[gradsusr] Low Level I/O Error: Seek error on data file [GrADS] -
Leopoldo Alvarez
lap.bolsa at gmail.com
Mon Sep 23 16:39:31 EDT 2013
Hi Matias,
I could do this calculation before without problems. A year or more
data. I think now is the architecture I use.
Linux login1 2.6.16.53-0.16-ppc64 #1 SMP Tue Oct 2 16:57:49 UTC 2007
ppc64 ppc64 ppc64 GNU/Linux
The GrADS can handle files larger than 2 GB in size.
Leopoldo
> Hi Leopoldo,
>
> One solution could be to split big MM5 output file into smaller
files. You
> can do this with the following fortran program:
> ftp://ftp.ucar.edu/mesouser/MM5V3/Util/SplitMMFiles.f
> Then you execute mm5_to_grads to each smaller files.
> I hope this helps you.
> Regards
> Matias
El 23/09/13 21:35, Leopoldo Alvarez escribió:
> Dear Jennifer and James,
>
> The size file that I generated with my mm5_to_grads.csh is ->
> [login1 17] /gpfs/projects/XXXXX/XXXXXX/mesoescala/MM5toGrADS > du -k
> test.dat
> *68926752* 2013-09-22 11:11 test.dat
>
> I think it can be that grad.exe is bad compiled? Is this file
> (test.dat) bad generated? Is it really a problem of file size? how I
> fix it?
> >>>> I'm on a machine that works -> uname -a
> >>>> Linux login1 2.6.16.53-0.16-ppc64 #1 SMP Tue Oct 2 16:57:49 UTC
> 2007 ppc64 ppc64 ppc64 GNU/Linux
>
> > When you are using PDEF, the xsize and ysize come from the first two
> args in the PDEF entry. Here is my estimate of what the file size
> should be: sfc_vars * x * y * t * 4 bytes = 11* 60 * 60 * 1036 * 4 =
> 164102400
> > plus
> > z-varying-vars * z * x * y * t * 4 bytes = 22 * 25 * 60 *60 * 1036 *
> 4 = 8205120000
> > equals 8369222400 bytes.
>
> I paste another GrADS descriptor file, sorry.
> I can see the data well, until the time of step 1064 (set t 1064, then
> 'd t' it`s ok), but when i exceeded from 1064 time of step gives me
> that error ...
>
> /ga-> set t *1065*
> Time values set: 2012:2:15:8 2012:2:15:8
> ga-> d t
> Notice: Automatic Grid Interpolation Taking Place
> Low Level I/O Error: Seek error on data file
> Data file name = test.dat
> 650 rc=-1 fpos=-2145713296 pfi->fhdr=0
> Error occurred when seeking to byte 7081408
> Data Request Error: Error for variable 't'
> Error ocurred at column 1
> DISPLAY error: Invalid expression
> Expression = t/
>
> My full GrADS descriptor file is OK to *8737 *time values (one year) ->
>
> dset ^test.dat
> options big_endian
> title MM5 data
> undef -9999.
> pdef 60 60 lcc 32.63 -112.54 1.00 1.00 30.00 60.00
> -110.34 6000. 6000.
> xdef 149 linear -112.65 0.0270
> ydef 123 linear 32.63 0.0270
> zdef 25 levels
> 0.99689
> 0.99191
> 0.98944
> 0.98818
> 0.97377
> 0.94500
> 0.91000
> 0.87000
> 0.82500
> 0.77500
> 0.72500
> 0.67500
> 0.62500
> 0.57500
> 0.52500
> 0.47500
> 0.42500
> 0.37500
> 0.32500
> 0.27500
> 0.22500
> 0.17500
> 0.12500
> 0.07500
> 0.02500
> tdef*8737 *linear 00:00Z02JAN2012 60MN
> vars 33
> rc 0 99 accum conv pcn (cm)
> rn 0 99 accum non-c pcn (cm)
> ter 0 99 ter elevation (m)
> xlat 0 99 cross lat (degree)
> xlon 0 99 cross lon (degree)
> lu 0 99 land use
> t2m 0 99 2 m temperature (K)
> q2m 0 99 2m mix ratio (kg/kg)
> u10 0 99 10 m u wind (m/sec)
> v10 0 99 10 m v wind (m/sec)
> pslv 0 99 sea level prs (mb)
> u 25 99 u wind (m/s)
> v 25 99 v wind (m/s)
> w 25 99 vertical vel (m/s)
> pp 25 99 prs pert (Pa)
> t 25 99 temperature (C)
> q 25 99 mixing ratio (kg/kg)
> clw 25 99 cloud water (kg/kg)
> rnw 25 99 rain water (kg/kg)
> rtnd 25 99 rad tend (K/day)
> z 25 99 height AGL (m)
> h 25 99 geopot height (m)
> td 25 99 dewpoint temp (C)
> rh 25 99 rel humidity (%)
> th 25 99 potential temp (K)
> the 25 99 theta-e (K)
> prs 25 99 pressure (Pa)
> vor 25 99 vorticity (s-1)
> pv 25 99 potential vort (pvu)
> dbz 25 99 reflectivity (dbz)
> div 25 99 divergence (s-1)
> dir 25 99 dir (degrees)
> tadv 25 99 temp advection (K/s)
> endvars
>
>
> Leopoldo
>
>
>
> On Sep 23, 2013, at 3:03 PM, James T. Potemra wrote:
>
> >
> > Leopoldo:
> >
> > You have a couple problems here. First, your descriptor file
> specifies 1036 time values, but you are trying to access t=1064 (?).
> You should be sure about the number of time steps in your data set.
>
> > Second, as Jennifer suggested, x*y*z*t*4 should be your file size.
> In this case you have 22 variables with 25 levels, and 11 with 1; 149
> x-points; 123 y-points and 1036 time points. I think your file should
> therefore be (22*25+11)*149*123*1036*4 = 426,063,116,368 bytes.
>
> When you are using PDEF, the xsize and ysize come from the first two
> args in the PDEF entry. Here is my estimate of what the file size
> should be:
> sfc_vars * x * y * t * 4 bytes = 11* 60 * 60 * 1036 * 4 = 164102400
> plus
> z-varying-vars * z * x * y * t * 4 bytes = 22 * 25 * 60 *60 * 1036 * 4
> = 8205120000
> equals 8369222400 bytes.
>
> If you have 1064 time steps instead of 1036 then you would have a file
> size of 8595417600 bytes.
>
> --Jennifer
>
>
> >
> > Jim
> >
> > On 9/23/13 8:24 AM, Leopoldo Alvarez wrote:
> >> Hi Jennifer,
> >>
> >> >> and my data file .dat weighs 32 Gb
> >> >A precise file size is required, down to the last byte.
> >>
> >> Now my data file .dat weighs 66 Gb
> >> 68926752 - 66G 2013-09-22 11:11 test.dat
> >>
> >> I can see the data well, until the time of step 1064 (set t 1064,
> then 'd t' it`s ok), but when i exceeded from 1064 time of step gives
> me that error ...
> >>
> >> My full GrADS descriptor file ->
> >>
> >> dset ^test.dat
> >> options big_endian
> >> title MM5 data
> >> undef -9999.
> >> pdef 60 60 lcc 32.63 -112.54 1.00 1.00 30.00 60.00
> -110.34 6000. 6000.
> >> xdef 149 linear -112.65 0.0270
> >> ydef 123 linear 32.63 0.0270
> >> zdef 25 levels
> >> 0.99689
> >> 0.99191
> >> 0.98944
> >> 0.98818
> >> 0.97377
> >> 0.94500
> >> 0.91000
> >> 0.87000
> >> 0.82500
> >> 0.77500
> >> 0.72500
> >> 0.67500
> >> 0.62500
> >> 0.57500
> >> 0.52500
> >> 0.47500
> >> 0.42500
> >> 0.37500
> >> 0.32500
> >> 0.27500
> >> 0.22500
> >> 0.17500
> >> 0.12500
> >> 0.07500
> >> 0.02500
> >> tdef 1036 linear 00:00Z02JAN2012 60MN
> >> vars 33
> >> rc 0 99 accum conv pcn (cm)
> >> rn 0 99 accum non-c pcn (cm)
> >> ter 0 99 ter elevation (m)
> >> xlat 0 99 cross lat (degree)
> >> xlon 0 99 cross lon (degree)
> >> lu 0 99 land use
> >> t2m 0 99 2 m temperature (K)
> >> q2m 0 99 2m mix ratio (kg/kg)
> >> u10 0 99 10 m u wind (m/sec)
> >> v10 0 99 10 m v wind (m/sec)
> >> pslv 0 99 sea level prs (mb)
> >> u 25 99 u wind (m/s)
> >> v 25 99 v wind (m/s)
> >> w 25 99 vertical vel (m/s)
> >> pp 25 99 prs pert (Pa)
> >> t 25 99 temperature (C)
> >> q 25 99 mixing ratio (kg/kg)
> >> clw 25 99 cloud water (kg/kg)
> >> rnw 25 99 rain water (kg/kg)
> >> rtnd 25 99 rad tend (K/day)
> >> z 25 99 height AGL (m)
> >> h 25 99 geopot height (m)
> >> td 25 99 dewpoint temp (C)
> >> rh 25 99 rel humidity (%)
> >> th 25 99 potential temp (K)
> >> the 25 99 theta-e (K)
> >> prs 25 99 pressure (Pa)
> >> vor 25 99 vorticity (s-1)
> >> pv 25 99 potential vort (pvu)
> >> dbz 25 99 reflectivity (dbz)
> >> div 25 99 divergence (s-1)
> >> dir 25 99 dir (degrees)
> >> tadv 25 99 temp advection (K/s)
> >> endvars
> >>
> >>
> >> El 21/09/13 09:35, Leopoldo Alvarez escribió:
> >>> Hi Jennifer,
> >>>
> >>> My binary data grad.exe file is 210 Kb. and data file .dat weighs
> 32 Gb.
> >>> xsize*ysize*zsize*tsize*esize*4 , Where I see these parameters? i
> don´t know how to fix it, where ?, I need to recompile ?.
> >>>
> >>> best regards,
> >>>
> >>> Leopoldo ÁLVAREZ
> >>> Tenerife - Canary Islands
> >>>
> >>>
> >>>
> >>>
> >>> El 20/09/13 23:13, Leopoldo Alvarez escribió:
> >>>> Hi jennifer,
> >>>>
> >>>> I am running a script (nuevo.gs) to extract info domain mm5 model
> data. I use GrADS 2.0.2 version. I use 'options big_endian' but my
> data file weighs 32 GB and
> >>>> shows multiple mutiple on screen- ->
> >>>>
> >>>> Low Level I/O Error: Seek error on data file
> >>>> Data file name = test.dat
> >>>> 650 rc=-1 fpos=-172300 pfi->fhdr=0
> >>>>
> >>>> I think the problem is when running the mm5_to_grads.csh and
> creates the .ctl and .dat and not script. I can see the data well,
> until the time of step 1062 (set t 1062, then 'd t' it`s ok), but when
> i exceeded from 1062 time of step gives me that error ... How i can
> run my script saving this problem???. One clue
> http://gradsusr.org/pipermail/gradsusr/2008-April/006883.html
> >>>>
> >>>> I'm on a machine that works -> uname -a
> >>>> Linux login1 2.6.16.53-0.16-ppc64 #1 SMP Tue Oct 2 16:57:49 UTC
> 2007 ppc64 ppc64 ppc64 GNU/Linux
> >>>>
> >>>>
> >>>> nuevo.gs script
> >>>> ***************
> >>>> ***************
> >>>> 'set lon -110.1829'
> >>>> 'set lat 34.6518'
> >>>> 'set lev 0.99199'
> >>>>
> >>>> #8737
> >>>> timeend = 4198
> >>>> i = 1
> >>>> while(i <= timeend)
> >>>> 'set t 'i''
> >>>> 'q time'
> >>>> timehr = substr(result, 8, 3)
> >>>> timedy = substr(result, 11, 2)
> >>>> timemo = substr(result, 13, 3)
> >>>> timeyr = substr(result, 16, 4)
> >>>>
> >>>> #velocidad viento
> >>>> 'tg1=mag(u,v)'
> >>>> 'd tg1'
> >>>> a.nr1 = subwrd(result,4)
> >>>> #direccion viento
> >>>> 'd dir'
> >>>> a.nr11 = subwrd(result,10)
> >>>> # tv de Jeffrey iowa
> >>>> 'tv=t+t*0.61*q'
> >>>> 'rho=prs/(286.9*tv)'
> >>>> 'd rho'
> >>>> a.nr = subwrd(result,4)
> >>>> 'd t-273.15'
> >>>> a.nr2 = subwrd(result,10)
> >>>> 'd rh'
> >>>> a.nr3 = subwrd(result,10)
> >>>> 'd prs/100'
> >>>> a.nr4 = subwrd(result,10)
> >>>>
> >>>> val = timeyr%"-"%timemo%"-"%timedy"-"%timehr" "%a.nr1" "%a.nr11"
> "%a.nr" "%a.nr2" "%a.nr3" "%a.nr4
> >>>> rec = write('drylake_80v.txt',val, append)
> >>>> i = i + 1
> >>>> endwhile
> >>>> *******************
> >>>> *******************
> >>>>
> >>>> best regards,
> >>>>
> >>>> Leopoldo ÁLVAREZ
> >>>> Tenerife - Canary Islands
> >>>
> >>
> >>
> >>
> >> _______________________________________________
> >> gradsusr mailing list
> >> gradsusr at gradsusr.org
> >> http://gradsusr.org/mailman/listinfo/gradsusr
> >
> > _______________________________________________
> > gradsusr mailing list
> > gradsusr at gradsusr.org
> > http://gradsusr.org/mailman/listinfo/gradsusr
>
> --
> Jennifer M. Adams
> Center for Ocean-Land-Atmosphere Studies (COLA)
> 111 Research Hall, Mail Stop 2B3
> George Mason University
> 4400 University Drive
> Fairfax, VA 22030
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://gradsusr.org/pipermail/gradsusr/attachments/20130923/eef8cf9f/attachment-0001.html
More information about the gradsusr
mailing list