TPump: - if you use multiple sessions, be shure to SERIALIZE on the PI columns - use ARRAYSUPPORT ON -In worst case you have to set PACK to 1
Can't you combine both jobs into 1 or run them serially?
MLoad: Of course you could try to change your script to MLoad, but there's probably a reason why TPump over MLoad was choosen initially. But if you know both jobs it would be better to control when they run, one after the other.
MultiLoad reference: "The TABLEWAIT runtime option in which hours is the number of hours that Multiload continues trying to start Multiload when one of the target tables is being loaded by some other job
When the Multiload utility tries to start MultiLoad, and the Teradata Database indicates that the target table is being loaded (Error 2574), the Multiload utility: • Waits for 6 minutes, by default, or for the amount of time specified by the SLEEP runtime option. • Then tries to start MultiLoad again. Multiload repeats this process until it has either acquired the table or exceeded the TABLEWAIT (hours) time period."