More Free Memory Available on a specific database node than others
We are highly parallized environment, as are all Exadata environments. Recently, we have experienced some free memory issues that have caused some node evictions to occur...We are running at 72 GB of memory per node, with SGA's totally about 28 GB of that memory for the two databases, ASM instance, and the DBFS instance....
We have one node, in this case, node 3 which always has significantly more free memory available than any other database node.
Reviewing the amount of parallel slaves running on the nodes, the distribution of workload, etc all appear to be the same...