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

[Support #FAC-765998]: [conduit] Slow connection/big lag from conduit.ncep.noaa.gov

Hi Pete,

re:
> That is interesting..  I ran an ldmping to idd.unidata.ucar.edu and also
> to conduit.ncep.noaa.gov, and the idd.unidata ones were consistantly 0.1
> second or less.

OK, but the ldmping packets are very small, so small numbers do not
necessarily imply that a lot of data can be transferred.  The classic case
of this is when there is volume based packet shaping (artificial volume
limiting) being done on the upstream side.

re:
> The ones to conduit.ncep.noaa.gov were mostly 0.09 but every 6th one or
> so was 0.35, and every once in a while there was a 4 second one.
> Unfortunatlely I closed the window that had them in it so I can't show
> exactly what I saw.. Right now ldmping conduit.ncep.noaa.gov is not
> returning pings..

ldmping is good for showing that an upstream LDM is running.  I would not
rely on it meaning that the latencies for products being relayed by
the upstream will be low.

re:
> Our conduit feeds are split.  There are 5 feeds, currently
> 
> REQUEST CONDUIT "[05]$" idd.unidata.ucar.edu
> REQUEST CONDUIT "[16]$" idd.unidata.ucar.edu
> REQUEST CONDUIT "[27]$" idd.unidata.ucar.edu
> REQUEST CONDUIT "[38]$" idd.unidata.ucar.edu
> REQUEST CONDUIT "[49]$" idd.unidata.ucar.edu

OK.  You should be redundantly feeding from both NCEP upstream sites,
conduit.ncep.noaa.gov and ncepldm4.woc.noaa.gov.  You can, of course,
also continue your redundant feed REQUESTs to idd.unidata.ucar.edu.

NB: if you REQUEST the same feed(s) from more than one upstream, you
should make sure that your REQUESTs are identical, or, at least, the
union of the REQUESTs is identical.

re: 
> I used to have them split 09, 18, 27, 36, 45 but it seemed to me to go
> every 5th product made more sense. Maybe not?

This shouldn't make any difference as the sequence numbers are probably
distributed uniformly (meaning just as many end with a '0' as end with
'1', '2', '3', etc.).

re:
> Will let you know if I hear any more from the NCEP folks..

I am in contact with Carissa Klemmer of NCEP.  They are eager to look
into the problem, but they wanted to understand the real-time stats
pages better before proceeding.

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: FAC-765998
Department: Support CONDUIT
Priority: Normal
Status: Closed