Chapter 19. Changes in Data Grid 7.3

The Data Grid server is based on EAP 7.2 in this release. Make sure you consult the right version of the EAP documentation and that any underlying EAP configuration or setup is supported in 7.2.

For information about supported versions of EAP in this release, see the supported configurations at https://access.redhat.com/articles/2435931.

Data Grid now provides Library and Java Hot Rod Client modules for EAP as a single package. In previous releases, Data Grid provided Library modules and Java Hot Rod Client modules as separate packages. Download the modules for EAP from the customer portal.

For Apache Camel integration, the Red Hat Fuse team actively develop and maintain a camel-infinispan component that supersedes the camel-jbossdatagrid component. Red Hat recommends that you use the camel-infinispan component that is available with Red Hat Fuse 7.3 and later.

Refer to Red Hat Fuse documentation for more information.

Data Grid 7.3 introduces changes to internal marshalling functionality that are not backwardly compatible with previous versions of Data Grid. As a result, Data Grid 7.3.x and later cannot read cache stores created in earlier versions of Data Grid. Additionally, Data Grid no longer provides some store implementations such as JDBC Mixed and Binary stores.

Use StoreMigrator.java to migrate cache stores. This migration tool reads data from cache stores in previous versions and rewrites the content for compatibility with the current marshalling implementation.

For more information, see Store Migrator.

The Data Grid Memcached endpoint no longer stores keys as java.lang.String. For better compatibility Data Grid stores keys as byte[] arrays that represent UTF-8 encoded strings.

If you use the Memcached endpoint, you should reload data in the cache to store keys as byte[]. Either perform a rolling upgrade or load the data from an external source.

The default memcache-connector is disabled for security considerations. To enable the memcache-connector you must configure the endpoint subsystem. See Memcached Connector Configuration.

Distributed scripts with a text-based data type no longer return null when the result from each server is null. The response is now a JSON array with each individual result, for example: [null, null].

The threads that handle the child Netty event loops have been renamed from *-ServerWorker to *-ServerIO.

You cannot configure cache stores as shared and then enable passivation for the stores. This configuration can result in data inconsistencies because it is not possible to synchronize data across nodes. For this reason, Data Grid now throws an exception at startup if it detects this configuration.

The default number of shards is now 4. In previous releases, the default number of shards was equal to the number of segments in the cache.

The AdvancedCacheLoader SPI allows Reactive Streams-based publishKeys and publishEntries methods that improve performance, threading, and ease of use. This change affects custom CacheLoader implementations.

This release deprecates features and functionality that affect migration to Data Grid 7.3. Notably compatibility mode is deprecated by configuring the MediaType for key/value pairs and storing data in binary format. The LevelDB cache store is also deprecated and replaced with the RocksDB.

For the complete list, see Features and Functionality Deprecated in Data Grid 7.3.