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

[McIDAS #EUT-864314]: mcidas 2008 build problems



Hi Bill,

Thanks for the update.  Please read through to the end of this reply to see a) 
I made
a new addendum release for Unidata McIDAS-X/-XCD v2008 and b) we are 
experiencing some
problems on adde.ucar.edu that are a result of what we believe to be your 
access of
U and V wind components from the RTGRIBS/GFS-ALL dataset.

re:
> Just an update with mostly success to report:
> 
> 1) After considering your comment about fonts not being installed I trashed 
> around
>    and eventually found through 'yum' that fonts listed in the FC 10 package 
> were
>    there but not 'installed'--compared it to my server machine (FC8 at 
> school) where
>    they were listed as 'installed'.  Now, I'm the true gorilla in a glass 
> blowing shop
>    with this stuff, and x11 fonts stuff gives me the cold shivers.  But sort 
> of
>    following along I managed to do a yum install on a bunch of fonts, and now 
> that
>    startup message from mcidas has gone away and I have very easy to read 
> help screens.
>    Problem 1 fixed.

Excellent!

> 2) The GUI wouldn't start (as I suspected) because I had DATALOC info that 
> was not appropriate
>    for my machine.  Once I cleaned up LOCDATA.BAT and deleted some entries 
> from the running
>    mcidas all is well...the GUI will start.

When I logged onto your machine last week, I noticed that a number of the 
DATALOCs were set
to the string that was designed to be replaced in ~mcidas/data/LOCDATA.BAT:

fully_qualified_host_identifier

> 3) Unfortunately, my first sortie with the GUI was only moderately 
> successful.  I made an
>    image of NA (GINICOMP) with a map on it.  Magnified it a bit.  Then 
> started examining point
>    source options.  After putting the RPTSRC temp (twice, once in the default 
> Kelvin, once in F)
>    and dewpoint temp on the map (in graphics default, very ugly overprinting 
> everything), the
>    data selection buttons for point source data hung.  I could not select 
> another parameter to
>    plot even though the data set, date and time were correct and the same.  
> Anyway, as a newbie
>    to the GUI I'm sure I did something wrong.  I'll play with it over the 
> coming months.

There are two GUIs provided in Unidata McIDAS:

GIU   - the GUI from SSEC
MCGUI - the GUI from Unidata

Both GUIs use the routine 'mcwish' that is created as part of the McIDAS build. 
 The test I
was interested in you running was for the Unidata GUI, MCGUI, but running 
SSEC's GUI would
exercise the same thing.  Apparently, since you were able to run SSEC's GUI, 
there was no
problem with the copy of 'mcwish' that was built as part of your Unidata 
McIDAS-X v2008
installation under Fedora 10.  This is useful information for me as I was 
having LOTS of
problems on my 64-bit Fedora 10 machine.

Just so you know, you can bring up the MCGUI by running:

mcidas -config

- click on the radiobutton for MCGUI
- click Yes on Save configuration values to defaults file
- click on Start

Update:

I released addendum v2008c this morning.  This addendum contains the latest 
version of
Tcl/Tk, v8.5.6, and some reworked code used by the Unidata GUI, MCGUI.  I now 
have no
problem running the MCGUI on my 64-bit Fedora 10 machine.  Installation of the 
addendum
is the same as installation of the previous one:

<as 'mcidas'>
cd ~mcidas/mcidas2008/src
make clobber
cd ~mcidas/mcidas2008/update
ftp ftp.unidata.ucar.edu
  <user> .....
  <pass> .....
  cd unix/2008/bugfix
  binary
  get mcupdate.tar.gz
  get mcupdate.list.2008c
  quit
./mcupdate
cd ~mcidas/mcidas2008/src
make mcx
make install.mcxall

> Just wanted to let you know that, from my end and to my limit of 
> understanding,
> all works fine.

Very good, thanks for the update.

Item #2:

We are experiencing problems serving particular sets of data with the ADDE GRIB 
server
on adde.ucar.edu to a (your?) machine at Missouri State that has an IP address 
of
146.7.96.190.

Every day we see 9 'gribgget' processes (the ADDE GRIB server) "hang" (not 
terminate).
The invocations appear to be attempts to display RTGRIBS/GFS-ALL surface wind 
speeds
or wind barbs, or wind streamlines over the continental US (LAT=15.56 62.49 
LON=47.57 150.63).
The "hung" processes drive the load average up to around 17 on adde.ucar.edu, 
and
must be killed by 'root'.

Questions:

- are you kicking off ADDE accesses to RTGRIBS/GFS-ALL surface winds

- if yes, can you send me the invocations that you are running so that we can
  better track down the cause for the hangs of the ADDE GRIB server?

Thanks in advance for your help in troubleshooting this problem!

> Thanks again.

No worries.

Cheers,

Tom
--
****************************************************************************
Unidata User Support                                    UCAR Unidata Program
(303) 497-8642                                                 P.O. Box 3000
address@hidden                                   Boulder, CO 80307
----------------------------------------------------------------------------
Unidata HomePage                       http://www.unidata.ucar.edu
****************************************************************************


Ticket Details
===================
Ticket ID: EUT-864314
Department: Support McIDAS
Priority: Normal
Status: Closed