Opened 9 years ago

Closed 9 years ago

#676 closed help (fixed)

Error in reconfiguration with CCN mods

Reported by: bparkes Owned by: um_support
Component: UM Model Keywords:
Cc: Platform:
UM Version: 6.1

Description

I'm running v6.1 of the climate model (coupled atmos-ocean) and have run the model successfully with a user generated ancillary file. When I modify this ancil to investigate the effects of changing CCN in the model it fails during the reconfiguration. Attached is the leave file for the compilation and the normal running of the model with failure in the reconfiguration. The experiment name is xgfog.

Cheers

Ben

Attachments (1)

xgfog000.xgfog.d11230.t115407.leave (132.6 KB) - added by bparkes 9 years ago.
Leave file from model run

Download all attachments as: .zip

Change History (11)

Changed 9 years ago by bparkes

Leave file from model run

comment:1 Changed 9 years ago by bparkes

Note, the compilation leave file is too large but can be provided if necessary.

comment:2 Changed 9 years ago by willie

Hi Ben,

The reconfiguration seems to have worked, but the model has encountered some MPI_Recv errors and there is no output. There are some things you can do to progress,

  1. add /home/n02/n02/wmcginty/modsets/flush.mf77 to the model modsets - this should get some output.
  2. There are some UMUI check set up errors that should be resolved - just by visiting the appropriate page and doing 'close'
  3. repeat the run

Which is the successful run? I assumed it was xgfof, but these have MPI errors too.

Regards,

Willie

comment:3 Changed 9 years ago by bparkes

Hi Willie,

I've added in your modset and corrected the errors in the check setup. The working run is job xfgod. The run is now queuing on hector.

Cheers

Ben

comment:4 Changed 9 years ago by bparkes

Hi Willie,

The run has completed and crashed, however I'm unable to fathom anything from the output. The output leave files are both in my work directory /home/n02/n02/bparkes/work .

Cheers

Ben

comment:5 Changed 9 years ago by willie

Hi Ben,

The model has completed six time steps. In the leave file you have

 Atm_Step: Timestep  7
  RHS zero so GCR( 2 ) not needed 

So the model has become unstable on the seventh. On the diagnostics print page (Scientific sections > section 13) could you change to output every time step instead of every 24, and change the value to 10m/s from 0.4? This will help find when the model goes unstable more precisely. It is possible that it is the ancillary file. The 'good' ancillary from xgfod seems to have vanished so I can't compare them.

Regards,

Willie

comment:6 Changed 9 years ago by bparkes

Hi Willie,

Thanks for the feedback, I've put in the settings you've requested. Also the good ancil file is in /home/n02/n02/bparkes/work/ancil/ and is called ccn_ancil1

Cheers

Ben

comment:7 Changed 9 years ago by bparkes

Hi Willie,

The newest set of leave files are now in my /work directory.

Cheers

Ben

comment:8 Changed 9 years ago by willie

Hi Ben,

The good ancillary file /work/n02/n02/bparkes/ancil/ccn_ancil1 seems to have values between 0 and 400; the current ancillary file /work/n02/n02/bparkes/ancil/ccn_ancil_pr3 has values ranging between zero and 1020. You can check this in xconv. So it looks like you need to re-create your ancillary file.

Regards,

Willie

comment:9 Changed 9 years ago by bparkes

Hi Willie,

I've managed to find and fix the issue with the ancillary files and the model is currently running.

Cheers

Ben

comment:10 Changed 9 years ago by willie

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