I am trying to restart HadGEM3-GC3.1-LL from the end of one of the CMIP6 historical runs, u-bg467 (2015-01-01). I am using suite u-cn716, which is similar to PI suite u-as037 (but with restart dumps/GHG/ozone etc the same as u-bg467 in 2014).
u-cn716 keeps failing at the first coupled step (20150101T000Z) and I don’t know what the problem is:
UM: files like cylc-run/u-cn716/log/job/20150101T0000Z/coupled/01/job.err
only have a standard MPICH error.
CICE: diagnostic file cylc-run/u-cn716/work/20150101T0000Z/coupled/ice_diag.d is empty
NEMO: The last lines of cylc-run/u-cn716/work/20150101T0000Z/coupled/ocean.output has a warning that also shows up in suites that run fine (e.g. u-ck767: cylc-run/u-ck767/work/18700401T0000Z/coupled/ocean.output):
===>>> : W A R N I N G
===============
W A R N I N G: end of record or file while reading namelist nambias in configuration namelist iostat = ****
I am now using suite u-cn949. The changes I have made to this suite from PI suites that run (e.g. u-as037, u-ck767) are:
(1) The CFC and GHG concentrations (e.g co2_mmr)
(2) Some ancillary files (e.g. ukca_em_files)
(3) The NEMO, CICE and UM restart files.
u-cn949 runs fine with changes (1) and (2) only, but as soon as I add (3) the model fails in the first coupled step, with the same initialisation problem as the previous post. I also ran u-cn949 with fcm_make_um > u-cn949/app/fcm_make_um/rose-app.conf:optimisation_level=debug, but this didn’t give any more information.
Is there a way to get more information about why the model is failing?
Try setting PRINT_STATUS to Extra Diagnostic messages (re run) - then look in the file in /home/n02/n02/radiam24/cylc-run/u-cn949/work/20150101T0000Z/coupled/pe_output