HadGEM-GC3.1-MM workflow

Dear CMS help desk,

I am working on a HadGEM-GC31.-MM simulation (u-c916), which is originally copied from u-aq281 (the one for CMIP6 piControl).
I did some changes to the NEMO source code, etc, and have successfully built it and run it on meto_monsoon. However, there are a few tasks after housekeeping that I do not understand. These are coupled_crun/lrun/nrun. Although these went through OK, but what are they for?

There are also the tasks nemo_rebuild_restart_lrun/crun/nrun after coupled_crun/lrun/nrun. Again, I do not understand why these are here, since I thought you do not need to rebuild the model?
The other problem is that I got errors here as attached below. It is a python3 problem I guess, and I can edit the icb_combrest.py file to resolve this issue. However, is there a way to solve this universally?

Would be great if you could help me around.
Cheers, Alcide

File “/home/d04/alzha/cylc-run/u-ci916/bin/icb_combrest.py”, line 81
print ‘Error: unable to open input file: ’ + pathstart+nn+’.nc’
^
SyntaxError: Missing parentheses in call to ‘print’. Did you mean print(‘Error: unable to open input file: ’ + pathstart+nn+’.nc’)?
[FAIL] rebuild_nemo_restart_files # return-code=1
2021-10-28T10:58:18Z CRITICAL - failed/EXIT

Hi Alcide,

In suite conf → restart tests switch off “Test restartability”, this will switch off all th extra development tests that you don’t need. This will include the coupled_crun/lrun/nrun tasks and the associated nemo_rebuild_* tasks.

Regards,
Ros.

Hi Ross,

Super, that was magic and all sorted.

thank you sososo much!

Cheers, Alcide

Hi Ros,

I am having the another issue.
So the suite u-ci916 was fine and I run it for a few months. I then copied it into u-cj025. Now, u-cj025 halts even at the very initial stage of validate_suite_info with the errors attached below.
This is again a python issue, but strangely it worked previously for u-ci916.
Any suggestions?

Cheers, Alcide

File “/home/d04/alzha/cylc-run/u-cj025/bin/validate_suite_info.py”, line 271, in
main()
File “/home/d04/alzha/cylc-run/u-cj025/bin/validate_suite_info.py”, line 251, in main
warnings, errors = check_experiment(suite_info, cv_experiment_id)
ValueError: need more than 1 value to unpack
[FAIL] python2.7 $CYLC_SUITE_RUN_DIR/bin/validate_suite_info.py $CYLC_SUITE_RUN_DIR <<‘STDIN
[FAIL]
[FAIL] ‘STDIN’ # return-code=1
/usr/lib64/python2.6/site-packages/requests/packages/urllib3/connection.py:337: SubjectAltNameWarning: Certificate for xcslc1 has no subjectAltName, falling back to check for a commonName for now. This feature is being removed by major browsers and deprecated by RFC 2818. (See Feature: Remove support for common names in certificates · Issue #497 · urllib3/urllib3 · GitHub for details.)
SubjectAltNameWarning

Hi Alcide,

validate_suite_info doesn’t work for me in either of the suites. It is not an unusual occurrence, we usually just suggest turning off the task or ignoring it and manually set the task to succeeeded. This task only checks some of the suite metadata with what’s held in the online CMIP6 controlled vocabulary so won’t affect the running of the suite at all.

Regards,
Ros.

Hi Ros,

I believe I have done that by editing suite.rc.
However, there is then an error with fcm_make_um:
Seems there is some variable definitions that are missing?
Cheers, Alcide

[FAIL] /working/d04/alzha/cylc-run/u-cj025/work/18500101T0000Z/fcm_make_um/fcm-make.cfg:9: reference to undefined variable
[FAIL] extract.location{diff}[shumlib] =
[FAIL] undef($shumlib_sources)

Hi Alcide,

Sorry for the delay in getting back to you. Looks like you have managed to solve this problem now is that correct?

Regards,
Ros.

Hi Ross,
no worries at all.
I solved the issue partly by linking the built um model from the u-ci916. However, I still want to know how I can resolve this issue properly if possible.

Cheers, Alcide

Hi Alcide,

Back in u-ci916 at revision 209452 you added in lines to the fcm_make config files to extract shumlib and casim, this is the cause of the problem. Remove these lines and the UM will build ok again.

See u-ci916 changeset diff on MOSRS.

Regards,
Ros.

Thanks Ros for spotting out the issue. However, I do not remember to have changed these files.
Do I need to remove these lines from fcm_make_um/file/fcm-make_main/slash.cfg as well ?

Cheers, Alcide

Hi Alcide,

Yes, remove those 2 lines from the 3 .cfg files.

Cheers,
Ros.