Opened 10 years ago

Closed 9 years ago

#590 closed help (fixed)

vn7.6 on hector

Reported by: anmcr Owned by: willie
Component: UM Model Keywords: vn7.6, LAM
Cc: ros Platform:
UM Version: 7.6

Description

Hi,

I copied over a vn7.6 4km run to HECToR which Mark Weeks of the MetOffice? has been running on MONSooN — specifically to take advantage of the improvements it contains for running over steep orography. I saw that there are vn7.6 datasets on HECToR and some vn7.6 jobs on PUMA. Is this version therefore supported?

My job id is xffwf. The model fails in the model build. Not surprisingly, as I have not included any fixes in 'include modifications from branches'. I am also unsure whether the FCM is setup properly, as I copied the format from my prvious vn7.1 job. The output is in /home/n02/n02/anmcr/output.

Thanks,

Andrew

Change History (9)

comment:1 Changed 10 years ago by willie

  • Owner changed from um_support to willie
  • Status changed from new to assigned

Hi Andrew,

I am currently working on getting the Met Office Parallel Suite 25, which includes a vn7.6 global job, working on HECToR. So far my global build job xfupa is working and I am moving on to the reconfiguration job. I will keep you posted.

Regards,

Willie

comment:2 Changed 10 years ago by ros

  • Cc ros added

Hi Andrew,

If you want to try running your job copied from Mark then where FCM is concerned you will need to change the following things:

In UMUI window FCM Configuration → FCM Options for UM Model and Reconfiguration you need to

  1. Modify Container File name and location (UM_CONTAINER) to be $UM_SVN_BIND/container.cfg@vn7.6_cfg
  1. Switch on Include modifications from branches and specify the NCAS branch fcm:um_br/pkg/Config/VN7.6_ncas/src and put a Y in Use Y/N column. In FCM UM Versions you always need to include the appropriate VNx.y_ncas branch to run on HECToR.

To ensure that you get a clean build and extract, rather than an incremental one, having changed some FCM options, I'd also advise you go into window FCM Configuration → FCM Extract directories and Output levels and switch on Force Full Extract and Force Full Build.

For reference all the UM Version that we support on HECToR and MONSooN are listed on the UM Trac Wiki

Regards,
Ros.

comment:3 Changed 10 years ago by anmcr

Dear Ros and Willie,

Thank you for suggestions.

I made the changes suggested by you, Ros. And I made a number of further changes so that when I 'differenced' my job against xfupa, the differences were number of model levels, physics set up, etc, and the options required to build/compile the model matched.

However, the building fails with the error

'mkdir: cannot create directory '/xffwf': Read-only file system'

This output file is the most recent in /home/n02/n02/anmcr/output/.

I would be very grateful if you could have a look.

Best wishes,

Andrew

comment:4 Changed 10 years ago by ros

Hi Andrew,

It looks like, for some reason, it's been unable to expand $DATADIR to be your work directory. I can investigate, but won't be able to do so until Thursday, as I'm in Exeter tomorrow. In the meantime you could try explicitly setting $DATAW and $DATAM to /work/n02/n02/anmcr/$RUNID in UMUI panel:
Sub-independent Output and resources → Time convention and script environment variables
and resubmitting to compile.

Another way around this, and that will be quicker, as the executable has compiled ok, it just hasn't been copied to the $RUNID/bin directory, is to copy the bin directory under umbase into the top level directory (/work/n02/n02/anmcr/xffwf), and also copy qxum.m4 from ummodel/bin into the top level bin. Then you should be able to submit the job to run.

Regards,
Ros.

comment:5 Changed 10 years ago by anmcr

Hi Ros,

I done as you suggested and copied over the executable file. However, I got the error along the lines of 'number of fields not correct' or something, which meant that the startdump needed to be reconfigured.

But on reconfiguring I get the error in the .rcf file that

'Ancillary file does not exist.
File : /work/n02/n02/anmcr/ofcap/ancil/qrclim.ozone_L70_035'

However, this file does exist. So I don't really know what is going on.

Could you please help again?

Thanks,

Andrew

comment:6 Changed 10 years ago by ros

Hi Andrew,

Well that confused me for a while….. Sometimes I hate 0s and Os.

The filename in your /work directory is qrclim.ozone_L70_O35 (letter O before the 35 not the number zero which your job is looking for)

Regards,
Ros.

comment:7 Changed 10 years ago by anmcr

Dear Ros,

Sorry, I should have spotted that one.

The reconfiguration executable is created. However, the resulting reonfigured startdump (~/work/xffwf/ant4.start) has a problem as the 'surface temperature after timestep' field has been overwritten by the 'snow amount over land' field.

I had this problem once before with version 7.1. Myself and Willie couldn't figure why this was occurring, but I got a new 7.1 basis file and the problem never occurred.

Could you help?

Thanks,

Andrew

comment:8 Changed 10 years ago by willie

Hi Andrew,

I have just noticed that my PS25 Global reconfiguration (UM7.6) suffers the same problem, with a different start dump.

Regards,

Willie

comment:9 Changed 9 years ago by willie

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