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

19990218: gpfront...



Maureen,

If you are drawing your images to the console, its probably a good idea
to launch ntl once on the console, and leave it running. This should really
help with color map problems since all gempak programs will share the same
colormap that ntl has allocated.

The other alternative to draving maps on the console is to create a virtual 
screen using the X11R6 program Xvfb (virtual frame buffer), however, if you
don't have that program, keeping ntl on your desktop is probably the best 
thing. 

The most frequent problems with hung gempak programs is having conflicts with
obtaining the message queues. If you find that your programs have gotten
hung up, check for old copies of gplt as well as device drivers (gf/xw etc)
in your process tables. After killing those programs, you will generally have to
clear out the message queues (using ipcs to list the queues, and ipcrm to
remove them). Usually a program gets hung when either it can't connect to
the console for some reason, or there is not enough memory to fire up the
gplt program, or several gempak program have a contention for the same
message quueue (possible if 2 crons start up at nearly the same time
since the both check to see if a queue is used, and both try to
start gplt with the same queue id).

If you have crons that tend to launch gempak programs while others are still 
running, 
you may want to arrange a semaphore system in your scripts to prevent them from
stepping on each other. I use semaphores with the automatic generation of NIDS 
images
I have here running from the LDM, since when the LDM first starts up, there is
a tendency to get a bunch of backlogged images all at once which can lead
to race conditions for lots of programs being fired up.

Steve Chiswell
Unidata User Support



>From: Maureen Ballard <address@hidden>
>Organization: UK Ag Weather Center
>Keywords: 199902182100.OAA16345

>Steve,
>
>Thanks for the info on the changes to gpfront - I will do that in just a minut
> e.
>
>> The topo images can be displayed an any mapping programs so long as
>> proj=sat. Eg, the topo images are AREA files just like the McIDAS
>> satellite images. Since they are the same as satellite images, they
>> require the same number of colors as a GOES-10 image would- 128 colors.
>> Define the topo file with satfil=... and set clear=y to display the
>> image with the fronts/contours etc overlayed.
>>
>> If you are having trouble with color allocation, launch ntl first
>> if you don't already have that running. Ntl will either allocate all the col
> ors
>> you will need, or tell you howmany you need. If you don't have enough
>> colors, try launching ntl with the -s option to reduce the number of
>> satellite colors- for instance, ntl -s 96 or ntl -s 64.
>>
>
>Okay, now here's a question for you. We run all of our maps through crontab en
> tries
>and I have never used ntl. Should I have the crontab launch ntl at the start o
> f
>each map or should we launch ntl once and leave it running?
>
>Thanks...
>Maureen
>
>> Steve Chiswell
>>
>> >From: Maureen Ballard <address@hidden>
>> >Organization: UK Ag Weather Center
>> >Keywords: 199902182012.NAA15031
>>
>> >Steve,
>> >
>> >I saw the messages posted about the weird looking warm fronts - we were
>> >beginning to wonder! Hope NCEP gets that all straightened out soon!
>> >
>> >Any way, I was wondering if the fronts can be placed on a topo map. I
>> >tried a couple times and got various errors (partially because there
>> >were problems allocating colors). I did a reboot of the system since I
>> >couldn't figure out what processes had hung (seem to have that problem
>> >quite a bit here) and then tried but still can't get it to work. For
>> >that matter I can't seem to get it to work with gdcntr either. Since the
>> >topo maps are so colorful, are they a resource hog when it come to color
>> >allocation?
>> >
>> >Any help would be greatly appreciated.
>> >
>> >Thanks!
>> >
>> >Maureen
>> >
>> >--
>> >========================================================================
>> >
>> >Maureen Moore Ballard                   address@hidden
>> >Staff Meteorologist                           ph:  606-257-3000ext244
>> >Ag. Weather Center                         fax: 606-257-5671
>> >243 Ag. Engineering Bldg
>> >Dept. of Biosystems and Ag. Engr.
>> >University of Kentucky
>> >Lexington, KY 40546-0276
>> >HOMEPAGE   http://wwwagwx.ca.uky.edu
>> >
>> >You don't stop laughing because you grow old;
>> >you grow old because you stop laughing.
>> >=========================================================================
>> >
>> >
>> >
>>
>> ****************************************************************************
>> Unidata User Support                                    UCAR Unidata Program
>> (303)497-8644                                                  P.O. Box 3000
>> address@hidden                                   Boulder, CO 80307
>> ----------------------------------------------------------------------------
>> Unidata WWW Service                        http://www.unidata.ucar.edu/     
>> ****************************************************************************
>
>
>
>--
>========================================================================
>Maureen Moore Ballard                   address@hidden
>Staff Meteorologist                           ph:  606-257-3000ext244
>Ag. Weather Center                         fax: 606-257-5671
>243 Ag. Engineering Bldg
>Dept. of Biosystems and Ag. Engr.
>University of Kentucky
>Lexington, KY 40546-0276
>HOMEPAGE   http://wwwagwx.ca.uky.edu
>
>You don't stop laughing because you grow old;
>you grow old because you stop laughing.
>=========================================================================
>
>