Scheduled Downtime
On Friday 21 April 2023 @ 5pm MT, this website will be down for maintenance and expected to return online the morning of 24 April 2023 at the latest

(RESOLVED) Segmentaition fault when running wrf.exe

yutinghe

New member
Hello all,
I'm running WRFV4.5 with sf_urban_physics=1 (i.e., SLUCM), unfortunately, the model stopped after integrating several hours. The error information which I got is as follows:

[ib0414:86410:0:86410] Caught signal 11 (Segmentation fault: address not mapped to object at address 0xfffffffe07e73c40)
==== backtrace (tid: 86410) ====
0 0x000000000004d455 ucs_debug_print_backtrace() ???:0
1 0x0000000002d58bd1 module_sf_sfclayrev_mp_psim_stable_() ???:0
2 0x0000000002d53fd4 module_sf_sfclayrev_mp_sfclayrev1d_() ???:0
3 0x0000000002d51ba6 module_sf_sfclayrev_mp_sfclayrev_() ???:0
4 0x00000000025e0d76 module_surface_driver_mp_surface_driver_() ???:0
5 0x0000000001e7287c module_first_rk_step_part1_mp_first_rk_step_part1_() ???:0
6 0x00000000016e731c solve_em_() ???:0
7 0x0000000001505328 solve_interface_() ???:0
8 0x00000000005b9ac3 module_integrate_mp_integrate_() ???:0
9 0x00000000005ba0e0 module_integrate_mp_integrate_() ???:0
10 0x0000000000417591 module_wrf_top_mp_wrf_run_() ???:0
11 0x000000000041754f MAIN__() ???:0
12 0x00000000004174e2 main() ???:0
13 0x0000000000022555 __libc_start_main() ???:0
14 0x00000000004173e9 _start() ???:0
=================================
forrtl: severe (174): SIGSEGV, segmentation fault occurred
Image PC Routine Line Source
wrf.exe 000000000322B07A for__signal_handl Unknown Unknown
libpthread-2.17.s 00002BA37A900630 Unknown Unknown Unknown
wrf.exe 0000000002D58BD1 Unknown Unknown Unknown
wrf.exe 0000000002D53FD4 Unknown Unknown Unknown
wrf.exe 0000000002D51BA6 Unknown Unknown Unknown
wrf.exe 00000000025E0D76 Unknown Unknown Unknown
wrf.exe 0000000001E7287C Unknown Unknown Unknown
wrf.exe 00000000016E731C Unknown Unknown Unknown
wrf.exe 0000000001505328 Unknown Unknown Unknown
wrf.exe 00000000005B9AC3 Unknown Unknown Unknown
wrf.exe 00000000005BA0E0 Unknown Unknown Unknown
wrf.exe 0000000000417591 Unknown Unknown Unknown
wrf.exe 000000000041754F Unknown Unknown Unknown
wrf.exe 00000000004174E2 Unknown Unknown Unknown
libc-2.17.so 00002BA37AD33555 __libc_start_main Unknown Unknown
wrf.exe 00000000004173E9 Unknown Unknown Unknown

I've looked up forums for solutions to segmentation fault, and set w_damping=1 and epssm=0.4. But these settings don't fix the error. However, when I turn off the urban mode (i.e., sf_urban_physics=0), the model is able to integrate continuously until the end time.

This confuses me, at this stage of WPS the static data I used is all default, I don't know why there is such a problem. Perhaps there is something in the latest version of WRF that needs to update the static files? I have attched my namelist and rsl files, I hope you can provide some help.
Thanks,
yuting
 

Attachments

  • namelist.input
    4.9 KB · Views: 11
  • rsl.tar.gz
    271 KB · Views: 3
Hi everyone,

I recently tested version 4.4.2 and version 4.2.2 separately. Under the condition that the remaining namelist parameters remain the same, I found that version 4.2.2 can normally use the SLUCM urban physics, but neither version 4.4.2 nor 4.5 can, but these two versions can be normally coupled with BEP.

So I think it might be that UCM is conflicting with some physics schemes, but I don't know which physics scheme.
 
Hi @yutinghe. Can you please check this link here? It explains how to fix a bug that might cause your error.
Anyway, are you running the model with the new hybrid 100-m global land cover dataset?
I'm facing an issue similar to yours. Here is the post: Running WRF V-4.5 with the new hybrid 100-m global land cover dataset.
Thanks a lot for your advice, my problem doesn't seem to be related to this bug. Still, I modified the source code accordingly, but the same error persists.
As for the new hybrid 100-m global land cover dataset you said, I haven't used it. Sorry for not being able to help you.
 
I had similar trouble with the urban canopy model in WRF 4.4, I ran it in January with no problems, but in the summer run, it kept getting segmentation errors. After fixing the issue at the interruption, it would break into another place for segmentation reasons due to something else, and it still hasn't been resolved.
 
I had similar trouble with the urban canopy model in WRF 4.4, I ran it in January with no problems, but in the summer run, it kept getting segmentation errors. After fixing the issue at the interruption, it would break into another place for segmentation reasons due to something else, and it still hasn't been resolved.
Yes, this error has not been resolved, so I back to use version 4.2.
 
We will contact the developer and hopefully to find the solution. It may take some time. Thanks for your attention.
 
I am also having an issue when using SLUCM. After days of simulation I get a segmentation fault. But when running without SLUCM there are no issues.
 
@yutinghe , @sabina12 , and @jkqian , which land surface model are you using coupled with the UCM?
I found out from the WRF User's Guide that UCM can be coupled with Noah LSM (sf_surface_physics = 2) ONLY.
With this fix, I managed to run the model without any other issues.
 

Attachments

  • Screenshot 2023-08-06 at 11.27.43.png
    Screenshot 2023-08-06 at 11.27.43.png
    93.9 KB · Views: 27
@yutinghe , @sabina12 , and @jkqian , which land surface model are you using coupled with the UCM?
I found out from the WRF User's Guide that UCM can be coupled with Noah LSM (sf_surface_physics = 2) ONLY.
With this fix, I managed to run the model without any other issues.
Thanks for your reminder, in fact, I'm using Noah LSM. I would like to know how many simulation days your model was run, becase I can run the model for several days successfully, but went wrong for one month simulation.
 
Thanks for your reminder, in fact, I'm using Noah LSM. I would like to know how many simulation days your model was run, becase I can run the model for several days successfully, but went wrong for one month simulation.
I ran the model for a couple of weeks. Before changing the LSM option, I couldn't even run it for an hour.
I'm planning to run several-months-simulations from the middle of September. I'll keep you posted, but I hope your model will be okay by then.
 
I ran the model for a couple of weeks. Before changing the LSM option, I couldn't even run it for an hour.
I'm planning to run several-months-simulations from the middle of September. I'll keep you posted, but I hope your model will be okay by then.
Thanks for your kindness and looking forward to your new results.
 
@yutinghe , @sabina12 , and @jkqian , which land surface model are you using coupled with the UCM?
I found out from the WRF User's Guide that UCM can be coupled with Noah LSM (sf_surface_physics = 2) ONLY.
With this fix, I managed to run the model without any other issues.
I used sf_surface_physics=2 and when using WRF version 4.5 while it runs for several days after that it doesn't. Using WRF4.2 the simulation completes with no issue.
 
Top