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

20030805: differences between LDM 5 and 6

Karen,

>Date: Tue, 05 Aug 2003 15:23:18 -0500
>From: "Karen Cooper" <address@hidden>
>Organization: CIMMS at NSSL
>To: Steve Emmerson <address@hidden>
>Subject: Re: 20030805: differences between LDM 5 and 6

The above message contained the following:

> I can't give you access to the originating machine, because it's not 
> mine.  It's one of Carl Sinclair's CRAFT machines, and he and I have 
> been working together on the problem.

The CRAFT data-products are known to have wildly inaccurate
product-creation times.  See, for example

    
http://www.unidata.ucar.edu/cgi-bin/rtstats/iddstats_nc?CRAFT+thelma.ucar.edu

The large positive and negative latencies are due solely to inaccurate
product-creation times.

> The reason we suspect LDM (or tcp/rpc) is that we are missing entire 
> packets.  For instance, yesterday I got a volume scan that had only 300 
> radials for one of it's elevations.  The final packet of data that 
> should have contained the 60+ radials left in the elevation was 
> apparently never received (or didn't get written to file by pqact, or 
> something...).  However, it is included in the LDM output file on the 
> machine that directly ingest the raw data.  None of the downstream 
> machines received it.

Look at the logfile of the downstream LDM for that time period.  Was
there a reconnection?

Regards,
Steve Emmerson
LDM Developer