TOB mainly considers the maximum amount of memory available to the host JVM, and how much out of this amount is still free.
On start up, TOB pre-loads as many persistent objects as possible,
until the free memory percents drops under the value configured by
theobjectbase/mem-policy/@targetFreePercentage
. And object loading order is determined by each object's
OnlinePriority
number. The smaller an object's priority number is, the prior it tends
to be pre-loaded.
During runtime, objects with higher priorities (smaller priority numbers) are more sticky to memory.
So the maximum memory would made available to the JVM hosting TOB instance should better be adjusted to an optimal value regarding the persistent object amount, their OnlinePriority , and memory demand by other programs run in the same JVM.
