Re: metar2nc behaviour

On Mon, 26 Jul 2004, Chris Peters wrote:

> Hello,
>
> I've noticed some unusual behaviour with metar2nc.  Here are the relevant
> details:
>
> ldm version - 6.0.15
>
> decoder version - 3.0.4
>
> OS: Red hat linux kernel 2.4.21-9
>
> Behaviour:   the decoder is working nominally but I've noticed that upon
> starting ldm, a decoder process (metar2nc) starts up and stays alive for days.
> Right now I can see this in the process table:
>
>
> ps auxw | grep metar2nc
> ldm       5008  0.3  0.9 83924 37272 ?       S    Jul23  14:07
> /usr/local/bin/perl decoders/metar2nc etc/metar.cdl data/decoded/metar 200407
>
>
> My previous experience with ldm and metar2nc on other systems is that this 
> does
> not happen.  Decoder processes can be spawned by ldm and stay alive for a 
> while,
> on the order of 10-20 minutes, but not the same single process staying alive 
> for
> days.  Is this a new feature or perhaps a bug?
>
>

Chris,

Actually the metar2nc decoder is written to stay alive for a whole month
as long as some data arrives every 20 minutes. The idea was to decrease
the number of decoder start ups, most expensive part of the perl decoders.
The decoders open/sync/close  netCDF files as needed and do report
checking internally.  The metarLog shows the status of netCDF file and
metar.lst shows what reports were processed in the last 24 hours on shut
down so they are not reentered into the netCDF files.

robb...






> Thanks,
>
> Chris Peters
> =======================================================
> Christopher Peters
> Senior Wx Systems Developer - The Weather Channel
> address@hidden
> Phone:  770-226-2039
>
>
>

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