SIGSEGV, segmentation fault error

[jgriley]

Hi,

I am been trying to run WRFV3.7.1 on Cheyenne, I am getting a segmentation fault error. I have tried shorting my time steps and running with more processors, but that have not work. I read online that it might be a stacksize, but doing ulimit -s unlimited did not work either. Attach is my namelist.input and below is the error being shown.

INITIALIZE THREE Noah LSM RELATED TABLES
Tile Strategy is not specified. Assuming 1D-Y
WRF TILE 1 IS 35 IE 42 JS 24 JE 25
WRF NUMBER OF TILES = 1
forrtl: severe (174): SIGSEGV, segmentation fault occurred

 

[kwthomas]

Crashing before even acknowledging the initial conditions is rather unusual. Is there anything useful in any
of the rsl* files?

You might try settingi the "debug_level" in "namelist.input" to 9999 and see what happens.

You might try using twice the nodes, but only half the cores per node. I won't think this would cause SIGSEGV,
but you never know.

You might try recreating your data. I'll make an educated guess that some of it is corrupt.

 

[jgriley]

I tried recreating the data and I still got the same error. I set the debug_level in the namelist.input to 9999 and now I get:

d01 2007-11-04_00:30:00 CALL rrtmg_lw
forrtl: severe (174): SIGSEGV, segmentation fault occurred
Image PC Routine Line Source
wrf.exe 0000000003136691 Unknown Unknown Unknown
wrf.exe 00000000031347CB Unknown Unknown Unknown
wrf.exe 00000000030EE664 Unknown Unknown Unknown
wrf.exe 00000000030EE476 Unknown Unknown Unknown
wrf.exe 0000000003077AE9 Unknown Unknown Unknown
wrf.exe 000000000307F356 Unknown Unknown Unknown
libpthread-2.19.s 00002AAAAB1E2870 Unknown Unknown Unknown
wrf.exe 0000000001F11DC9 Unknown Unknown Unknown
wrf.exe 0000000001EB3A9F Unknown Unknown Unknown
wrf.exe 0000000001EAE44E Unknown Unknown Unknown
wrf.exe 0000000001EA3F3C Unknown Unknown Unknown
wrf.exe 0000000001832A2A Unknown Unknown Unknown
wrf.exe 0000000001919542 Unknown Unknown Unknown
wrf.exe 0000000001304571 Unknown Unknown Unknown
wrf.exe 0000000001189924 Unknown Unknown Unknown
wrf.exe 00000000004FFBFF Unknown Unknown Unknown
wrf.exe 0000000000405EC1 Unknown Unknown Unknown
wrf.exe 0000000000405E7F Unknown Unknown Unknown
wrf.exe 0000000000405E1E Unknown Unknown Unknown
libc-2.19.so 00002AAAABBF8B25 __libc_start_main Unknown Unknown
wrf.exe 0000000000405D29 Unknown Unknown Unknown

 

[Ming Chen]

For a grid interval of 182.9km, time step can be at least 1080. There is no reason to set it to 30.
I notice that the grid number is 100 x 32. is there any reason why the grid number is so small in the Y-direction? Where is the domain located?
Please look at the map_factors to make sure they are reasonable.

If the model crashed immediately, it often indicates the input data are wrong. Please double check.