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

20000404: unidata contact



>From: David Harper <address@hidden>
>Organization: NCAR/RAP
>Keywords: 200003012325.QAA04248 McIDAS-X GINI netCDF ADDE server convert Linux 
>RedHat Debian

Steve,

>My name is Steven Webb.  I'm the system administrator for the ATEC project
>here at NCAR in Colorado.  I received the attached e-mail from Dave Harper who
>works with me at NCAR.
>
>Here's some information about our hardware/os:
>
>We're running the current "frozen" debian (Debian ver 2.2 with 2.2.14
>kernels).  We're running the libc6 (glibc2.1) libraries, which should be the
>same as RedHat (perhaps off by a minor version number, but basicially the
>same).

OK, this should work.

>I'd be happy to give you a normal user account on one of our debian machines
>here at NCAR, but note, you will receive no e-mail at address@hidden
>because this is a non-user account - will this be a problem?

This sounds good; This account doesn't need to receive email.

>I suggest that we try out the RedHat binaries first, and if that doesn't work
>I'll give you the account and you can try compiling them on our debian
>machine.
>
>Will this do?

Yes, this should do nicely.

>On Tue, 4 Apr 2000, David Harper wrote:
>
>> Tom with Unidata
>> Unidata Support <address@hidden>
>> 
>> We do not need the remote server he talks about in the second paragraph.
>> The meat of the last email I got from him:
>> 
>> Yes, I understand.  The fact is that giving you this subset of code as
>> precompiled binaries is relatively easy.  Figuring out the small set of
>> source code that would be needed to build the binaries is something
>> that is a LOT harder and is something that I would rather not have to do.
>> 
>> In order for me to figure out exactly which binaries I need to give you
>> I will need access to the machine that would run the code.  This access
>> would be best done as a user named 'mcidas'.  If you eventually want
>> the remote McIDAS ADDE server to work, the setup will require that
>> either I or someone else logon as 'root' to set it up.
>> 
>> In order to produce the binaries you need, I need to build them on the
>> platform on which they will run.  Since we use RedHat 6.1 Linux here,
>> we may not be able to simply produce the binaries and have them run
>> on your (I am guessing) Debian system.  The issue here is the newer
>> version of libc that comes with RedHat 6.x.  Again, I don't know enough
>> about your target system to know if things can/will work.
>> 
>> Thanks,
>> Dave
>> 
>> PS Let me know what happens.
>> 
>
>--
>EMAIL: (h) address@hidden  WEB: http://badcheese.com/~steve
>       (w) address@hidden
>           address@hidden

Tom Yoksas

>From address@hidden  Tue Apr  4 13:59:52 2000
>To: Steve Webb <address@hidden>
>cc: address@hidden, David Harper <address@hidden>
>Subject: 20000404: account information on RAP machine
--
>From: David Harper <address@hidden>
>Organization: NCAR/RAP
>Keywords: 200003012325.QAA04248 McIDAS-X GINI netCDF ADDE server convert Linux 
>RedHat Debian

Steve,

>ACCOUNT INFO:

Thanks

>Please use ssh for all of your sessions.

Will do.  I also have to use ncftp since RAP is behind the security
perimeter and Unidata is outside.

>You may send e-mail from this
>account, but will not be able to receive any e-mail to this account.

I have no need to email into or out of this account.

>This machine is running debian 2.2 (unstable/frozen).

OK.  A couple of quick tests indicates that the McIDAS code compiled
on our RedHat 6.1 Linux box works on your Debian 2.2 installation.

>Any other sysadmin requests should be sent to me.

OK, where can I find GINI data files?  The object of me putting
McIDAS executable code on a RAP machine is to:

o find out what the smallest subset of McIDAS that is needed to convert
  NOAAPORT GINI imagery to netCDF

o to setup the account so that conversions can be made

For the latter objective, I will need to be able to do test conversion
runs, so I will need to know where the source GINI files are located.

Thanks for the help...

Tom

>From address@hidden  Tue Apr  4 14:17:29 2000
>cc: David Harper <address@hidden>
>Subject: Re: 20000404: account information on RAP machine

David -

Can you help Tom find the data that he needs?

His account is on 4dwx-dev.

- Steve

>From address@hidden  Tue Apr  4 14:50:44 2000
>To: David Harper <address@hidden>
>Subject: 20000404: account information on RAP machine (cont.)

>What machine?  you don't seem to be on the one I expected.

4dwx-dev.rap.ucar.edu

Tom

>From address@hidden  Tue Apr  4 15:15:40 2000
>Subject: 20000404: account information on RAP machine (cont.) 

>You need to give write permissions to the directory.  I'm not considered part 
>of your group.

Done.  The directory is world read/write/execute now.

>I have 4 Puerto-rico files directly from the NOAAport.  I can also get others,
>but that should be a start.

I would rather have 4 different kinds of images.  Let's say

1 Puerto Rico Regional 10.7 um
1 Super National composite at 3.9 um
1 GOES-West VIS (big)
1 Alaska Regional WV

One of the primary reasons for the diversity is that I have to setup
McIDAS ADDE datasets for the GINI imagery.  The datasets are organized
by the product type and band, so I need different types to set things up.

I was hoping that I would be doing this on a system that had direct
access to all of the GINI imagery so that I could setup all of the
dataset names that were needed and then be able to do random conversion
tests.  On whatever is the final system, the datasets will have to be
setup in order to access the GINI imagery AND write to netCDFs.

Tom

>From address@hidden  Tue Apr  4 15:22:05 2000
>Subject: Re: 20000404: account information on RAP machine (cont.)

>I got the Puerto-rico moved over, I'm still digging out the other files.

Dave

>From address@hidden  Tue Apr  4 15:41:29 2000
>Subject: Re: 20000404: account information on RAP machine (cont.)

Tom,

>I just dropped in conus visible.

Dave

>From address@hidden  Tue Apr  4 15:50:30 2000
>Subject: Re: 20000404: account information on RAP machine (cont.)

Tom,

>I think we have a problem.  I have several gigs of data, but it looks
>like only a few sets of the gini data.  I'm trying to see if this is
>correct, but I think their tape ran out as we got to the good stuff.

>We will be doing conversions on files dealing with each local area.

>Here are the files I currently have access to:
>35 visible conus (goes_west)
>Puerto_rico: 11micron  12micron  4micron   7micron   visible
>w_atlantic: 11micron  7micron   visible

-Dave

>From address@hidden  Tue Apr  4 16:13:32 2000
>Subject: Re: 20000404: account information on RAP machine (cont.)

I'm leaving for a trip tomorrow.  I'll be talking to the guys that gave us the
Gini data and I will be getting the missing data.  Can you CC my home address:
address@hidden

Does it work properly on the files I gave you?  Unfortunately the rest will have
to wait until Monday.

-Dave