[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: ldm problems and separating feed types..

Thanks Stonie..I appreciate it very much.

-----Original Message-----
From: Stonie R. Cooper [mailto:address@hidden]
Sent: Monday, August 06, 2001 12:42 PM
To: Robert Mullenax
Cc: Cannon, Declan; Wells, Timothy; address@hidden
Subject: Re: ldm problems and separating feed types..


Robert,

I'll take this off-line due to the obvious ramifications that it may appear 
as self-promotion - and that is not my intent.

I'm assuming you have either SDI straight from SSEC, or from the now defunct

Alden.  We are only familiar with them in-as-much as we have been brought in

to replace such systems with ours - and LDM is one of the reasons why.

I'll copy Declan, also, as I know he has seen queries in mail lists about
the 
same problem as it occurs with Unisys's system - i.e. the pqing simply dies 
with a "resource unavailable" error.  You can ask Dan Vietor, also, but I'm 
pretty sure that they inject data into the LDM pretty much the same way as 
the Alden/SSEC system.

I don't know that there is a resolution around this, as of yet.

From Unidata/UCAR mail archives, my exposure to the SDI system was that
Alden 
was trying to make this look exactly like a FOS feed; NRS data, though, as 
you know, is not FOS - and does not have the inherit distinction of 
"feedtype".  If the SDI software allows you establish fifos beyond the ones 
you have listed, you may be able to segment the data further - but from your

list below, all messages coming in from the /tmp/jmb.fifo.4 are bitwise
or'ed 
as both IDS and DDPLUS - thus you can't separate them - all messages have 
both feedtypes.

If SDI allows you to segment further, make the data you want to be only IDS 
or only DDPLUS go to separate fifo's and add pqing entries for those, and 
changing the feedtype parm passed to be either just IDS or DDPLUS:

exec "pqing -f ids|ddplus /tmp/jmb.fifo.4"
exec "pqing -f hrs /tmp/jmb.fifo.5"
exec "pqing -f ids /tmp/jmb.fifo.6"
exec "pqing -f ddplus /tmp/jmb.fifo.7"

Again, thought, this assumes you can do some product definition upstream 
withint the SDI software.  Unidata has a great reference on-line about how 
feedtypes are supposed to be defined - with links to NOAA/WMO where it 
applies:

http://www.unidata.ucar.edu/packages/ldm/feedtypes/index.html

And, yes, we do our LDM implementation completely different (for the above 
reasons), and then set our NRS to stream data to the LDM with feedtypes to 
match the current IDD (as much as we can).

Stonie

On Monday 06 August 2001 16:36, you wrote:
> We have an SDI NOAAPORT box that is ingesting text/binary and the GINI sat
> imagery.  We have an ldm
> running on the machine that is using the ingest from the fifos:
>
> exec "pqing -f ids|ddplus /tmp/jmb.fifo.4"
> exec "pqing -f hrs /tmp/jmb.fifo.5"
>
> We get occasional periods where the ldm just quits receiving data (the
> McIDAS socket connection keeps working)
> and an error says:    pqinsert..file name..resource unavailable.  Has
> anyone else seen this before?  The ldm
> has to be stopped and restarted.
>
> Also, is there a way we can split the feed type so the IDS is separate.
On
> our downstream machines we'd
> like to file it separately and right now, even with separate DDPLUS and
IDS
> pqact entries..the same data
> is in each stream.
>
> Thanks,
> Robert Mullenax
>
> Robert Mullenax
> Weather Systems Administrator
> Universal Weather and Aviation

-- 
Stonie R. Cooper,
Science Officer
Planetary Data, Incorporated
3495 Liberty Road
Villa Rica, Georgia  30180
ph. (770) 456-0700; pg. (888) 974-5017; fx. (770) 459-0016

From address@hidden Mon Aug  6 11:41:52 2001
Received: from kiwi.pyro.net (address@hidden [207.7.10.2])
        by unidata.ucar.edu (UCAR/Unidata) with SMTP id f76Hfp118289
        for <address@hidden>; Mon, 6 Aug 2001 11:41:52 -0600 (MDT)
Keywords: 200108061741.f76Hfp118289
Received: (qmail 6270 invoked from network); 6 Aug 2001 17:56:57 -0000
Received: from unknown (HELO helium.fieldlab.planetarydata.com) 
(216.126.169.251)
  by mail.pyro.net with SMTP; 6 Aug 2001 17:56:57 -0000
From: "Stonie R. Cooper" <address@hidden>
Reply-To: address@hidden
Organization: Planetary Data, Incorporated
Date: Mon, 6 Aug 2001 17:41:42 +0000
X-Mailer: KMail [version 1.1.99]
Content-Type: text/plain;
  charset="iso-8859-1"
To: Robert Mullenax <address@hidden>
References: <address@hidden>
In-Reply-To: <address@hidden>
Subject: Re: ldm problems and separating feed types..
Cc: "Cannon, Declan" <address@hidden>,
   "Wells, Timothy" <address@hidden>,
   address@hidden
MIME-Version: 1.0
Message-Id: <address@hidden>
Content-Transfer-Encoding: 8bit

Robert,

I'll take this off-line due to the obvious ramifications that it may appear 
as self-promotion - and that is not my intent.

I'm assuming you have either SDI straight from SSEC, or from the now defunct 
Alden.  We are only familiar with them in-as-much as we have been brought in 
to replace such systems with ours - and LDM is one of the reasons why.

I'll copy Declan, also, as I know he has seen queries in mail lists about the 
same problem as it occurs with Unisys's system - i.e. the pqing simply dies 
with a "resource unavailable" error.  You can ask Dan Vietor, also, but I'm 
pretty sure that they inject data into the LDM pretty much the same way as 
the Alden/SSEC system.

I don't know that there is a resolution around this, as of yet.

From Unidata/UCAR mail archives, my exposure to the SDI system was that Alden 
was trying to make this look exactly like a FOS feed; NRS data, though, as 
you know, is not FOS - and does not have the inherit distinction of 
"feedtype".  If the SDI software allows you establish fifos beyond the ones 
you have listed, you may be able to segment the data further - but from your 
list below, all messages coming in from the /tmp/jmb.fifo.4 are bitwise or'ed 
as both IDS and DDPLUS - thus you can't separate them - all messages have 
both feedtypes.

If SDI allows you to segment further, make the data you want to be only IDS 
or only DDPLUS go to separate fifo's and add pqing entries for those, and 
changing the feedtype parm passed to be either just IDS or DDPLUS:

exec "pqing -f ids|ddplus /tmp/jmb.fifo.4"
exec "pqing -f hrs /tmp/jmb.fifo.5"
exec "pqing -f ids /tmp/jmb.fifo.6"
exec "pqing -f ddplus /tmp/jmb.fifo.7"

Again, thought, this assumes you can do some product definition upstream 
withint the SDI software.  Unidata has a great reference on-line about how 
feedtypes are supposed to be defined - with links to NOAA/WMO where it 
applies:

http://www.unidata.ucar.edu/packages/ldm/feedtypes/index.html

And, yes, we do our LDM implementation completely different (for the above 
reasons), and then set our NRS to stream data to the LDM with feedtypes to 
match the current IDD (as much as we can).

Stonie

On Monday 06 August 2001 16:36, you wrote:
> We have an SDI NOAAPORT box that is ingesting text/binary and the GINI sat
> imagery.  We have an ldm
> running on the machine that is using the ingest from the fifos:
>
> exec "pqing -f ids|ddplus /tmp/jmb.fifo.4"
> exec "pqing -f hrs /tmp/jmb.fifo.5"
>
> We get occasional periods where the ldm just quits receiving data (the
> McIDAS socket connection keeps working)
> and an error says:    pqinsert..file name..resource unavailable.  Has
> anyone else seen this before?  The ldm
> has to be stopped and restarted.
>
> Also, is there a way we can split the feed type so the IDS is separate.  On
> our downstream machines we'd
> like to file it separately and right now, even with separate DDPLUS and IDS
> pqact entries..the same data
> is in each stream.
>
> Thanks,
> Robert Mullenax
>
> Robert Mullenax
> Weather Systems Administrator
> Universal Weather and Aviation

-- 
Stonie R. Cooper,
Science Officer
Planetary Data, Incorporated
3495 Liberty Road
Villa Rica, Georgia  30180
ph. (770) 456-0700; pg. (888) 974-5017; fx. (770) 459-0016