Understanding Dynamic JDB Table Cache Refresh

The JDB (JD Edwards Database) Cache feature improves performance by caching specific tables in the Call Object Kernel, Subsystem UBEs and UBEs. With tables cached in JDB, any update or delete of a record in a cached table requires a clear of the cache on all kernels across all enterprise servers for the data to be available to all users.

Dynamic cache refresh allows a specific JDB table cache registered in the F98613 table to be cleared across kernel processes and across all enterprise servers as long as they are managed by the same server manager. All enterprise servers no longer need to be bounced in order to clear cache. Because there is no table caching in JAS, there isn't any need to clear cache in the JAS servers.

For example (for a single table), at the end of an accounting period a person in the finance department needs to update the current period in the Company Constants application. The current period value is important for all financial processes within EnterpriseOne and must be accurate to ensure proper processing. Application P0010 updates the Company Constants (F0010) table, which is a table that is normally cached. Currently within EnterpriseOne, a system administrator must log off all current users and restart the enterprise server in order for all users to have the most current data available for their applications. Each user must sign off the system while it is being reset, then sign back in when it is completed. This causes a disruption of normal business operations. A user may have hundreds of companies that need to have their current period changed on a monthly basis, causing even more disruption.

There are two methods available for a user to dynamically refresh cache for a specific table:

1. Reset database table cache for a specific table using P986116D.

2. Reset database table cache using a pre-configured application.