search

NOAA-EMC / EMC_verif-global

Global Forecast System (GFS) verification package using MET and METplus
8 stars 14 forks source link

Update the location of the verif fix files. #109

Closed DavidHuber-NOAA closed 1 year ago

DavidHuber-NOAA commented 1 year ago

The location of the verif fix files on each HPC has been updated to reflect that the new directory structure and dated control system. Fixes #108

I tested this on Jet and verified file locations on Hera, Orion, and S4. Mallory verified locations on WCOSS2.

malloryprow commented 1 year ago

@KateFriedman-NOAA When the files get linked in sorc/link_workflow.sh, would the verif files be linked as $FIXgfs/fix_verif?

KateFriedman-NOAA commented 1 year ago

@KateFriedman-NOAA When the files get linked in sorc/link_workflow.sh, would the verif files be linked as $FIXgfs/fix_verif?

No, we dropped the "fix_" bit in a reorg last fall. It would end up in the global-workflow clone as fix/verif/....

https://github.com/NOAA-EMC/global-workflow/blob/develop/sorc/link_workflow.sh#L101

We also began using a fix.ver version file to set the timestamp in the symlink so the symlink from global-workflow /fix looks like this:

kate.friedman@dlogin01 fix> ll | grep verif
lrwxrwxrwx 1 kate.friedman global   60 Feb 16 17:01 verif -> /lfs/h2/emc/global/noscrub/emc.global/FIX/fix/verif/20220805

https://github.com/NOAA-EMC/global-workflow/blob/develop/versions/fix.ver

Apologies for not communicating these changes. Not sure if you wanna use something like the fix.ver too. Up to you of course. :)

DavidHuber-NOAA commented 1 year ago

@KateFriedman-NOAA @malloryprow I can fix that as well and give this a cycled test.

malloryprow commented 1 year ago

@KateFriedman-NOAA All good, I was busy with EVS development for the Alpha test. I may have missed it.

@DavidHuber-NOAA We need make sure that https://github.com/NOAA-EMC/EMC_verif-global/blob/69fd27c8cf3bd387b91a00064916d331f3c00969/ush/run_verif_global_in_global_workflow.sh#L232 is correct. Sounds like it should be $FIXgfs/verif.

malloryprow commented 1 year ago

@DavidHuber-NOAA Thanks! Would it be a lot of work to give this a cycled test?

DavidHuber-NOAA commented 1 year ago

@malloryprow No, I don't think so. I will let a 192/96 test run overnight, which should be enough to gauge all three jobs.

DavidHuber-NOAA commented 1 year ago

@malloryprow I ran a cycled test on S4 from 2022113018 through 2022120300 and the verif global jobs all ran successfully. Investigating the .stat files generated, they all had expected lead times with valid entries, so I believe we are good to go.