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

[Datastream #IZJ-689237]: Additional Datafeeds

Hi Jeff,

re:
> Well, my data levels on Whistler are creeping back up, but they're well below 
> 100%.

OK.

> We're still seeing the "holes" in our data.

In order to pinpoint exactly why you have gaps in your decoded data, it would 
be useful
to get a comprehensive list of the products that are missing.  This will help 
us determine
if the missing products are ones from a particular IDD datastream (e.g., 
CONDUIT, HDS, etc.)
or if they are all over the board.

> I responded to a message from a Gempak-support person (Michael James?) about 
> what's happening.
> I haven't heard back yet, so I've been searching for myself.

Michael is working as hard as he can to put together the GEMPAK 5.11.4 release.

> Basically, what we're running into is the same thing described here -
> http://www.unidata.ucar.edu/support/help/MailArchives/idd/msg04127.html,

The advice in this exchange is _exactly_ the subject I broached with you at the 
beginning
of our exchanges:  your LDM queue was much, much too small to store all of the 
data you
were/are ingesting long enough for the decoders to be able to process it out of 
your
LDM queue.

> but I'm not sure how to go about tracking down why it's doing this.

I commented in an earlier email that the latencies that are being seen for 
CONDUIT
data products may be an indication that you are not receiving all of that data.
I further commented that the same situation does not appear to hold for the data
in the HDS datstream.  My comment above about documenting the list of missing
products (the holes in the data) will help pinpoint which datastream(s) are not 
being
fully decoded.  This may also determine the next steps to take.

One thing that I would recommend in this testing phase is to disable the 
ingestion
of all CONDUIT data.  This would completely free-up network resources and lessen
the load on the machine enough to determine if other decoding is working 
correctly.

> Is this due to too much data coming in still?  Or the latency problems?

I am not sure.  Turning off CONDUIT would test the first possibility and 
possibly
the second.

Cheers,

Tom
--
****************************************************************************
Unidata User Support                                    UCAR Unidata Program
(303) 497-8642                                                 P.O. Box 3000
address@hidden                                   Boulder, CO 80307
----------------------------------------------------------------------------
Unidata HomePage                       http://www.unidata.ucar.edu
****************************************************************************


Ticket Details
===================
Ticket ID: IZJ-689237
Department: Support IDD
Priority: Normal
Status: Closed