[Met_help] [rt.rap.ucar.edu #92411] History for point_stat segementation fault?

Wed Oct 2 14:26:37 MDT 2019

----------------------------------------------------------------
  Initial Request
----------------------------------------------------------------

MET Help,

I'm trying to get MET running on a new computer. I've had great experience
running it on a Cray XC30, but now I'm trying to run it on a more generic
Intel-based computer. I'm running the same code that I'm using on the Cray
XC30, but I'm finding segmentation faults when running point_stat. Here is
the actual command being run:

/home/tyndalld/met/bin/point_stat
/tmpsd/tyndalld/cylc-scratch/ctrl_predator/share/20150815T0000Z/verif/fc_pre
_n1_2015081500_f0000.grib2
/tmpsd/tyndalld/cylc-scratch/ctrl_predator/share/pbnc_obs/prepbufr_201508150
0.nc
/gpfs/home/tyndalld/coamps_cylc_suite/met_config/PointStatConfigPresLevs
-outdir
/tmpsd/tyndalld/cylc-scratch/ctrl_predator/share/20150815T0000Z/verif/pre/ne
st1/ -log
/tmpsd/tyndalld/cylc-scratch/ctrl_predator/share/20150815T0000Z/run/verif/lo
g.2015081500_nest1_fc0000_pre

The output of the log file isn't exactly helpful (note that I've changed
verbosity up to 10):

DEBUG 1: Default Config File:
/home/tyndalld/met/share/met/config/PointStatConfig_default

DEBUG 1: User Config File:
/gpfs/home/tyndalld/coamps_cylc_suite/met_config/PointStatConfigPresLevs

DEBUG 4: Met2dDataFileFactory::new_met_2d_data_file() -> created new
Met2dDataFile object of type "FileType_Gb2".

With the standard error reporting:

/var/spool/pbs/mom_priv/jobs/4441019.a4au-m1.SC.orig: line 173: 666: Memory
fault

2019-10-02T01:01:58Z CRITICAL - failed/ERR

Any idea what I might be doing wrong? I'm wondering if it might be related
to the libraries listed in my LD_LIBRARY_PATH-could you tell me every
library that needs to be on that path in order to point stat to run? If it
helps any, pb2nc completed successfully. Thanks,

-Dan

Dr. Daniel P. Tyndall

United States Naval Research Laboratory

7 Grace Hopper Avenue Stop #2

Monterey, CA 93943

Phone: (831) 656-4054 - Fax: (831) 656-4769 - Cell: (630) 440-8622

Email:  <mailto:dan.tyndall at nrlmry.navy.mil> dan.tyndall at nrlmry.navy.mil

After hours email:  <mailto:dan.tyndall at outlook.com> dan.tyndall at outlook.com

----------------------------------------------------------------
  Complete Ticket History
----------------------------------------------------------------

Subject: point_stat segementation fault?
From: George McCabe
Time: Tue Oct 01 20:08:05 2019

Hi Dan,

I see you are getting a seg fault when running point_stat. I am
assigning
this ticket to John. He will get back to you in the near future.

Thanks,
George

On Tue, Oct 1, 2019 at 7:54 PM Tyndall, Dr. Dan via RT
<met_help at ucar.edu>
wrote:

>
> Tue Oct 01 19:54:14 2019: Request 92411 was acted upon.
> Transaction: Ticket created by Dan.Tyndall at nrlmry.navy.mil
>        Queue: met_help
>      Subject: point_stat segementation fault?
>        Owner: Nobody
>   Requestors: Dan.Tyndall at nrlmry.navy.mil
>       Status: new
>  Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=92411 >
>
>
> MET Help,
>
>
>
> I'm trying to get MET running on a new computer. I've had great
experience
> running it on a Cray XC30, but now I'm trying to run it on a more
generic
> Intel-based computer. I'm running the same code that I'm using on
the Cray
> XC30, but I'm finding segmentation faults when running point_stat.
Here is
> the actual command being run:
>
>
>
> /home/tyndalld/met/bin/point_stat
>
> /tmpsd/tyndalld/cylc-
scratch/ctrl_predator/share/20150815T0000Z/verif/fc_pre
> _n1_2015081500_f0000.grib2
>
> /tmpsd/tyndalld/cylc-
scratch/ctrl_predator/share/pbnc_obs/prepbufr_201508150
> 0.nc
>
/gpfs/home/tyndalld/coamps_cylc_suite/met_config/PointStatConfigPresLevs
> -outdir
>
> /tmpsd/tyndalld/cylc-
scratch/ctrl_predator/share/20150815T0000Z/verif/pre/ne
> st1/ -log
>
> /tmpsd/tyndalld/cylc-
scratch/ctrl_predator/share/20150815T0000Z/run/verif/lo
> g.2015081500_nest1_fc0000_pre
>
>
>
> The output of the log file isn't exactly helpful (note that I've
changed
> verbosity up to 10):
>
>
>
> DEBUG 1: Default Config File:
> /home/tyndalld/met/share/met/config/PointStatConfig_default
>
> DEBUG 1: User Config File:
>
/gpfs/home/tyndalld/coamps_cylc_suite/met_config/PointStatConfigPresLevs
>
> DEBUG 4: Met2dDataFileFactory::new_met_2d_data_file() -> created new
> Met2dDataFile object of type "FileType_Gb2".
>
>
>
> With the standard error reporting:
>
>
>
> /var/spool/pbs/mom_priv/jobs/4441019.a4au-m1.SC.orig: line 173: 666:
Memory
> fault
>
> 2019-10-02T01:01:58Z CRITICAL - failed/ERR
>
>
>
> Any idea what I might be doing wrong? I'm wondering if it might be
related
> to the libraries listed in my LD_LIBRARY_PATH-could you tell me
every
> library that needs to be on that path in order to point stat to run?
If it
> helps any, pb2nc completed successfully. Thanks,
>
>
>
> -Dan
>
>
>
> Dr. Daniel P. Tyndall
>
> United States Naval Research Laboratory
>
> 7 Grace Hopper Avenue Stop #2
>
> Monterey, CA 93943
>
> Phone: (831) 656-4054 - Fax: (831) 656-4769 - Cell: (630) 440-8622
>
> Email:  <mailto:dan.tyndall at nrlmry.navy.mil>
dan.tyndall at nrlmry.navy.mil
>
> After hours email:  <mailto:dan.tyndall at outlook.com>
> dan.tyndall at outlook.com
>
>
>
>
>

------------------------------------------------
Subject: point_stat segementation fault?
From: John Halley Gotway
Time: Wed Oct 02 10:33:23 2019

Hi Dan,

I see that you're getting a segfault when processing GRIB2 data in
MET.  I have definitely seen runtime problems in the past related to
GRIB2 data, and those problems were solved by recompiling the GRIB2
library.

Now I'm making an assumption here that the underlying problem really
is caused by GRIB2.  It's possible that there is another underlying
problem, but I do think it's likely caused by GRIB2.

Please take a look here at the instructions for compiling the GRIB2
library:
https://dtcenter.org/met-online-tutorial-metv8-0/tutorial-
setup/compilation/required-libraries

Specifically, please remove the "-D__64BIT__" flag from the makefile.

After you've recompiled the GRIB2 library, please recompile MET (make
clean; make install).  And then check to see if Point-Stat runs
without error.

Please let me know how it goes.

Thanks,
John

------------------------------------------------
Subject: point_stat segementation fault?
From: Tyndall, Dr. Dan
Time: Wed Oct 02 13:02:33 2019

Awesome John,

That seemed to do the trick! Thank you for your help.

-Dan

-----Original Message-----
From: John Halley Gotway via RT [mailto:met_help at ucar.edu]
Sent: Wednesday, October 2, 2019 9:33 AM
To: Tyndall, Dr. Dan
Subject: [rt.rap.ucar.edu #92411] point_stat segementation fault?

Hi Dan,

I see that you're getting a segfault when processing GRIB2 data in
MET.  I
have definitely seen runtime problems in the past related to GRIB2
data, and
those problems were solved by recompiling the GRIB2 library.

Now I'm making an assumption here that the underlying problem really
is caused
by GRIB2.  It's possible that there is another underlying problem, but
I do
think it's likely caused by GRIB2.

Please take a look here at the instructions for compiling the GRIB2
library:
https://dtcenter.org/met-online-tutorial-metv8-0/tutorial-
setup/compilation/required-libraries

Specifically, please remove the "-D__64BIT__" flag from the makefile.

After you've recompiled the GRIB2 library, please recompile MET (make
clean;
make install).  And then check to see if Point-Stat runs without
error.

Please let me know how it goes.

Thanks,
John

------------------------------------------------