Re: 20000602: pqact entry?

Chris Herbster wrote:

> Anne,
>
> I was wondering if I might be able to stop by on the 14th, 15th or 16th to 
> work
> with you (or someone else) on this problem.  I'll be at the Unidata workshop
> next week but I added a couple of extra days to try and work on these 
> problems.
>
> Please reply to both emails as I might not see this one before I leave town on
> Wed. morning.
>
> I arrive in Denver at noon on Wed.  Could we meet on Wednesday afternoon or
> anytime Thursday?
>
> Thanks!!
>
> Chris H.
>
> ______________________________ Reply Separator 
> _________________________________
> Subject: Re: 20000602: pqact entry?
> Author:  Anne Wilson <address@hidden> at Internet-Mail
> Date:    6/7/00 4:20 PM
>
> Unidata Support wrote:
>
> > ------- Forwarded Message
> >
> > >To: Bob Lipschutz <address@hidden>
> > >cc: "Herbster, Chris" <address@hidden>,
> > >cc: address@hidden
> > >From: Chris Herbster <address@hidden>
> > >Subject: Re: pqact entry?
> > >Organization: Embry Riddle
> > >Keywords: 200006022237.e52MbaT04107 LDM pqexpire SIGALARM Linux
> >
> > Thanks Bob!
> >
> > Yup, our machine just isn't coping well.  I think it is under RAMed,
> > only 112 MB, which is low for Linux....  The LDM runs fine with a
> > trickle of data, but dies when I try to increase to flow.
> >
> > Right now I am only ingesting data from FSL and SUNY Albany.  SUNY
> > appears to be down, so only the ACARS data are entering the queue.
> >
> > "ldmadmin tail" (with pqact in debug mode) sees:
> >
> > Jun 02 22:20:46 opwx pqact[26675]: SIGALRM
> > Jun 02 22:20:46 opwx pqact[26675]: End of Queue
> > Jun 02 22:20:55 opwx pqact[26675]: SIGCONT
> > Jun 02 22:20:55 opwx pqact[26675]: 00a10cf2baa21126d902380af81c3186
> > 384732 20000602222033.057    PCWS 000  FSL.NetCDF.ACARS.QC.20001542220.*
> > Jun 02 22:20:55 opwx pqact[26675]: End of Queue
> > Jun 02 22:21:10 opwx pqact[26675]: SIGALRM
> > Jun 02 22:21:10 opwx pqact[26675]: End of Queue
> >
> > I assume that the "SIGALRM" comes from scanning the queue and not
> > finding anything to do, but that's a guess.
> >
> > "ldmadmin watch" sees:
> >
> >  Jun 02 22:20:56 pqutil:   384732 20000602222033.057    PCWS 000
> > FSL.NetCDF.ACARS.QC.20001542220.*
> >
> > No decoders are fired off and no data are being written to disk.  (Other
> > DDPLUS data do get written to disk when I add FSU to the request.)
> >
> > Here are the pqact.conf entries:
> >
> > # FSL ACARS data
> > PCWS    ^FSL\.NetCDF\.ACARS\.QC\.([0-9]+)\..*
> >         PIPE    -close  /usr/local/ldm/decoders/bin/dcacars -v -b 30
> >         -e GEMTBL=/home/gempak/NAWIPS-5.4/gempak5.4/tables
> >         -l logs/dcacars.log data/gempak/acars/YYYYMMDDHHNN_acars.gem
> >
> > PCWS    ^FSL\.NetCDF\.ACARS\.QC\.([0-9]+)\..*
> >         FILE  -close data/netcdf/acars/\1.acars
> >
> > The beginning of line characters are tabs.
> >
> > Any ideas?  I've Cc'd the support folks at Unidata to see if they can
> > help.
> >
> > Thanks again for the help!
> >
> > Chris H.
>
> Hi Chris,
>
> I wanted to make sure somebody got back to you on this...
>
> Your original message had lots of included messages in it, so I may not be 
> clear
> on
> the problem.  Is it simply that your ldm dies when you request more data?  If
> so,
> please send me more info from the log.
>
> 112Mb  does indeed sound like a small amount of RAM, which could definately
> affect
> the LDM as the product queue is a memory mapped file.
>
> The SIGALRM doesn't by itself indicate a problem -  quoting from our support
> database: "It simply means that no data got into the queue, or the pqact 
> process
> has
> not been told about it, and it alarmed out of its
> sleep to check on its own.  As you are seeing a End of Queue, then no data
> is getting into the queue."  But, your ldmadmin watch indicates that some 
> ACARS
> data
> did indeed make it into the queue, so I'm not sure what to make of this...
>
> Again, more info from the log might help.
>
> Anne
> --
> ***************************************************
> Anne Wilson                     UCAR Unidata Program
> address@hidden                  P.O. Box 3000
>                                   Boulder, CO  80307
> ----------------------------------------------------
> Unidata WWW server       http://www.unidata.ucar.edu/
> ****************************************************

Hi Chris,

Sure, I'm happy to meet with you.   So far, Wednesday afternoon or Thursday 
would
be fine.   Can you narrow that down some?

>Please reply to both emails as I might not see this one before I leave
>town on
>Wed. morning.

Is there more than one email?  I'm not sure what you mean.   I hope you get 
this!

Anne

--
***************************************************
Anne Wilson                     UCAR Unidata Program
address@hidden                  P.O. Box 3000
                                  Boulder, CO  80307
----------------------------------------------------
Unidata WWW server       http://www.unidata.ucar.edu/
****************************************************