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

[Support #VXV-940104]: Re: [conduit] missing and incomplete GFS grids on CONDUIT feed

Hi David and Pete,

Pete wrote:
> Are you using a split feed to get conduit data or a single request line? 
> We've found
> that splitting the feed into multiple request lines can have a huge impact on 
> getting
> CONDUIT data with much smaller lags than you are seeing on freshair2.
> 
> We're using a 10-way split and our lags have recently been less then 30-60s 
> even with
> the additional data from the GFS FV3 update.

I believe that Pete's comment may be in the right neighborhood for figuring out 
what
may be going on.

Compare rtstats volume plots for two UWashington machines that are REQUESTing
CONDUIT with our machine that REQUESTs CONDUIT from NCEP and then relays
the data to the accumulators for our top level IDD relay clusters:

freshair1.atmos.washington.edu
http://rtstats.unidata.ucar.edu/cgi-bin/rtstats/iddstats_vol_nc?CONDUIT+freshair1.atmos.washington.edu

freshair2.atmos.washington.edu
http://rtstats.unidata.ucar.edu/cgi-bin/rtstats/iddstats_vol_nc?CONDUIT+freshair2.atmos.washington.edu

conduit.unidata.ucar.edu
http://rtstats.unidata.ucar.edu/cgi-bin/rtstats/iddstats_vol_nc?CONDUIT+conduit.unidata.ucar.edu

A visual comparison in the receipt volumes between the machines freshair1 and 
conduit
indicates that the volumes are essentially the same.  The same comparison 
between
freshair2 and either/both freshair1 and conduit shows that freshair2 is 
receiving
less volume.  This lead me to compare receipt latencies between freshair1 and
freshair2:

freshair1.atmos.washington.edu
http://rtstats.unidata.ucar.edu/cgi-bin/rtstats/iddstats_nc?CONDUIT+freshair1.atmos.washington.edu

freshair2.atmos.washington.edu
http://rtstats.unidata.ucar.edu/cgi-bin/rtstats/iddstats_nc?CONDUIT+freshair2.atmos.washington.edu

This comparison shows that the receipt latencies on freshair1 are, for the most 
part,
very low, while the latencies on freshair2 are periodically high to the point 
that
products are likely to be missed.

Now the question is why are the latencies on freshair2 so much different
than the latencies on freshair1?  Pete's questioning of how CONDUIT is
being REQUESTed on each is the first place to start in this investigation.

I expect that the most likely answer will be that there are more splits
to the feed REQUEST on freshair1 than on freshair2.

David:

What are the REQUEST line(s) for CONDUIT both freshair1 and
freshair2?

Also, what are the LDM queue sizes on freshair1 and freshair2?

Comment that is somewhat aside:

It may be a good idea to also REQUEST CONDUIT from freshair1 on
freshair2 and visa versa as this will cause products received
on one earlier than the other to be sent to the other.  We
cross REQUEST on all of the accumulators (front ends) for our
IDD relay clusters as a matter of course.

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: VXV-940104
Department: Support CONDUIT
Priority: Normal
Status: Open
===================
NOTE: All email exchanges with Unidata User Support are recorded in the Unidata 
inquiry tracking system and then made publicly 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.