Does anyone have any experience they can share regarding database configuration for very large repositories?
Our custom object model has about 40 properties across several aspects (not all will be used on all documents - but assume worst case that they will be). There are 250million documents waiting to be loaded. This would generate over 250million rows in ALF_NODE but well over 10billion rows in ALF_NODE_PROPERTIES once system properties are also included.
What are the recommendations for managing a database this size? E.g. table partitioning
(The database would be MySQL or Oracle)
I've seen Alfresco have benchmarked over 1 billion documents but they gave no details about the DB configuration part from memory and CPU.