Re: 20040512: netCDF Decoders - Red Hat Linux - metar2nc errors

On Wed, 12 May 2004, Unidata Support wrote:

>
> ------- Forwarded Message
>
> >To: address@hidden
> >From: "Tom Baltzer" <address@hidden>
> >Subject: netCDF Decoders - Red Hat Linux - metar2nc errors
> >Organization: UCAR/Unidata
> >Keywords: 200405121544.i4CFibdf017469
>
> Institution: Unidata
> Package Version: 3.0.3
> Operating System: Red Hat Linux
> Hardware Information: AMD Dual 2.6 Gig processors, 1 Gig Memory
> Inquiry: Hi Robb,
>
> We're seen occasional metar2nc processes hanging out to dry on the LEAD test 
> bed machine.   When the log files are checked the messages indicate problems 
> with the output NetCDF file - for example:
>
> [ldm@lead1 surface]$ more metarLog.13430.log
> Opening data/pub/decoded/netcdf/surface/04051210_metar.nc with ncid 4
> Opening data/pub/decoded/netcdf/surface/04051209_metar.nc with ncid 5
> Opening data/pub/decoded/netcdf/surface/04051208_metar.nc with ncid 6
> Opening data/pub/decoded/netcdf/surface/04051207_metar.nc with ncid 7
> Closing data/pub/decoded/netcdf/surface/04051208_metar.nc with ncid 6, No 
> write for > 1200 Sec
> Closing data/pub/decoded/netcdf/surface/04051207_metar.nc with ncid 7, No 
> write for > 1200 Sec
> metar metadata file = 
> data/pub/decoded/netcdf/surface/../metadata/surface/metar/04051211_metar.meta
> Opening data/pub/decoded/netcdf/surface/04051211_metar.nc with ncid 6
> Closing data/pub/decoded/netcdf/surface/04051209_metar.nc with ncid 5, No 
> write for > 1200 Sec
> Opening data/pub/decoded/netcdf/surface/04051208_metar.nc with ncid 5
> Opening data/pub/decoded/netcdf/surface/04051209_metar.nc with ncid 7
> ncopen: filename "data/pub/decoded/netcdf/surface/04051011_metar.nc":

Not a netCDF file

Tom,

I've encountered this problem with gribtonc, the actual NetCDF file was
not created because there wasn't any space on the disk. Then the decoder
tries to write to the file which causes the following errors. Did you
actually look at 04051011_metar.nc file, is it corrupt?  What machine are
you working on now?

Robb...


> ncdimid: ncid -1: Not a netCDF id
> ncdiminq: ncid -1: Not a netCDF id
> Opening data/pub/decoded/netcdf/surface/04051011_metar.nc with ncid -1
> ncrecinq: ncid -1: Not a netCDF id
> NetCDF::recput result = -1
> Caught SIGTERM --shutting down
> Closing data/pub/decoded/netcdf/surface/04051011_metar.nc with ncid -1
> ncclose: ncid -1: Not a netCDF id
> Closing data/pub/decoded/netcdf/surface/04051210_metar.nc with ncid 4
> Closing data/pub/decoded/netcdf/surface/04051208_metar.nc with ncid 5
> Closing data/pub/decoded/netcdf/surface/04051211_metar.nc with ncid 6
> Closing data/pub/decoded/netcdf/surface/04051209_metar.nc with ncid 7
> [ldm@lead1 surface]$ more metarLog.11505.log
> Opening data/pub/decoded/netcdf/surface/04051209_metar.nc with ncid 4
> Opening data/pub/decoded/netcdf/surface/04051208_metar.nc with ncid 5
> Opening data/pub/decoded/netcdf/surface/04051207_metar.nc with ncid 6
> metar metadata file = 
> data/pub/decoded/netcdf/surface/../metadata/surface/metar/04051210_metar.meta
> Opening data/pub/decoded/netcdf/surface/04051210_metar.nc with ncid 7
> ncopen: filename "data/pub/decoded/netcdf/surface/04051010_metar.nc": Not a 
> netCDF file
> ncdimid: ncid -1: Not a netCDF id
> ncdiminq: ncid -1: Not a netCDF id
> Opening data/pub/decoded/netcdf/surface/04051010_metar.nc with ncid -1
> ncrecinq: ncid -1: Not a netCDF id
> NetCDF::recput result = -1
>
>
> By grepping for the netCDF file name in all log files, it appears that this 
> is occuring on the first open of the netCDF file (this assumes that process 
> id is sequential).  Additionally, the process does appear to still be "live" 
> after this - I was able to send it a sigterm which it caught and responded 
> to.  However, there was another (actually two at the time) metar2nc process 
> for the same month running.  The "older" processes are consuming large 
> amounts of CPU.
>
> Is it possible that on this type of error the metar2nc decoder is indicating 
> termination to pqact but then not terminating?
>
> Thanks,
> Tom.
>
>
>
> --
> NOTE: All email exchanges with Unidata User Support are recorded in the
> Unidata inquiry tracking system and then made publically available
> through the web.  If you do not want to have your interactions made
> available in this way, you must let us know in each email you send to us.
>
> ------- End of Forwarded Message
>

===============================================================================
Robb Kambic                                Unidata Program Center
Software Engineer III                      Univ. Corp for Atmospheric Research
address@hidden             WWW: http://www.unidata.ucar.edu/
===============================================================================