Copyright © TIBCO Software Inc. All Rights Reserved
Copyright © TIBCO Software Inc. All Rights Reserved


Chapter 27 Berkeley DB Shared All Persistence : Berkeley DB Shared All Persistence Overview

Berkeley DB Shared All Persistence Overview
This form of data persistence uses TIBCO BusinessEvents DataGrid as the cache provider and Oracle Berkeley DB Java Edition software as the data store (not bundled with the TIBCO BusinessEvents software).
With this form of persistence, cache servers act as persisters, and interact with the persistence layer. All cache nodes must have access to a reliable shared file system. The Berkeley DB data store is created within this shared file system.
The Berkeley DB Shared All Persistence option uses stores on a shared file system, such as NFS. This can result in better performance than use of a traditional DBMS product accessed over the network.
Reading from the Data Store
Persisters load data in bulk from the data store at startup. The following functions are used to load individual concepts into cache:
CacheLoadConceptByExtIdByURI()
CacheLoadConceptById()
Writing to the Data Store
Data is flushed to the operating system buffers for every write operation. TIBCO BusinessEvents uses WRITE_NO_SYNC durability by default, which means that Berkeley DB will flush every write to the operating system’s buffers immediately but not call fsync. In case of an application crash, there will be no data loss as long as the underlying operating system synchronizes its buffers to disk.
TIBCO BusinessEvents internally sets durability to WRITE_NO_SYNC durability. This behavior can be overridden by setting the je.txn.durability property in the je.properties file. See Configuring the Berkeley DB Shared All Persistence Option.
The CacheLoadConceptByExtId() function is not supported for use with this feature.

Copyright © TIBCO Software Inc. All Rights Reserved
Copyright © TIBCO Software Inc. All Rights Reserved