For OMI, an algorithm has been implemented to better coordinate concurrent data loading. It is now possible to have one load job per project (regardless of whether online or in batch); these jobs run concurrently with each other.
This algorithm makes use of user-definable lock handling directly on the Oracle server and dramatically decreases the likelihood of deadlock situations, while at the same time import performance is improved because it is no longer necessary for the OMI program to obtain row level locks on the import data. This global project lock handling needs to be configured with the following two parameters:
-
ZI_OMI_TBT - OMI batch lock timeout (maximum allowable lock wait time specified in minutes).
-
ZI_OMI_TON - OMI online lock timeout (maximum allowable lock wait time specified in seconds).