I discussed this with @WalterKolczynski-NOAA and we will move g-w develop to the newer cice/20240416 timestamp. That will resolve this issue and close #2480 in the process. We'll run CI tests to make sure things are ok. Let us know if you have any objections, thanks!