I’m trying to run the RNS with a modified SST ancillary file, but the SST ancillary doesn’t appear to be used. There is a note under the sstsmc_anc option (make extra SST and SMC ancillaries) that says ‘these extra ancillaries are not used by the default version of the nesting suite’.
You can add more ancil files under um > namelist > Configure ancils .... Right click in the table and add a new section, then right click the new entry and select ‘view this section’.
The setup should look something like this - SST goes to stash 24. For a climate run you’d set it to update the sst every so often from the ancil file, but that’s probably not needed in a regional run.
The modified SST ancillary will have daily mean values so will need to update the SST. I’ve set the update interval to 1 day but the um_fcst step is failing with an error message
INANCILA: Insifficient space for LOOKUP headers.
I’ve tried increasing nancil_lookupsa but it then fails with the following?
@BecJackson has this been resolved? If not, can you provide instructions on how to reproduce your error? Also just for future note, please try to avoid posting screenshots of errors and instead include them as plain text so that they are searchable.
I’ve added the SST ancillary to the Configure ancils table as @Scott mentioned above. The suite runs fine when update_anc = False, with either a monthly (created by the RAS) or daily mean (modified from monthly file) SST ancil, but when I try to update the SST field (update_anc = True) I get the below error regardless or which update period I use (1 day, month etc).
Any help would be really appreciated!
Cheers,
Bec
in job.err:
[14] exceptions: An exception was raised:11 (Segmentation fault)
[14] exceptions: the exception reports the extra information: Invalid permissions for mapped object.
[14] exceptions: whilst in a serial region
[14] exceptions: Task had pid=1452650 on host gadi-cpu-clx-0827.gadi.nci.org.au
[14] exceptions: Program is “um-atmos.exe”
[14] exceptions: calling registered handler @ 0x00416350
Warning in umPrintMgr: umPrintExceptionHandler : Handler Invoked
[9] exceptions: Done callbacks
[9] exceptions: *** GLIBC ***
[9] exceptions: Data address (si_addr): 0x149ad37f18c8; rip: 0x00534dc2
[0] exceptions: [backtrace]: ( 13) : Address: [0x004089de]
[0] exceptions: [backtrace]: ( 13) : line information unavailable error code: 4 (Program name & path is not absolute)
[0] exceptions:
[0] exceptions: To find the source line for an entry in the backtrace;
[0] exceptions: run addr2line --exe=</path/too/executable>
[0] exceptions: where address is given as [0x] above
[0] exceptions:
[0] exceptions: Done backtrace
Primary job terminated normally, but 1 process returned
a non-zero exit code. Per user-direction, the job has been aborted.
mpirun detected that one or more processes exited with non-zero status, thus causing
the job to be terminated. The first process to do so was:
There’s a limit to the number of 2d field slices the model will read from an ancil file, I can’t remember the exact figure but it should be at least a couple thousand. What’s in your file? If it’s like a decade of daily data or something like that you may need to break it up into multiple sections.
Thanks @Scott! The modified ancil file contains daily data for 1 year. I thought the issue might be the date format (eg 2018-01-01 instead of some reference year?), but I get the same error when using the default monthly ancil file created by the RAS, regardless of which update period I set.
Hello @BecJackson, thanks for your post, how’d you go?
If resolved, can you please mark one of the posts as a solution or summarise a resolution that worked for you? Cheers