Error outputting to non existent CICE file

ANCOMS Regional MOM
1

This is splitting a discussion off from this thread
When running in mom-only mode, @Paul.Gregory and a few other’s are getting this error:

replys to this are:

@Aidan @Paul.Gregory @anton @dougiesquire @mmr0

2

This was also a reply to this issue:

1 Like
3

Ok I got this working.

I was able to restart the run by purging my access-om3/archive directory access-rom3-MR.

This may be related to @mmr0 's restart issues.

Doing this in config.yaml

#userscripts:
    # setup: /usr/bin/bash /g/data/vk83/apps/om3-scripts/payu_config/setup.sh
    # archive: /usr/bin/bash /g/data/vk83/apps/om3-scripts/payu_config/archive.sh

Works. The job completes with stdout

======================================================================================
                  Resource Usage on 2025-03-11 14:20:42:
   Job Id:             136779491.gadi-pbs
   Project:            gb02
   Exit Status:        0
   Service Units:      8.17
   NCPUs Requested:    16                     NCPUs Used: 16              
                                           CPU Time Used: 02:32:58        
   Memory Requested:   100.0GB               Memory Used: 74.37GB         
   Walltime requested: 01:00:00            Walltime Used: 00:09:48        
   JobFS requested:    10.0GB                 JobFS used: 8.07MB          
======================================================================================

with stderr

Loading access-om3/pr30-5
  Loading requirement: access-om3-nuopc/0.3.1-fvr7qaw
Currently Loaded Modulefiles:
 1) access-om3-nuopc/0.3.1-fvr7qaw   3) nco/5.0.5                5) pbs  
 2) access-om3/pr30-5                4) openmpi/4.1.7(default)

It produces the following outputs in
access-rom3-MR/output000

 $ ls -lht *.nc
42M Mar 11 14:20 access-om3.mom6.h.native_2013_01.nc
612K Mar 11 14:20 access-om3.mom6.h.sfc_2013_01.nc
130K Mar 11 14:20 access-om3.mom6.h.static.nc
20M Mar 11 14:20 access-om3.mom6.h.z_2013_01.nc

Note - if you don’t do what @dougiesquire suggests and didn’t ENTIRELY remove the userscript section in config.yaml, i.e. this

userscripts:
    # setup: /usr/bin/bash /g/data/vk83/apps/om3-scripts/payu_config/setup.sh
    # archive: /usr/bin/bash /g/data/vk83/apps/om3-scripts/payu_config/archive.sh

It will generate a payu error:

Traceback (most recent call last):
  File "/g/data/vk83/prerelease/apps/base_conda/envs/payu-dev-20250220T210827Z-39e4b9b/bin/payu-run", line 8, in <module>
    sys.exit(runscript())
  File "/g/data/vk83/prerelease/apps/base_conda/envs/payu-dev-20250220T210827Z-39e4b9b/lib/python3.10/site-packages/payu/subcommands/ru
n_cmd.py", line 123, in runscript
    expt = Experiment(lab, reproduce=run_args.reproduce, force=run_args.force)
  File "/g/data/vk83/prerelease/apps/base_conda/envs/payu-dev-20250220T210827Z-39e4b9b/lib/python3.10/site-packages/payu/experiment.py"
, line 124, in __init__
    init_script = self.userscripts.get('init')
AttributeError: 'NoneType' object has no attribute 'get'

Evidently if you place userscripts in the config.yaml, payu expects a value.

I’ll now continue working with this configuration, changing the layouts, number of CPUs and resolution to try and figure out why I couldn’t get the 10x10 layout with 100 CPUs at 0.1 resolution working.

4

10x10 layout with 100 CPUs at 0.1 resolution working

Just noting that in general its a good idea to use entire nodes, e.g. multiples of 48 cores if you’re using the normal queue

1 Like
5

Thanks @dougiesquire - you have answered a question I was about to ask! I need to update this in the instructions.

6

Is this issue really fatal ?

It should just be a warning which could be ignored .

7

The CICE warning is not fatal and can be ignored. It does report itself as being fatal though, so is causing some confusion.

1 Like