forrtl: severe (174): SIGSEGV, segmentation fault occurred

[wonseok ko]

Hello, I would like to inquire about a WRF error.

forrtl: severe (174): SIGSEGV, segmentation fault occurred

This error appeared, and it only happens for the period of March 20, 2022. It does not occur for other periods in March.
What’s strange is that the simulation sometimes works on a different account on the same server, and even in my account, the simulation for this period worked well a month ago.

I heard it might be due to memory shortage, so I tried ulimit -s unlimited. Is it correct to add this to the .bash_profile?

Also, I’ve been running WRF using PBS, and it seems like this issue started after using PBS. Could this be related?

I am attaching the namelist.input file and the error file.

Thank you.

 

[kwerner]

What’s strange is that the simulation sometimes works on a different account on the same server, and even in my account, the simulation for this period worked well a month ago.

When this worked, was the simulation 100% identical to the one that fails? i.e., are you using the exact same namelist.wps and namelist.input, the same input data, same version of the model, compiled with the same compiler, libraries, and versions of those compilers and libraries, for the same dates, etc.?

If so, then the issue is likely specific to the system or the environment setup and I would suggest that you speak to a systems administrator about the problem. If the simulations are not identical, can you package all the rsl* files for the failed run into a single *.tar file and attach that? Please also share the namelists (.wps and .input) for a simulation that works, along with the namelist.wps file for the one that corresponds to the namelist.input file you attached above. Thanks!