[CIG-MC] ssh and autocombine.py

[Robert Moucha]

Hi Eh,

I checked and I do have my path variable set correctly so that is not
the problem. However I did figure out what is happening, but I don't
know enough python to figure how to fix it:

If I print the variable cmd in the function batchpaste in
batchcombine.py I get the following:

ssh c0-0
 pasteCitcomData.py /state/partition1/moucha2/citcoms cookbook6
coord,velo,visc 3 10 /home/moucha/convection/citcoms/exa

Note the return character after the node name (not an artifact of
wrapping)! So, the os.system call only executes the ssh c0-0.  This
explains why I am successively logged into the nodes whey I run
autocombine.py, but the command pasteCitcomData.py is not executed.
And when I type exit, the command pasteCitcomData.py is executed, but
now the file does not exist on the master node and so I get a file not
found error. I went back into batchcombine.py and added:

print node
print 'this is a test'

Now I get:

c0-0

this is a test

So the return character is coming from the variable node (note the
blank line between the print statements). My mpirun.nodes looks like
this:

c0-0
c0-34
c0-33
c0-46

Once again, thanks for your help on this.

Rob


On Wed, Nov 4, 2009 at 4:39 PM, Eh Tan <tan2 at geodynamics.org> wrote:
> Hi Rob,
>
> You need to add the visual/ directory to your PATH environment variable.
> If you are using bash, you can add this line in your ~/.bashrc:
>
> export PATH=$PATH:$HOME/convection/citcoms/visual
>
> Eh
>
> Robert Moucha wrote:
>> Hi Eh,
>>
>> Thanks for getting back to me so quickly. Our cluster is set-up with
>> password-less ssh into the nodes. I went into
>> batchcombine.py and added the line print cmd, to see what command was
>> being executed, here is what I get:
>> ------------------------
>> $ autocombine.py mpirun.nodes pid23816.cfg 10
>>
>> ssh c0-0  pasteCitcomData.py /state/partition1/moucha2/citcoms
>> cookbook6 coord,velo,visc 0 10
>> /home/moucha/convection/citcoms/examples/Cookbook6
>>
>> Last login: Wed Nov  4 16:04:40 2009 from geograppe.local
>> Rocks Compute Node
>> Rocks 5.0 (V)
>> Profile built 16:38 16-Jul-2008
>>
>> Kickstarted 16:50 16-Jul-2008
>> [moucha at compute-0-0 ~]$
>> -------------------------
>>
>> So I tried from my master node:
>>
>> $ ssh c0-0  pasteCitcomData.py /state/partition1/moucha2/citcoms
>> cookbook6 coord,velo,visc 0 10
>> /home/moucha/convection/citcoms/examples/Cookbook6
>>
>> and I get:
>>
>> bash: pasteCitcomData.py: command not found
>>
>> Now with the full path:
>>
>> $ ssh c0-0  ~/software/cig/bin/pasteCitcomData.py
>> /state/partition1/moucha2/citcoms cookbook6 coord,velo,visc 0 10
>> /home/moucha/convection/citcoms/examples/Cookbook6
>>
>> It works!  Funny though, because when I log into the node, I can
>> execute the command directly.  Do you know of a way to solve this
>> without editing the scripts, or is this just the way ssh works?
>>
>> Thanks
>>
>> Rob
>>
> --
> Eh Tan
> Staff Scientist
> Computational Infrastructure for Geodynamics
> California Institute of Technology, 158-79
> Pasadena, CA 91125
> (626) 395-1693
> http://www.geodynamics.org
>
>
>
>
> --
> Eh Tan
> Staff Scientist
> Computational Infrastructure for Geodynamics
> California Institute of Technology, 158-79
> Pasadena, CA 91125
> (626) 395-1693
> http://www.geodynamics.org
>
> _______________________________________________
> CIG-MC mailing list
> CIG-MC at geodynamics.org
> http://geodynamics.org/cgi-bin/mailman/listinfo/cig-mc
>



-- 
GEOTOP - Département des Sciences de la Terre et de l'Atmosphère
Université du Québec à Montréal
CP 8888, succursale Centre-Ville
Montréall, Québec
Canada  H3C 3P8
Tel:     (1-514) 987-3000, ext 1554#
FAX:     (1-514) 987-3635