...rerun.
It is possible to tell the parallel job to not rerun these jobs with the assumption that the problems with the node will be fixed.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Stanford Exploration Project
5/3/2005