5.2. Resolved Issues

When the Remote Query feature is used through the camel-jbossdatagrid component in EAP, and JDG modules for EAP are installed, the application throws a java.lang.LinkageError and the Remote Query feature does not work properly.

This issue is resolved as of Red Hat JBoss Data Grid 6.5.0.

When a near cache has not received remote events for some time, for instance if the cluster is inactive or the client does not load new data, the connection to remote JDG servers fails with java.net.SocketTimeoutException.

This issue is resolved as of Red Hat JBoss Data Grid 6.5.0.

When a node which was not part of previous view (e.g. due to network partition) joins a cluster, it may request retransmission of previous messages (JGroup's NAKACK). If some of those messages are prepare commands of transactions which have already finished on other nodes, then transactions can be prepared and committed on the joining node once again. This behavior may lead to inconsistent values across the cluster if the affected entries have been modified on nodes from the original view.

This issue is resolved as of Red Hat JBoss Data Grid 6.5.0.

The camel-jbossdatagrid component requires embedded cache dependencies when a remote cache is used. The required dependency is infinispan-embedded.jar as well as the expected infinispan-remote.jar. The same issue exists when using only an embedded cache, as the infinispan-remote.jar dependency is required in this case together with the infinispan-embedded.jar.

This issue is resolved as of Red Hat JBoss Data Grid 6.5.0.

Passing TCK test suite is an ultimate criterion of complying with the JCache specification. Some of the tests are still not passing with HotRod implementation (i.e. functionality is not provided yet).

These relate to JMX statistics, expiration and listeners, where correct functioning currently cannot be guaranteed.

This issue is resolved as of Red Hat JBoss Data Grid 6.5.0.

A new Maven repository, jboss-datagrid-$RELEASE-maven-repository-full.zip has been provided as a zip file; however, this repository is missing required dependencies.

This issue is resolved as of Red Hat JBoss Data Grid 6.5.0.

When starting a cluster with multiple nodes concurrently, a deadlock on GlobalComponentRegistry sporadically occurs which will prevent the nodes from starting completely.

This issue is resolved as of Red Hat JBoss Data Grid 6.5.0.

When datacenter replication is configured and the backup node is down messages similar to the following are seen in the log: However, these error messages are logged even with failure-policy set to IGNORE.

These messages indicate that the cluster can not communicate with the backup node, and do not represent immediate issues with the cluster itself.

The Backup Failure policy is now honored as of Red Hat JBoss Data Grid 6.5.0; to resolve this issue ensure that the remote backup is online and there are no communication errors.