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

20000203: Unidata McIDAS-X & Linux



>From: "Guillory, Anthony" <address@hidden>
>Organization: NASA/MSFC
>Keywords: 200002032214.PAA09948 McIDAS-X Linux

Anthony,

>On a completely unrelated note to the past week...I have known for a while
>that you got McIDAS running under Linux.  I think you told me eons ago that
>you had a change file for the mods you have to make to the UW version to
>make it run under Linux.  Is that correct?

Yes.  I have a compressed tar file in a public area on our FTP server that
you can grab and play with.   The information you need is:

machine: ftp.unidata.ucar.edu
user: anonymous
pass: your_full_email_address
directory: pub/mcidas
file: mclinux.tar.Z
mode: binary

mclinux.tar.Z contains several files:

zcat mclinux.tar.Z | tar tvf -
-rw-r--r-- 530 2000    2016 Jul 31 18:20:11 1999 ../src/mcar.sh
-rw-r--r-- 530 2000   28010 Jul 31 18:21:24 1999 ../src/mccomp.sh
-rw-r--r-- 530 2000    4871 Jul 31 18:20:33 1999 ../src/tclcomp.sh
-rwxr-xr-x 530 2000    6657 Jul 31 18:20:35 1999 ../src/tclconfig
-rw-r--r-- 530 2000   33421 Jun 01 18:07:33 1999 ../src/mct_ebuf.c
-rw-r--r-- 530 2000   70267 Sep 14 14:59:40 1999 ../src/decoder.cp
-rw-r--r-- 530 2000   31188 Sep 15 17:57:47 1999 ../src/mcserv.cp
-rw-rw-r-- 530 2000   41586 Sep 14 19:28:16 1999 ../netcdf/ncgen/ncgentab.c
-rw-rw-r-- 530 2000   54208 Sep 14 19:28:16 1999 ../netcdf/ncgen/ncgenyy.c
-rw-r--r-- 530 2000   27648 Sep 14 11:33:25 1999 ../tcl/generic/tclPosixStr.c
-rw-rw-r-- 530 2000    4668 Sep 14 19:55:55 1999 f2c.h
-rwxrwxr-x 530 2000  222656 Sep 14 19:42:45 1999 f2c
-rw-rw-r-- 530 2000  227078 Sep 14 19:42:30 1999 libf2c.a

The referenced by ../src, ../netcdf, and ../tcl are source files that
have been modified for Linux support.  The files f2c.h, f2c, and
libf2c.a are F2C support for RedHat 6.x Linux.  In particular, they are
needed for the newer implementation of libc on many Linux systems.  If
you are going to try to get McIDAS working on a pre-6.x version of
RedHat Linux, then you need a different set of these, but, then again,
they are included in the older RH distributions.  If you are going to
try and get McIDAS working on distibutions of Linux other that the ones
by RedHat, then you need to be concerned with the version of libc that
is in their distribution.  I can help more on this if needed.

>We have a need to consider
>(soon) running McIDAS-X under Linux so would it be possible to get a copy of
>your list of needed changes.  

Absolutely.  I can tell you that I have found the Linux version of
McIDAS to be especially useful.  We see quite a few of our sites moving
to Linux since it runs on the nice, fast, cheap, commodity PC
hardware.  The other thing is that I run McIDAS in 24-bit mode on my
Linux box at home.  Other than a few McIDAS code bugs (that are
supposedly being fixed at SSEC), this runs well.

Please see the information above, and let me explain a little about how
the compressed tar file is designed to be used:

o in Unidata McIDAS-X, I have an additional directory under the
  mcidas7.x root.  That additional directory is named update.  For
  my 7.6 distribution, this is then:

  ~mcidas/mcidas7.6/update

o the compressed tar file is meant to be FTPed to the update directory
  and unpacked with:

  cd ~mcidas/mcidas7.6/update
  ftp ftp.unidata.ucar.edu
    <user>
    <pass>
    cd pub/mcidas
    bin
    get mclinux.tar.Z
    quit
  cd ../src
  mv mcar.sh mcar.sh.ssec
  mv mccomp.sh mccomp.sh.ssec
  mv tclcomp.sh tclcomp.sh.ssec
  mv mct_ebuf.c mct_ebuf.c.ssec
  mv decoder.cp decoder.cp.ssec
  mv mcserv.cp mcserv.cp.ssec
  cd ../update
  zcat mclinux.tar.Z | tar xvf -

The McIDAS files that I modified will be unpacked into the ../src
directory.

The netcdf files that are included in the compressed tar file are ones
from our production release of netCDF 3.4 (the ones included in the
SSEC distribution of McIDAS-X 7.60 were from a beta distribution).

The file tclPosixStr.c is one that was patched by RedHat for their
bundled version of Tcl/Tk 8.0.4 on 6.0 Linux.

The files f2c, f2c.h, and libf2c.a are the copies of f2c that I make
available to my sites that are upgrading to RedHat 6.0 Linux.  These
are for the new libc stuff in the RedHat 6.x distributions.  They will
not work for pre-6.x versions of RH, but, then again, those came with
appropriate versions of f2c.h, f2c, and libf2c.a.

For some detailed information about things in Linux that you have to
worry about, please review the Linux sections in:

http://www.unidata.ucar.edu/packages/mcidas/mcx/warnings_mcx.html

>Back to your subject of Monday - hopefully in a few minutes or tomorrow
>things will calm down enough for me to take a look at your server.

OK.  Any chance of me getting access to the same ADDE server that you
gave Gilbert access to?  I am just curious about what sorts of data
sets that you have available, nothing else (yet :-).  If you need to
allow access to specific machines, then I would request that you all
all machines in 128.117.140.*  (this is Unidata only).  If you need to
allow a limited number of machines, then please allow 128.117.140.56
(zero.unidata.ucar.edu) and 128.117.140.73 (lenny.unidata.ucar.edu).

Please let me know if you need help out with your Linux tests.  So far,
I have provided the Linux stuff to a number of MUG sites including:
Australian Bureau of Meteorology and SATEPS (used to be VDUC).

Tom

>From address@hidden  Fri Feb  4 10:53:41 2000

Anthony,

>Thank you, Tom!

You are certainly welcome.

>The person who will work on it (Paul Meyer, address@hidden) will
>likely send you questions as he starts to dig into it next week.

OK, sounds good.  The mods are minor, so the amount of time/effort should
be small.

>We are
>interested in Linux for the exact reasons you named, cheap, fast hardware.
>SGI's are nice but if you aren't doing super duper graphics then who cares.

Right.  BUT, like I said, I am running in 24-bit mode on my RH 6.1
box at home and never having problems of no colors when things like
Netscape (tm) are running.

>Also, I have a request in the works to add the two machines you named to the
>server list.

By the time I looked at this, it had already had been done.  Thanks!

>And I'll consider the rest of Unidata, but for now I'll limit
>it to those two.

OK.

>Long story, but let's just say security for now.

This gets back to the issue that I and others raised with SSEC.  A
number of us want the ability to limit access to certain IP addresses
for selectable datasets.  I have added this capability into the ADDE
image sub servers that I have written (NIDS, NOWRAD (tm), GINI, ABoM
radar).  I will continue this trend as I develop ADDE servers for
additional imagery datasets.  Right now, I believe tht I can knock off
AVHRR Level 1B pretty quickly.  I also want to tackle imagery in
Terascan (tm) format and DMSP.

While on the subject of DMSP...  Did you have anything to do with the
WETNET project at MSFC?  If so, do you know how they turned DMSP
imagery into McIDAS AREA files?  What I am looking for is the
code/concepts for the imagery navigation/calibration.  I can take this
code and incorporate it into ADDE AGET and ADIR servers so that we can
access the imagery from McIDAS (and the new Unidata MetApp routines)
directly.  Any help (such as providing code or pointing me to the
appropriate person) you could provide on this would be most
appreciated.

>I'll let
>you know when they've been added.

Too late :-)

>I will say that the ADDE stuff is nothing
>more than the stuff on the anonymous FTP site - just "actively" within the
>McIDAS world.

OK, got it.

>You will need to DATALOC ADD (dataset from below) 198.116.59.198:
>G8-GHCC (contains VIS, IR2, IR3, IR4, IR5)
>G8-4KM (contains VIS, IR2, IR3, IR4, IR5)
>G10-4KM (contains VIS, IR2, IR3, IR4, IR5)
>GM5-4KM (contains VIS, IR1, IR2, IR3) - navigation is off!!
>ME7-4KM  (contains VIS, IR1, IR2)

Thanks again.

>Anthony
>Anthony R. Guillory 
>NASA/Marshall Space Flight Center
>Global Hydrology and Climate Center
>977 Explorer Blvd.
>Huntsville, AL 35806-2807
>(256) 922-5894
>(256) 922-5723 FAX
>address@hidden <mailto:address@hidden>   
>Geostationary Satellite Data:  http://www.ghcc.msfc.nasa.gov/GOES
><http://www.ghcc.msfc.nasa.gov/GOES> 

Tom

>From address@hidden  Fri Feb  4 13:05:35 2000

Tom,

I agree with you about the security and restricting access to certain
datasets.  It is definitely a necessity.

As for DMSP, I'm not sure if this is the right person but I would suggest
you contact Matt Smith (I think he is out today).  address@hidden
<mailto:address@hidden>   If he isn't the right person he may be
able to track down the code or point you to the right person.

Also, glad you discovered your machines were added because I still haven't
been told that they were. :-)

Anthony

Anthony R. Guillory 
NASA/Marshall Space Flight Center
Global Hydrology and Climate Center
977 Explorer Blvd.
Huntsville, AL 35806-2807
(256) 922-5894
(256) 922-5723 FAX
address@hidden <mailto:address@hidden>   
Geostationary Satellite Data:  http://www.ghcc.msfc.nasa.gov/GOES
<http://www.ghcc.msfc.nasa.gov/GOES>