If encounterting SEVERE errors, running jobs should be tried to get stopped and the heritrix engine should be shutdown. An NFS mount that contained heritrix's job directory went down last night. This lead to gazillion of Exceptions of the form:
The local disk that contains the heritrix_out.log ran full within minutes because of this flood of exceptions. In such non recoverable cases, the engine should be shutdown while trying to save as much state as possible.