Modification required to run a Monsoon suite on ARCHER2

,
21

Hi Ros,

I found the following information in the job.out file (near in the end of the job.out) in directory
/home/akpandeyjnu/cylc-run/u-cr175/log/job/20080901T0000Z/atmos_main/01

Error in EM_FOPEN: Permission denied - //work/y07/y07/umshared/ancil/atmos/n96e/ukca_emiss/biogenic/v1/DMS_land_spiro1992.nc

Is this is permission issues to read the file? Where I can ‘pe_output’ files?

Regards, Alok

22

Hi Ros,
I found I don’t have permission to read files in the directory ‘/work/y07/ y07’ on ARCHER2. I have permission to read ‘/work/y07/shared’.
Regards, Alok

23

Hi Alok,

There are no UM files under /work/y07/y07. The path is incorrect - see above.

Regards
Ros

24

Hi Ros,
Thanks for pointing out the path error. I have managed to run the model but it is struggling in the postproc (suite u-cr314). The job.err file has the following information:

Lmod is automatically replacing “cce/11.0.4” with “gcc/10.2.0”.
Due to MODULEPATH changes, the following have been reloaded:

  1. cray-mpich/8.1.4
    [WARN] file:atmospp.nl: skip missing optional source: namelist:moose_arch
    [WARN] file:nemocicepp.nl: skip missing optional source: namelist:moose_arch
    [WARN] Mule Module is not available. um-pumf will be used.
    [WARN] Iris Module is not available
    [FAIL] This code is compatible with versions of um-pumf from vn9.1 until its retirement at vn10.9. Currently attempting to use version 11.0
    Please ensure Mule is available as an alternative utiltity.
    [FAIL] Terminating PostProc…
    [FAIL] main_pp.py atmos <<‘STDIN’
    [FAIL]
    [FAIL] ‘__STDIN’ # return-code=1
    2022-10-10T18:22:20Z CRITICAL - failed/EXIT

I tried to use postproc_2.3 but that is creating issues in fcm_make2pp (suite u-cr175). Can you please point me to fix this postproc issue?

Additionally, I tried to do ‘fcm status’, ‘fcm diff’ and ‘fcm commit’ on pumatest and after ‘fcm commit’. I am unable to commit archer2.rc file which has been added.
-bash-4.1$ fcm status
? site/archer2.rc

What is the ‘?’ stands for? Is it due to new file addition? How to commit the suite including addition archer2.rc file?

Regards, Alok

25

Hi Alok,

Postproc_2.2 is not ARCHER2 compatible. You will need to upgrade the fcm_make_pp and postproc apps to postproc_2.3 using a similar process to what you did for the UM. (i.e. rose app-upgrade -a postproc_2.3)

Yes the ? is due to a new file that is not under revision control. You need to fcm add site/archer2.rc before you commit again.

Regards,
Ros.

26

Hi Ros,

Thanks for the clarification. I have upgraded the fcm_make_pp and postproc app of the suite ‘u-cr175’ but I got an error in fcm_make2pp. I thought I have done some irreversible error so created a new suite and modified it to run on ARCHER2 (u-cr314).

I have upgraded the fcm_make_pp and postproc app of suite ‘u-cr314’ and it also fails in fcm_make2_pp. The job.err message has the following information

The following have been reloaded with a version change:

  1. cce/11.0.4 => cce/12.0.3
    Use of uninitialized value in concatenation (.) or string at /mnt/lustre/a2fs-work1/work/y07/shared/umshared/software/fcm-2019.09.0/bin/…/lib/FCM/Util.pm line 281.
    [FAIL] atmos_transform.py: don’t know how to build specified target
    [FAIL] climatemean.py: don’t know how to build specified target
    [FAIL] fcm make -C /work/n02/n02/alok/cylc-run/u-cr314/share/fcm_make_pp -n 2 -j 128 # return-code=255
    2022-10-11T14:05:16Z CRITICAL - failed/EXIT

How can I fix this error?

Regards, Alok

27

Hi Alok,

You cannot use source code branches of a different version to the base code. This applies to all fcm_make apps. Remove the postproc_2.2 branch.

Whilst you’re there please add in the branch:

branches/dev/rosalynhatcher/postproc_2.3_pptransfer_gridftp_nopw@4557

which you’ll need if you’re going to be transferring data from ARCHER2 to JASMIN at somepoint.

Regards,
Ros.