ROMS random netcdf errors - fixed overnight??

Bug reports, work arounds and fixes

Moderators: arango, robertson

Post Reply
Message
Author
Timh37
Posts: 15
Joined: Thu May 09, 2019 3:25 pm
Location: NIOZ

ROMS random netcdf errors - fixed overnight??

#1 Unread post by Timh37 »

Dear all,

I've recently attempted to run a model that was working fine earlier this year. I'm using revision 966 and running ROMS in parallel on a HPC (currently through VPN). I'm using NETCDF4 and compiling with gfortran. Not having changed anything about the input or scripts, the model suddenly crashed instantly yesterday, giving random netcdf errors in the output files, ranging from
unable to open existing NetCDF file
to
error while reading variable: Uwind at TIME index = 56
. I noted this is similar to what is described in the FAQ. Running in debug mode resulted in the same errors. I'm quite sure pathing is right, as I can open these files using ncview just fine. Again, everything was working earlier.

This morning, I ran with exactly the same input files and scripts, and the model works fine. What happened? Why does my setup throw up errors at seemingly random moments in time and run fine at others? Is it possible that my code version got changed? Or more likely that something happened to the HPC? I really don't get it.

Best,
Tim

Post Reply