mpirun error

General scientific issues regarding ROMS

Moderators: arango, robertson

Post Reply
Message
Author
wrh
Posts: 23
Joined: Fri Nov 14, 2008 4:26 pm
Location: zhejiang ocean university

mpirun error

#1 Post by wrh » Wed Jul 09, 2014 11:46 pm

dear all:
after one month running ,the mode got the following error. I search for all the posts of this website but got nothing .
so could you help me ?
mpirun noticed that process rank 3 with PID 23850 on node node13 exited on signal 11 (Segmentation fault)

wrh
Posts: 23
Joined: Fri Nov 14, 2008 4:26 pm
Location: zhejiang ocean university

Re: mpirun error

#2 Post by wrh » Thu Jul 10, 2014 11:28 am

if #undef AVERAGES then the model running is ok ! but #define AVERAGES will get the same error once again! so what's problem?

lanerolle
Posts: 157
Joined: Mon Apr 28, 2003 5:12 pm
Location: NOAA

Re: mpirun error

#3 Post by lanerolle » Thu Jul 10, 2014 2:20 pm

How big is your averages.nc file? With some NetCDF libraries and on some systems, only files as large as ~2.18Gb can be written and if you attempt to write files bigger than this, two things could happen: (a) your run could crash or (b)ROMS stops writing to the averages.nc file which continues to remain the same size for the rest of the simulation.

User avatar
kate
Posts: 3780
Joined: Wed Jul 02, 2003 5:29 pm
Location: CFOS/UAF, USA

Re: mpirun error

#4 Post by kate » Thu Jul 10, 2014 4:58 pm

What I know of segmentation faults is listed here. Is this right at the beginning of the run?

wrh
Posts: 23
Joined: Fri Nov 14, 2008 4:26 pm
Location: zhejiang ocean university

Re: mpirun error

#5 Post by wrh » Fri Jul 11, 2014 12:06 am

thank you all for the reply!
yes ,the mode was right till at the time step 7200 when average file would be created then got the error!
if USE_DEBUG=on then it seems everything is ok!

it is most likely that the problem has something to do with this issue viewtopic.php?f=14&t=2511.so i will try

Post Reply