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

[LDM #VAP-368514]: Primary/alternate switching mixes in old data with new

This archive contains answers to questions sent to Unidata support through mid-2025. Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated. For the most up-to-date information on the use of NSF Unidata software and data services, please consult the Software Documentation first.

  • Subject: [LDM #VAP-368514]: Primary/alternate switching mixes in old data with new
  • Date: Mon, 23 Nov 2009 11:04:19 -0700
Hi Art,

> On the latency chart for CONDUIT data for system iddrs3.meteo.psu.edu at:
> 
> http://www.unidata.ucar.edu/cgi-bin/rtstats/iddstats_nc?CONDUIT+iddrs3.meteo.psu.edu
> 
> I've been investigating the latency spikes.

Some of those spikes are pretty big (over 6000 seconds).  What's the maximum 
data-product latency that you allow on your LDM system?  Do you invoke 
"ldmadmin start" with a "-m maxLatency" option?  What is the output of the 
command "ldmadmin config | grep latency"?

>  System iddrs3 pulls CONDUIT
> .* from idd-ingest.meteo.psu.edu primary and idd.cise-nsf.gov secondary.
> idd-ingest.meteo.psu.edu pulls CONDUIT .* from only idd.cise-nsf.gov.  It
> appears that when these spikes occur, reception switches from
> idd-ingest.meteo.psu.edu to idd.cise-nsf.gov as primary and, for some
> reason, starts feeding very old data interspersed with new data.  I don't
> believe we're losing data, just that old data is being resent (I think).
> It doesn't do this every time it switches, just sometimes.  I don't
> believe this occurs when it switches to idd-ingest.meteo.psu.edu as
> primary (at least not that I've observed) which would agree with the
> graphic.  Any ideas what could be causing this?

I have some ideas, but they need to be verified.  Please send me the answer to 
my question and also the output from the following command:

    fgrep '[<pid>]' $HOME/logs/ldmd.log | fgrep 'LDM-6 desired product-class'

where <pid> is the process identifier of the downstream LDM that receiving 
CONDUIT data from "idd.cise-nsf.gov".

> FYI, I upgraded to LDM V 6.8.1 yesterday (Nov 20th) at ~16:30Z to see if
> that would help.  Prior to that I was running LDM V 6.7.0.
> 
> 
> Thanks...
> 
> Art
> 
> Arthur A. Person
> Research Assistant, System Administrator
> Penn State Department of Meteorology
> email:  address@hidden, phone:  814-863-1563

Regards,
Steve Emmerson

Ticket Details
===================
Ticket ID: VAP-368514
Department: Support LDM
Priority: Normal
Status: Closed