Your post is fairly nonspecific but this sounds similar to a problem I ran
into on a Red Hat 3 update 3 machine.  This machine would panic when certain
tasks were run out of the crontab; the culrpit was eventually found to be
the 'audit' service.  If you're running that service and don't require it,
try to shut it down ('service audit stop')  and see if the problem still
persists.  Note, if you DO shut down the audit service, you'll find a lot of
warning messages showing up in the logs.  You can get rid of these messages
by removing the /dev/audit device node.