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

[LDM #UEA-117531]: LDM data feed type recycle request



Wilson,

> As discussed this morning, hopefully below is what you need to determine
> the data feed type characteristics.
> 
> Thanks,
> Wilson Fong
> 301-713-0882 x123
> 
> David S Alden wrote:
> > Hi Wilson,
> >     Wow, you are a steely-eyed missle man!  That is REALLY good news!
> > The product ID we are currently sending is:
> >
> > wsr88dHHMM#.tbz2
> >
> > where:
> > HH = UTC hour 00 - 23
> > MM = UTC minute 00 - 59
> > #  = UTC sequence # 0 - 5
> >
> > These are bzip2 compressed tar archives containing all the NexRad88D Level
> > III RPG and SPG product sets from all 200 DOD, NWS, and FAA Doppler radars.
> >
> > Here are the configuration lines from ldmd.conf, pqact.conf and the 
> > arguments
> > to 'pqinsert' respectively (of course the FT15 will be changed to FT11):
> >
> > REQUEST     EXP|FT15        "^wsr88d.*\.tbz2" radar3src.nws.noaa.gov
> > FT15        ^(wsr88d.*) FILE -overwrite -close /var/opt/ldm/NEXRD3/incoming
> > pqinsert -fFT15 wsr88d$fid.tbz2

I feel bad, but I'm afraid that I have to disappoint you.

I talked your proposal over with my colleagues who know more about
the data than I do and the consensus is that the NNEXRAD feedtype
should be used rather than the FT15 feedtype.  The reasons for this
are the following:

1) The data products are NEXRAD Level III and that's
what the NNEXRAD feedtype means;

2) The 32 bits in the feedtype mask are a very scarce resource that
we must husband carefully until the next-generation LDM obviates this
limitation (we gave away a chunk of feedtypes early in the LDM's
history and that decision has come back to bite us---hard);

3) The pattern for the NEXRAD-III data-products you're creating
doesn't interfere with the existing NEXRAD-II data-products using
that feedtype (yours can be matched with "^wsr88d" and the other
data-products are matched with "^SDUS"); and

4) The GEMPAK and McIDAS pqact(1) configuration-files that ship
with those packages use "^SDUS" to select the current NEXRAD-III
data-products in the NNEXRAD stream, so no mixing can occur.

We do have some questions.  The files you're creating must be huge.
What's a typical size?  What's the meaning of the sequence number?
Who will you be sending these data-products to?  Will distribution
be restricted or might they wind up on the Unidata IDD?

Also, I notice that the timestamp in the current pattern has the
hour and minute, but not the date.  Based on our experience here
with data files, we recommend adding the date.  If the format
"YYYYMMDDhhmm" is used, then the files can be easily sorted into
temporal order (which often makes things easier).

Again, sorry to disappoint you.

Regards,
Steve Emmerson

Ticket Details
===================
Ticket ID: UEA-117531
Department: Support LDM
Priority: Normal
Status: Closed