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

20031114: McIDAS build at Western Michigan (cont.)



>From: Karthik Samudram Jayaraman <address@hidden>
>Organization: Western Michigan
>Keywords: 200310251920.h9PJKRgf026027 McIDAS install

Karthik,

>Because of the space constraint , our sysadmin asked us to use 
>the /cutrim1/mcidas as the location to download the files and install 
>them.

OK.

>I think the installation problem was with the location of 
>the "cc" and "f77". 

??

>The location of these compilers were in /opt/SUNWspro/bin/cc 
>&  /opt/SUNWspro/bin/f77, now it has been changed to 
>the /opt/Vendor/SUNWspro/bin/cc & /opt/Vendor/SUNWspro/bin/f77 
>resplectively.

OK.

>I have changed the path for CC and FC environment variables.
>
>I did run the make file once again. The installation went past the 
>point where it used to get stuck up the previous time, but i had an 
>fatal error at another point. So i think the system should now be 
>fine with the parameters ( atleast CC and FC is now fine).

I logged onto s400 and took a look at the .cshrc file for the user
'mcidas'.  The problem I found was that 'path' was being redefined down
in the file:

set path=(${MCGUI} ${syspath} ${BIN} /usr/sbin .)

'syspath' was defining /usr/local/bin to come before /usr/bin.  This
was causing /usr/local/bin/uname to be used in stead of the Sun
version, /usr/bin/uname.  I commented out this line since it is not
needed.

The other things I did in .cshrc are:

- add definition of F90 to be blank:

setenv F90

- took out unneeded directories in the PATH specification close to the
  top of the file.  Here is how it stands now:

# Unidata McIDAS-X definitions
setenv PATH  
/usr/bin:.:$HOME/bin:/usr/openwin/bin:/usr/ccs/bin:/usr/local/bin:/home2/autocad/autocad/acad:/home2/software/ideasadm/ideas/ms2/bin:/opt/moldflow/mf90/bin:/home3/software/proengin/15.0/bin:/opt/Vendor/SUNWspro/bin

I made these changes active in my session with:

source ~/.cshrc

The next thing I did was to cd to the McIDAS-X v2003 installation and
start a build from scratch (as I outlined in previous emails):

cd /cutrim1/mcidas/mcidas2003/src
make clobber
make all
make install.all

The build went without any problems, so I followed it up with the 'make
install.all' I listed above.

So, McIDAS-X v2003 is now installed on s400.  In order to exercise the
code a little bit, I setup McIDAS to look at various datasets on a
remote server running at the University of Nebraska-Lincoln:

cd /cutrim1/mcidas/data
cp DATALOC.BAT LOCDATA.BAT
<edit LOCDATA.BAT to set DATALOCs to point at papagayo.unl.edu for
 a variety of datasets and one other at SSEC>

cd /cutrim1/mcidas/workdata
batch.k LOCDATA.BAT                <- makes the settings in LOCDATA.BAT active

dataloc.k LIST
Group Name                    Server IP Address
--------------------         ----------------------------------------
AMRC                         UWAMRC.SSEC.WISC.EDU
BLIZZARD                     ADDE.UCAR.EDU
CIMSS                        PAPAGAYO.UNL.EDU
GINICOMP                     PAPAGAYO.UNL.EDU
GINIEAST                     PAPAGAYO.UNL.EDU
GINIWEST                     PAPAGAYO.UNL.EDU
GOESEAST                     UNIDATA2.SSEC.WISC.EDU
ME7                          IO.SCA.UQAM.CA
MYDATA                       <LOCAL-DATA>
NEXRCOMP                     PAPAGAYO.UNL.EDU
RTGRIDS                      PAPAGAYO.UNL.EDU
RTIMAGES                     PAPAGAYO.UNL.EDU
RTNEXRAD                     PAPAGAYO.UNL.EDU
RTPTSRC                      PAPAGAYO.UNL.EDU
RTWXTEXT                     PAPAGAYO.UNL.EDU
TOPO                         <LOCAL-DATA>

<LOCAL-DATA> indicates that data will be accessed from the local data directory.
DATALOC -- done

dsinfo.k I GOESEAST

        Dataset Names of Type: IMAGE in Group: GOESEAST

Name         NumPos   Content
------------ ------   --------------------------------------
IMAGES         250    METEOFORUM GOES EAST IMAGE SECTORS
WH-1KVIS        50    METEOFORUM GOES EAST 1 km 0.65 um VIS
WH-4K39         50    METEOFORUM GOES EAST 4 km 3.9 um SHORT IR
WH-4KCO2        50    METEOFORUM GOES EAST 4 km 13.3 um CO2
WH-4KIR         50    METEOFORUM GOES EAST 4 km 10.7 um THERMAL IR
WH-4KWV         50    METEOFORUM GOES EAST 4 km 6.5 um WATER VAPOR

DSINFO -- done

imglist.k GOESEAST/WH-1KVIS
Image file directory listing for:GOESEAST/WH-1KVIS
 Pos Satellite/         Date       Time      Center   Band(s)
     sensor                                 Lat  Lon
 --- -------------  ------------  --------  ---- ---- ------------
  43  G-12 IMG      14 NOV 03318  20:45:00     0   72 1
imglist.k: done


>We don't intend on using the old version of the McIDAS though. So i 
>think i will probably ( with the consent of my professor ) will 
>remove the old version.

It should be removed.

One of the main reasons I asked about the 7.6 installation of McIDAS
was that by removing it, there was plenty of room to install McIDAS in
the recommended manner, under ~mcidas.  In fact, even without removing
the old version of McIDAS, there was plenty of room to install McIDAS
under ~mcidas:

df -k /import/home01
Filesystem            kbytes    used   avail capacity  Mounted on
ifs2.wmich.edu:/export/home01
                     52428800 8215084 41452217    17%    /import/home01

So, it seems to me that the McIDAS installation could have been done in
the recommended way.  The only reason I mention this is that the closer
a McIDAS installation is to what is recommended, the less problems a
site will have when configuring and running the package.

So, what is the next step for McIDAS at Western Michigan?  Is the
intention to run an LDM to ingest data and then decode it with the
McIDAS-XCD package?  If yes, has there been consideration as to where
to put that data?  I see that there is plenty of space under either
/cutrim1 or /cutrim2, so those would be the logical place to put data.

Tom

>From address@hidden Sat Nov 15 09:08:56 2003

Tom,

Thats wonderful and my hearty thanx.
I will go thro what is been done and will keep you updated of the 
status here.

Sincerely,

Samduram Jayaraman Karthik.