It would be good to make sure that in case a job is removed from condor queue using condor_rm , the stdout of the job comes back.

In vanilla condor case, that seems to be the behavior where a long running job stdout is staged back when user does condor_rm

In Pegasus case, all jobs are wrapped by kickstart, and a LIGO user reports that the stdout is not staged back.

It maybe that pegasus-kickstart may need to setup appropriate signal handler and pass it to the running job.