Re: Bad read of grib file

From: Mary Haley (haley AT XXXXXX)
Date: Mon Mar 29 2004 - 21:58:50 MST

  • Next message: Guido Vettoretti: "CMPS iso utility"

    >
    > Hi,
    >
    > I have a set of grib files to open. They're climate model data,
    > downloaded from the IPCC DDC here:
    >
    > http://ipcc-ddc.cru.uea.ac.uk/dkrz/dkrz_index.html
    >
    > The files come in at between 1.5 and 6MB, so hardly huge.
    >
    > The problem is that one set of files - from the CCSR/NIES model (last
    > one in the table), give an error when I try loading them:
    >
    > ncl 0> a = addfile("CCSRNIES_SRES_A2_PREC_1990-2019.grb", "r")
    > fatal:NclMalloc Failed:[errno=12]:Not enough space
    > Segmentation fault (core dumped)
    >
    > Is this a known/recognised sequence of events? Grib files from all
    > other models load without problem. The problem is with four separate
    > files so I don't think it's a download/corruption issue; wgrib also
    > reads them without trouble.
    >
    > ncl version 4.2.0.a027, Tru64 Unix on an Alpha.
    >
    > (it also fails with 4.2.0.a028 on OS X, but no error message, just an
    > infinite wait :-/ )
    >
    > Cheers for any thoughts,
    >
    > Harry
    > Ian "Harry" Harris
    > Climatic Research Unit
    > University of East Anglia
    > Norwich NR2 4HG
    > United Kingdom

    Dear Harry and others,

    One of our developers created a work-around for this problem, and it
    will be available in the next version of NCL (no scheduled date
    yet).

    Here's some info on the work-around, In the developer's own words:

        This is a workaround fix for Grib files from the IPCC Data
        Distribution Center. (http://cera-www.dkrz.de/IPCC_DDC/)

        Several unrecognized centers are used for these files. The
        work-around code is activated when an unrecognized center is
        encountered. I have only looked at a limited number of these
        files, so I do not know how comprehensive this fix is.

        User-defined parameter tables are still needed and I do not
        attempt to supply a name for the center.

    If anybody wants to try this new version of NCL now, send me a
    personal email (haley AT ucar.edu) and I'll see if I can provide you with
    a precompiled binary.

    --Mary
    _______________________________________________
    ncl-talk mailing list
    ncl-talk AT ucar.edu
    http://mailman.ucar.edu/mailman/listinfo/ncl-talk



    This archive was generated by hypermail 2b29 : Mon Mar 29 2004 - 22:24:44 MST