Right now the query system doesn't do a good job of checking whether the operation has been killed using

killCurrentOp.checkForInterrupt()

This can make it impossible to kill long-running queries. Runners should check for interrupt inside their getNext() method.