Typical UM Errors

Note: This page is still a stub. Please add to this page.

Errors during Extract


When compiling the UM, the first step is for the UMUI(x) to extract the source code. If an error occurs, check for these things:

Step 1: Did the error occur during the base, model, or reconfiguration extraction?
Step 2: Check for the extraction log, it is either in
~/UM_OUTPUT/<JOBID>/<umbase or ummodel or umrecon>/ext.out
or
/scratch/users/$USER/<JOBID>/<umbase or ummodel or umrecon>/ext.out

The final lines typically show at which command the error occurred.

SVN errors


If you get a permission denied error from an SVN, it might be because svn needed your password, and didn't get it. The solution is to run

$ svn ls https://access-svn.nci.org.au/svn/um/branches/

If you get asked for a password, you need to supply your NCI password. And then it will ask you whether it should store it in plain text.
As much as I hate to say it: you need to answer 'yes'.

And immediately afterwards, you have to run the command:

$ chmod -R go-rwx $HOME/.subversion

SCP and RSYNC errors


If you get an error like this:

mkdir: cannot create directory `/abc123': Permission denied

(where abc123 is your username), the most likely explanation is that you haven't set the DATAOUTPUT environment variable on accessdev.
If you are using bash, add this line to your $HOME/.bashrc file:

export DATAOUTPUT=/short/${PROJECT}/${USER}/UM_ROUTDIR

If you are using tcsh or csh, add this line to your $HOME/.login

setenv DATAOUTPUT /short/${PROJECT}/${USER}/UM_ROUTDIR

Then log out of accessdev and log back in, open up umuix again, and check whether the error went away.

Errors during Submit


Errors during Compile


Errors during Run


REPLANCA: PP HEADERS ON ANCILLARY FILE DO NOT MATCH


The model is trying to update an ancillary field, but it is looking for a date that is past the end of the ancillary file. For instance, say an ancillary file is valid for dates between 1850-01-01 and 2001-01-01 and the model is trying to update fields for the date 2001-02-01. The model can't find this data in the file, so it crashes.

Near the end of the .leave file there will be a block that looks like

  STASH code in dataset                    122   STASH code requested
                    58
 'Start' position of lookup tables for dataset in overall lookup array
                   481
                   122                    58                    39
  UP_ANCIL : Error in REPLANCA.
 
 *********************************************************************************
 UM ERROR (Model aborting) :
 Routine generating error: UP_ANCIL
 Error code:                    239
 Error message:
REPLANCA: PP HEADERS ON ANCILLARY FILE DO NOT MATCH
 *********************************************************************************

The important bit here is 'STASH code requested', which is in this case was 58. Internally the UM refers to each model field (U, V, temperature, etc.) by a numeric STASH code. You can check field is represented by a code by going to https://accessdev.nci.org.au/umdocs/7.3/stash_browse. The STASH codes are divided into different sections, for errors like this you'll find the variable in section 0) Prognostic Variables. Scroll down the item list to find the name for the error code - the code 58 corresponds to SULPHUR DIOXIDE EMISSIONS.

To find out which file this corresponds to you'll need to go back to the UMUI, and check the screen 'Atmosphere -> Ancillary and Input data files -> Climatologies'. It should be fairly simple to work out the entry to use (in our example you'd look at 'Sulphur-Cycle Emissions') but if you have trouble check with the helpdesk. You'll need to replace the file here with one that covers the next part of your model run.

Its common for this error to occur when running an AMIP-type simulation and going past 2001, where the data changes from Historical to the RCP scenarios. If this is the case you'll need to change Ozone (files starting with SPARCO3), 2d Sulphur Cycle (sulp) Soot (BC_hi), Biomass (Bio), and OCFF (OCFF). You can find ancillary files for the different scenarios under /projects/access/data/ancil/CMIP5, or ask the CMS team at climate_help@nf.nci.org.au and we'll help you locate the right ancillary file.

DRLANDF1 : Error in FILE_OPEN


Symptom: The UM aborts with the error message:
 *********************************************************************************
 UM ERROR (Model aborting) :
 Routine generating error: UM_SHELL
 Error code:                      1
 Error message:
 DRLANDF1 : Error in FILE_OPEN.
 *********************************************************************************

This indicates that the start dump (NRUN) or restart dump (CRUN) could not be opened. Search the leave file for the string "Unit 21". You should find a line like this:

OPEN:  File <somefile> to be Opened on Unit 21 does not Exist

<somefile> might be empty, in which case there will be 2 spaces between "File" and "to be Opened".
Otherwise, check that the file exists and that you have read permissions.

If this happens during an NRUN, check the settings in
Model Selection -> Atmosphere -> Ancillary and input data files --> Start dump
Check the settings under "Specify initial dump"

If the file name is an empty string, and you are running a CRUN, open
Model Selection -> Atmosphere -> Control -> Post processing, Dumping & Meaning --> Dumping and meaning
and ensure that the restart dump frequency is consistent with your per-submission run length.
(That is, if your run length is 1 year, you may have monthly or yearly restart dumps, but not 5-yearly restart dumps.)