#2688 closed help (answered)

North/South halos too small for advection error during switch to Monsoon

Reported by: ebubb Owned by: um_support
Component: UM Model Keywords:
Cc: Platform: Monsoon2
UM Version: 11.1

Description

I'm working on suite u-bd494 (which is a copy of u-an983), The original suite was running on a different platform and I'm running it on monsoon. I've made a few small changes to be able to run it on monsoon but am now stuck with this error:

?  Error code: 15
?  Error from routine: LOCATE_HDPS
?  Error message: North/South halos too small for advection.
?        See the following URL for more information:
?        https://code.metoffice.gov.uk/trac/um/wiki/KnownUMFailurePoints
?  Error from processor: 252
?  Error number: 47

[252] exceptions: An non-exception application exit occured.
[252] exceptions: whilst in a serial region
[252] exceptions: Task had pid=58850 on host nid03800
[252] exceptions: Program is "/home/d05/embub/cylc-run/u-bd494/share/fcm_make/build-atmos/bin/um-atmos.exe"
[252] exceptions: calling registered handler @ 0x0047bfa0
Warning in umPrintMgr: umPrintExceptionHandler : Handler Invoked
[252] exceptions: Done callbacks
Rank 252 [Wed Nov 28 09:28:41 2018] [c5-1c2s6n0] application called MPI_Abort(MPI_COMM_WORLD, 9) - process 252

I've had a look at the common fail points web site and have run it with "l_print_max_wind=.true." as suggested. I looked at the windmax.dat file and it only has 2 entries, is this expected or abnormal?

I found a setting "extend_halo_ew" and "extend_halo_ns" and have increased these from 4. I thought that this would make the halo bigger and fix the too small problem. But this didn't work.

Do you have any suggestions as to how I could fix this error?

I think possible while changing the platform to monsoon I have a setting wrong somewhere, as I have not changed any other science settings.

I have reached out to the owner of the original suite and he suggested updating the code version to 11.2 to see if that helps and gave me some instructions.

Change History (3)

comment:1 Changed 10 months ago by grenville

Hi Emma

I don't see this error in the most recent log file (job killed: walltime 312 exceeded limit 300)

Looks like you have fixed the problem?

Grenville

comment:2 Changed 10 months ago by ebubb

Hi Grenville

I contacted the owner of the original suite and he suggested lowering the number of longitudinal nodes. This fixed the issue I opened the ticket for, it can be closed now.

Thank you,
Emma

comment:3 Changed 10 months ago by grenville

  • Resolution set to answered
  • Status changed from new to closed
Note: See TracTickets for help on using tickets.