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

Re: Motherlode FNEXRAD latency (fwd)

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 Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated.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: Re: Motherlode FNEXRAD latency (fwd)
  • Date: Mon, 21 Apr 2003 10:18:20 -0600 (MDT)
On Mon, 21 Apr 2003, Steve Emmerson wrote:

> Robb,
>
> Can you shed any light on this?
>
> --Steve
>
> ------- Forwarded Message
>
> Date: Mon, 21 Apr 2003 09:42:55 -0600 (MDT)
> From: Steve Chiswell <address@hidden>
> To: Steve Emmerson <address@hidden>
> cc: address@hidden
> Subject: Re: Motherlode FNEXRAD latency
>
> Steve,
>
> Not me. The floater stuff is Robb's domain. When the floater domain is 
> changed,
> it will go back in time and insert some past data. You probably need to see 
> how
> robb send's the change information and inserts the floaters.


Steve,

Chiz is correct.   When the new floater sites are selected from the web
page, the info is sent to motherlode via of ldmsend.  When motherlode
receives the product, it runs the mkfloat program using the new selection.
It kills the old floater program and starts a new one with  new selection.
It scans the queue from the begining looking for the sites, converting
them to FNEXRAD products. I believe the initial scan is causing the high
latency. Since the new queue is much larger now, I don't believe the
floater process needs to go back to the beginning of the queue for
products. I'm going to add the -o flags, (std ldm flag) to go back maybe 4
hours so users don't get such a backlog of data.

Robb...


>
> Chiz
>
> On Mon, 21 Apr 2003, Steve Emmerson wrote:
>
> > Hi,
> >
> > I'm trying to understand a 6000 s latency spike in Motherlode's FNEXRAD
> > plot at about 2003-04-21 12:30 UTC (today).
> >
> > Motherlode's logfile indicates that process 10926 was reaped by pqact(1)
> > due to a SIGKILL (-9) at 12:27:01. Shortly thereafter at 12:27:21, a
> > "floater" process started and processed a lot of old data. Did either of
> > you send the SIGKILL?
> >
> > --Steve
> >
>
> ------- End of Forwarded Message
>
>

===============================================================================
Robb Kambic                                Unidata Program Center
Software Engineer III                      Univ. Corp for Atmospheric Research
address@hidden             WWW: http://www.unidata.ucar.edu/
===============================================================================