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

Re: sirocco's requests to thelma

Robert Leche wrote:
> 
> Hello Anne,
> 
> Glad to hear about your trip to New Orleans was interesting and you
> had a great time to boot!. It's good place to have fun.
> 
> Your concern about network performance is timely. Recently, products
> have been added allowing our new Climatologist to  utilize more of
> Gempak's capability. Unfortunately, this has the effect of increasing
> loading on our computers and the network involved.
> 
> I looked at the ldmd.conf on Sirocco and made some changes. NEXRAD is
> now only feeding from Datoo ( our NOAA port). 

Great!  Thank you.


> After incorporating
> changes from your suggestions, a snippet of Sirocco's ldmd.conf shows:
> 
> request
> NEXRAD|IDS|UNIDATA|HRS|WSI|WMO|DDS|HDS|FNEXRAD|NPORT
> ".*"            datoo.srcc.lsu.edu
> #
> request
> DIFAX|FSL2|NMC2|MCIDAS
> ".*"            thelma.ucar.edu
> 
> If you have any more ideas or suggestions that will optimize the LDM
> traffic let me know.
> 

Please see below regarding NMC2.

> As it is, we are interested in the NOGAPS product line, so contacts to
> the proper agency would be helpful in requesting this service.
> NEXRAD is now set to feed from Datoo. As for NMC2, our Climatologist
> is requesting these products, but after reviewing the different
> products, we will trim back unneeded products. Is NMC2 available via
> the NOAAport?
> 

No, NMC2 is not available via NOAAPORT.  If your climatologist could
select a subset of products, that would be great. 

In the meantime, you should split up the NMC2 feed across multiple
connections.  Without a very high speed connection, it is difficult if
not impossible to get the entire CONDUIT feed relayed without falling
behind.  This could affect thelma as it keeps trying to feed you what
you're asking for when it doesn't have all the data available because
it's been expired from the queue.

You split the connection by making aliases in /etc/hosts, then
requesting from those aliases in ldmd.conf.  Here's a relevant email
from the archives: http://www.unidata.ucar.edu/glimpse/ldm/4924.  Please
give this a try.  Let me know if you have any questions about how to do
it.

To request the NOGAPS data, you should contact Phil Sharfstein at
address@hidden.  They will get you set up.


> On different but related topic. Sirocco's system performance.
> 
> In Sirocco's case, it appears that the biggest hit on Sirocco's system
> performance is coming from the decoding of the LDM streams. This has
> lead me to distribute the decoding load to other machines. My plan is
> to continue using Sirocco as LSU's IDD server. Sirocco will continue
> feed from Thelma (or Motherlode or where ever). Sirocco will continue
> feeding the down stream sites, but the decoding of the Gempak products
> will soon be handled by seistan.srcc.lsu.edu. Seistan will feed from
> Datoo and Sirocco.
> 

We have had the same experience, causing us to make motherlode a
decoding machine and thelma a relay only machine.  Your plan sounds
good.

> Lastly, under the category of just wanted to let you know:
> 
> The Southern Regional Climate Center is entering a new phase of
> operations. By expanding our product line to better service the needs
> of emergency officials at LSU's hurricane center, our local and state
> government, and the federal government agencies as well, our aim is
> help with hurricane emergencies and other storm  emergencies by
> offering a range products to visualize and present  current weather
> events.  Gempak has become our main  graphical analytical tool and
> will be used in times of emergencies to aid these officials in making
> informed decisions. Having said this, you can see why I am interested
> in "hardening" the LDM IDD process and to make it more robust and
> efficient.
> 
> We have been asked to install our Gempak tools in systems at the
> emergency center (OEP), and so we will. The plan is to add a LDM
> server on site. This server will allow the use of the spectrum of
> Gempak products during time of disasters.
> 

This is interesting.  Thank you for letting me know.  While we always
like to see people use the LDM, you do understand that *we do not
support operations*.  We only support education and research.  If the
ldm got wedged, finding immediate help could be a problem.  Same for the
IDD  - I hope your emergency managers are not depending on receiving
products via the IDD.  I feel especially strongly about this given our
recent spate of hardware problems that we are still coping with.  Right
now the problem is that thelma is not keeping an hour's worth of data in
the queue, hence my message to you yesterday.  Coincidentally, SSEC, one
of our other top tier sites, recently experienced a significant
failure.  I'm just glad our problems didn't coincide with theirs.   You
have a dish, so you can provide NOAAPORT products, at least.  But, the
IDD delivers many products that do not come from NOAAPORT.

People are free to use the LDM in any way they wish.  However, I would
advise them to use it only as a backup and not a primary method for
receiving products.

If you would like to talk further about this or any other issue, let me
know.

Anne


> Bob
> 
> 
> --
> 
> ----------------------------------------------------------------
> Robert Leche
> System Administrator
> Louisiana State University - Southern Regional Climate Center
> 260 Howe-Russell Building
> Baton Rouge, La. 70803
> address@hidden
> 225 578 5023
> ----------------------------------------------------------------
> 
> 

-- 
***************************************************
Anne Wilson                     UCAR Unidata Program            
address@hidden                 P.O. Box 3000
                                  Boulder, CO  80307
----------------------------------------------------
Unidata WWW server       http://www.unidata.ucar.edu/
****************************************************