I've found a couple of problems with pegasus-run and pegasus-remove in my LDG DAX testing. It looks like pegasus-remove only works once. If I

pegasus-plan blah...
pegasus-remove /usr1/dbrown/log/H1L1V1-s6_lowmass_ihope-932169543-10457.ygi-jr/.
pegasus-run /usr1/dbrown/log/H1L1V1-s6_lowmass_ihope-932169543-10457.ygi-jr/.
pegasus-remove /usr1/dbrown/log/H1L1V1-s6_lowmass_ihope-932169543-10457.ygi-jr/.

(with the appropriate options) the second pegasus-remove fails, as it tries to remove the original condor dagman job id, not the new one from the re-run.

Also pegasus-run has been printing messages like

ERROR: Your database handle is undefined: ERROR: Please set property pegasus.catalog.work.db.driver or pegasus.catalog.*.db.driver to a database driver!

FATAL: Workflow was not entered in wf_work database
Warning: While trying to access work database
ERROR: Please set property pegasus.catalog.work.db.driver or pegasus.catalog.*.db.driver to a database driver!

These seem harmless as the DAG works fine, but "FATAL" and "ERROR" will cause LIGO users to panic.