Re: [Condor-users] About getting the output files after job completion

["Kewley, J \(John\)" <j.kewley@xxxxxxxx>]

 	Hi,
	 I have been trying to run a shell script say: job.sh and from
this shell script I am running a simple helloWorld program that
	prints 1-100. Now when my job completes I get all the outputs
printed within the shellscript with
	"echo" in my output file but the integers printed within the
helloWorld program are not in that output file.
	Could anybody shed somelight on what extra config parameter I
should set in the condor submit file?
	I am configuring it as:
	
	Universe = vanilla
	Executable = test.sh
	arguments = helloWorld

Your test.sh shellscript does not process any arguments, yet you provide
helloWorld as a parameter.

	output = helloWorld.output
	error = helloWorld.error
	should_transfer_files = YES
	when_to_transfer_output = ALWAYS
	Notification = ALWAYS
	Notify_user = myself@xxxxxxxxxxxx
	queue

You don't specify helloWorld as an input file for transferring. Does it
already exist on the target machines?
	
	and inside the shell script I am writing:
	
	#!/bin/sh
	setarch i386 -R ./helloWorld -_condor_D_ALL &

I am unfamiliar with setarch and the version on my machine doesn't have
a -R option. What I am suspicious of is the "&" at the end. This will
run that
line in the background, so any output from that prorgram is likely to be
printed
after the shellscript has terminated (if indeed it can find helloWorld)

	echo "Done!!"
	
	
	In the helloWorld.output file, I am only getting:
	Done!!

I presume that the setarch line is failing due to lack of ./helloWorld,
but does so in the background so you don't see any errors in
helloWorld.error
(or do you, you don't say)

	But the stuffs printed in helloWorld file (I am using simple
printf() by the way) are not in that output file.

Cheers

JK