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

20000829: Current Setup at University of Puerto Rico (cont.)

>From: McIDAS <address@hidden>
>Organization: University of Puerto Rico
>Keywords: 200008111642.e7BGgYN02913 McIDAS-X scripting

Karli,

>re: networking problems
>Apparetly the guy hasn't shown up :( I'll get Aamos to call him again. 

When I saw the connectivity to UPR get very good last Friday afternoon
and remain that way through the weekend, I had assumed (and hoped) that
the connectivity had improved dramatically (I could load NIDS images
from your site _very_ fast even at home overy my dial-up connection).
I noticed, however, that the connectivity had returned to the poor
category on Monday morning.

>Unfortunately(and fortunately), when we get the website up, we expect to
>get alot more traffic.  We hope it's internal traffic though so it
>shouldn't put that much of a strain on us.

It probably should be internal.  One quick comment on putting up displays
on web sites: you have to be aware of restrictions in data use:

o NLDN Lightning, NIDS

  Data Usage Guidelines
  http://www.unidata.ucar.edu/data/data_usage.html
  
o WMO Resolution 40 restrictions:
  http://www.nws.noaa.gov/oso/addata.html
  
I am not trying to discourage you; I just wanted to make sure that you
were aware of the above.

re: routing table update for NIDS will end when the decoding of NIDS products
into McIDAS AREA stops

>ok!

I turned off the decoding on Saturday.  I then removed the existing
AREA files in ~ldm/data/mcidasd AND the old FILEs in ~ldm/data/mcidasd/JUA.
By the way, the files under ~ldm/data/mcidasd/JUA were taking up some
200 MB of disk.  The update of the routing table for NIDS products is
now not active.

For reference, the NIDS products for JUA are being FILEed in the
~ldm/data/nexrad/NIDS/JUA directory.  I setup scouring in cron (invocation
of the shell script prune_nexrad.csh) to keep 60 of each kind of product
you are receiving from WSI Corporation.  This amounts to keeping between
5 and 10 hours of data online. The total time spanned by the products
will be determined by how often they are issued, which is a function
of the radar mode:

Mode          Update rate
-------------+--------------
clear air     10 minutes
precipitation  6 minutes
storm          5 minutes

Of course, this is also a function of the radar being operational :-)

>> IMGDISP RTNIDS/BREF1 LATLON=18:00 67:00 EU=BREF REFRESH='EG;MAP H'
>> BREF1, not BREF

>Oh! So that was it! And I thought it was this big huge problem with this
>complex solution... 

Nope, just a typo.

>> IMGDISP RTNIDS/VEL1 LATLON=18:00 67:00 EU=VEL REFRESH='EG;MAP H'

>cool. They all work now!!! 

Excellent.

>However the UNIDATA Menu option for radar is acting funny: not only is
>it printing the command wrong (it's missing a single quote at the end,

Hmm... this problem would be in the file imagloop.gui (gets installed
in ~mcidas/bin).  I'll have to check that one out.

>although I think I haven't tinkered with that file in ages, and then a
>perl script error afterwards:
>--------------------------------------------------------------------
>IMGDISP RTNIDS/BREF1 ID=JUA ALL=11 11 STA=JUA MAG=1 1 EU=BREF SF=YES
>REFRESH='MAP SAT 5 COUNTY=ALL GRA=(GRA); MAP VH 1 GRA=(GRA) WID=2;BAR
>ORI=VER C SU=X

From this output, the thing that is missing is the trailing single quote.
I just edited ~mcidas/bin/imagloop.gui on breeze and corrected this
problem.  I will correct this in my distribution.

>mcrun: Mismatched quotes: IMGDISP RTNIDS/BREF1 ID=JUA...
>Error in startup script: can't read "loobp": no such
>variable                        
>    while
>executing                                                                  
>"upcRunCommand "LB $loobp
>$loope""                                                   
>    invoked from
>within                                                              
>"if { $type == "NIDS" && $numradar == 1 && $loptn != "UD" }
>{                        
>      set nidsid [string range [lindex $radarinfo 0] 0
>2]                            
>      upcRunCommand "EG
>$bframe..."                                                  
>    (file "/unidata/mcidas/bin/imagloop.gui" line
>277)                               

This is a Tcl/Tk scripting error.

>which I thought was rather interesting... all users return this error on
>all NIDS products called from the menu.

Please give the Fkey menu action another try from a variety of accounts.

re: station ID use as value for STA= keyword on IMGDISP

>Oh so that was it! Yes, these sation codes do seem very familiar,
>especially TJSJ and I remember that being the code we used for Puerto
>Rico.

re: Please retry the commands above.
>Thank you they work beautifully now!

OK, good.

re: MCGUI is not ADDEized

>ok so right now it's the MCGUI the only one I'm concerned about, but for
>now we can do without it

OK.

re: MCGUI really should have been MCBIN

>oh...!!!

re: need to set McIDAS environment variables in shell scripts

>I'm sorry I actually meant to say that I "couldn't" do scripts without
>it.  If I hadn't defined the variable in the scripts, it couldn't get
>remote ADDE imagery.

Right.  The reason for this is that you need to tell the script where
to look to find the DATALOCations for the stuff you want to use in your
script.  The concept here is that McIDAS routines must have certain
things defined for them to work.  When one runs McIDAS commands out of
a script, the need stays the same.  The things that are needed to be
defined are the MCDATA, MCPATH, MCGUI, MCTABLE_READ, and MCTABLE_WRITE
environment variables (MCTABLE_WRITE is typically not needed, but it
does not harm to include it).

re: 5 seconds for a packet to get somewhere is outrageous!  We get _much_
better connectivity to Australia and Antarctica!!

>Wow! maybe we should get connected to an ISP from there! jeje 
>Our main problem isn't just bandwith, but more like a population
>problem: our bandwith is pretty decent: during the weekends and at
>2:00am.

It was spectacular this past weekend.

>The rest of the time we have to share it with thousands upon
>thousands of students surfing the web, checking their email and
>listening to online radio stations :(  Not that what their doing is bad,

It is not bad if the school has enough bandwidth to handle it.

>but we don't have a dedicated line outside. And what's worse is cutting
>through all of the red tape... but we'll see.

OK, good luck.

re: BATCH LSSERVE.BAT

>Yup, This works!

re: sorry for the typos

>I am actually the LUCKY sap that's been getting all this help from you! 
>Without it I wouldn't have figured out 75% of the things going on!  We
>all make typos.

Thanks.  I felt bad about leading you down at least two blind paths 
because of simple typos.

>I also have a couple of extra questions for you:
>
>1. I will be transferring the web page over to a Linux Mandrake Machine
>(much faster than this, so it'll take away alot of the overhead!) and
>we'll need a copy of McIDAS to be running in it.  I was wondering If we
>needed to buy another license to do this or if the license we have will
>do?

No, one license fits all McIDAS installations on a campus (not just a
department).

>Since I will be installing McIDAS from scratch, if you have any
>important compilation issues please let me know!

Please read the Notes and Warnings section of the McIDAS online
documents and pay particular attention to the sections related to
gcc/f2c/g77 (you can NOT use g77!) and Linux.   Also, we have not
compiled on Mandrake Linux, so there is a possibility that some things
may be different.

>2. Speaking of web page design, I was wondering how good of an idea
>would it be to have a web-based front end for McIDAS?

This has been tried before with varying success.

>What I mean is a
>set of perl or cgi scripts that would give you almost the same
>functionality as the McIDAS Menu but from the web.  Or are there such
>front-ends available?  

Hmm...  Jim Koermer at Plymouth State College put together a CGI front
end for putting up GINI imagery using McIDAS-X.  Jim's site uses WXP,
GEMPAK and some McIDAS; it is interesting to look at to get some
ideas:

Plymouth Home Page
http://www.plymouth.edu

Click on the Weather Center link and browse around.  The McIDAS GINI
displays can be found in:

Make your Own...
http://cyclone.plymouth.edu/make.html

Some other sites to check out can be found under:

Unidata McIDAS Home Page
http://www.unidata.ucar.edu/packages/mcidas

  Examples of McIDAS products

The listing is short and not up to date, but it will give you some
sites to look at.

>3. When I am done with the page would you be willing to take a look? :)
>Although I know your connection would be rather slow...

Sure.

>Thank you!

You are welcome.

Tom