[ncl-talk] fileattdef seg fault on CESM LE files

[David Brown]

Hi David,

Assuming you have just logged into yellowstone and you have not
modified the module status following the initial
"Restoring modules to system default" you (or anyone with access to
yellowstone) can run the latest dev version of NCL with the following
settings:

module swap intel gnu/4.7.2
export NCARG_ROOT=/glade/p/work/haley/dev/ncl-trunk-gnu472-opt
export PATH=$NCARG_ROOT/bin:$PATH

(bash syntax -- modify as appropriate for csh)

As far as the error is concerned, it was simply a typo in the function
that gets basic information from a file, including the names of the
global attributes. This function is specific to NetCDF4 or other
"advanced" file types. It is called from the fillattdef function.
Ironically, the error was in the part that gets the names of the
variables in the file -- it never reached the section that gets the
global attribute information.
 -dave



On Thu, Jun 4, 2015 at 1:08 PM, David B. Reusch <dreusch at ees.nmt.edu> wrote:
> I'm probably too swamped w/meeting prep and travel the rest of this
> month but I'd be curious to know how to access the development version
> for when I do have a chance.
>
> And I admit to also being curious about what the problem was :-)
>
> Thanks,
> Dave
>
> David Brown wrote on 6/3/15 9:36 AM:
>> Hi David,
>> I want to let you know that this issue has been fixed in the
>> development version of NCL. Let us know if you'd like to try it out on
>> yellowstone.
>>   -dave
>>
>> On Tue, May 26, 2015 at 2:11 PM, Mary Haley <haley at ucar.edu> wrote:
>>> Hi Dave,
>>>
>>> Thanks for providing the script.
>>>
>>> I was able to reproduce your results with NCL 6.2.1, NCL 6.3.0 (both intel
>>> and gcc versions) and a test version of NCL, which is our development
>>> version. I've created a ticket, NCL-2223.
>>>
>>> If/when we get this working, I can point you to a development version that
>>> you can use on yellowstone.
>>>
>>> --Mary
>>>
>>>
>>>
>>> On Tue, May 26, 2015 at 11:37 AM, David B. Reusch <dreusch at ees.nmt.edu>
>>> wrote:
>>>> I seem to have found an NCL seg fault when copying global attributes with
>>>> fileattdef from CESM Large Ensemble (LE) files.  This was part of a larger
>>>> script I've been using to regrid for multiple years. The attached script
>>>> just does the fileattdef step and it works on at least one of my other,
>>>> non-CESM LE files.  I've tried it on multiple CESM LE files with the same
>>>> results.  Copying the attributes individually works so I'm assuming it's
>>>> related to doing all the attributes together.
>>>>
>>>> This is on yellowstone with NCL 6.3.
>>>>
>>>> My temporary workaround is to just comment out this step so that I can
>>>> continue my file processing.  Since that means losing global metadata, a
>>>> better solution/fix would be preferred.
>>>>
>>>> Thanks,
>>>> Dave
>>>>
>>>> --
>>>> Associate Research Professor of Climatology
>>>> Dept of Earth and Environmental Science
>>>> MSEC 304; 801 Leroy Place
>>>> New Mexico Tech
>>>> Socorro, NM 87801
>>>>
>>>>
>>>> _______________________________________________
>>>> ncl-talk mailing list
>>>> List instructions, subscriber options, unsubscribe:
>>>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>>>>
>>>
>>> _______________________________________________
>>> ncl-talk mailing list
>>> List instructions, subscriber options, unsubscribe:
>>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>>>
>
> --
> Associate Research Professor of Climatology
> Dept of Earth and Environmental Science
> MSEC 304; 801 Leroy Place
> New Mexico Tech
> Socorro, NM 87801
>
> _______________________________________________
> ncl-talk mailing list
> ncl-talk at ucar.edu
> List instructions, subscriber options, unsubscribe:
> http://mailman.ucar.edu/mailman/listinfo/ncl-talk