When jobs vacate the pool, the condor_schedd can take a lot of time in cleaning up job spool directories. During this period, the schedd is unresponsive. Customer noted that breaking up the cleanup tasks into smaller chunks would help make the schedd more responsive.
This is related to https://bugzilla.redhat.com/show_bug.cgi?id=523801 Avoiding job spool creation might fix sufficiently fix this problem.
*** This bug has been marked as a duplicate of bug 523806 ***