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

20030512: Delayed data at U Washington

>From: Harry Edmon <address@hidden>
>Organization: University of Washington
>Keywords: 200305121626.h4CGQ07U025551 IDD

Hi Harry,

>sunny is has been over an hour behind on its data feed since Saturday night.  
>I am trying to figure out why.

We saw the latencies on sunny drop to near zero right after 18 UTC.
Did you find something at UW?  If so, would you share the information
with us?  Thanks.

On another note, I see that your data request from sunny to thelma is:

May 12 17:40:47 thelma.ucar.edu sunny(feed)[13756]: up6.c:299: Starting 
Up(6.0.10): 20030512164338.968 TS_ENDT {{NNEXRAD|DIFAX|FNEXRAD|FSL2|WMO,  ".*"}}

i.e., all of NNEXRAD|DIFAX|FNEXRAD|FSL2|WMO on a single rpc.ldmd

If you start to see latencies build, it would probably be a good idea
to split your data requests.  What I have in mind is something lik:

request NNEXRAD           ".*" thelma.ucar.edu
request DIFAX|FNEXRAD|HRS ".*" thelma.ucar.edu
request IDS|DDPLUS|DIFAX  ".*" thelma.ucar.edu

The idea here is to spit the total request into two or three separate
requests that partition the numbers (not volume) of product on each
feed more-or-less evenly.  This may not generally be necessary at UW
since your IDD reception is typically quite good.

Just a thought...

Tom