Automatic UM to netcdf conversion

General
, ,
1

We manually ran:

esm1p5_convert_nc for missing years so its all good. Just FYI incase it is of interest why these years were missed. Missing years:

/scratch/p66/jxs599/access-esm/archive/PI-case2b-expt-7ba86be5
output002/atmosphere/
output003/atmosphere/
output004/atmosphere/
output005/atmosphere/
output006/atmosphere/
output007/atmosphere/


/scratch/p66/jxs599/access-esm/archive/PI-case2a-expt-11659a9b/
output000/atmosphere/
2

It isn’t clear to me what happened that you didn’t expect? And it seems you solved the problem yourself, is that correct?

Do you still need assistance for this @Jhan?

3

HNY @Aidan,

It is resolved on my end. Just last time it happened Spencer went looking for the underlying cause (if there was one). Just thought you might want an extra example to check the logs if you want.

1 Like
4

Assigned to Spencer so he can take a look when he is back from leave. Does not require resolution as such.

5

Hi @Jhan,

I’ve taken a look at the output logs. In each of the failing years, the collation job ran out of walltime when combining the MOM output files. The UM netCDF conversion is done after this step, and so never ran for these years.

For most years, the collation appears to have finished in ~20 minutes, but for some reason it didn’t complete within an hour for the failing years. If this problem keeps coming up, it might be worth increasing the collation walltime in the config.yaml:

collate:
    exe: mppnccombine.spack
    restart: true
    mem: 4GB
    walltime: 1:00:00
    mpi: false
1 Like