One of our test cases /012-blackdiamond-invoke started failing for 4.5.0cvs after a fix for PM-829 was pushed out.

the reason for this failure is that now /012-blackdiamond-invoke test case results in a transfer job that run remotely and is invoked using the kickstart invoke functionality.

however, when the planner does it the stage in jobs arguments look like
arguments = " -n pegasus::transfer -N null -i - -R condorpool -L blackdiamond -T 2015-03-19T15:01:12-07:00 -w /nfs/ccg4/scratch-purge-no-backups/bamboo/scratch/bamboo/pegasus/blackdiamond/run0001 -I stage_in_remote_condorpool_1_0.arg "

and it also has
transfer_input_files = /nfs/ccg3/scratch/bamboo/012-blackdiamond-invoke/work/bamboo/pegasus/blackdiamond/run0001/stage_in_remote_condorpool_1_0.arg,/tmp/x509up_u550

however, the transfer_input_files transfers the files to condor scratch dir. kickstart tries to launch the job in workflow scratch dir and then the stage jobs fails.