> you should write something into the first-line of the code that looks > to see if there's a job already running; and quits if there is.
Of course. My daemon wrapper code checks if that particular job is already running, and if so skips it. If the job is still running after the Nth skip, it sends me a warning email > > the issue with that approach is that db blocking and long jobs can > leave you with 1,2,3,4,5,6,7,etc cron jobs ; and your box eventually > crashes Why are you using DB locking for such long periods? Almost none of my code requires locks, and if there is something that DOES require a lock, I lock, transform, unlock, making sure that the transform process is quick. Clint
