![]() The single GraphDB distribution requires at least Java 11. It is now simpler than ever to create a repository using a GraphDB Free license and upgrade to SE or EE at any time without even a restart. In line with that, we refactored the existing GraphDB repository types (Free, SE, and EE worker) into a single new repository called simply “GraphDB repository”. Unlike previous versions of GraphDB, version 10 is a single distribution that can run in Free, SE, or EE mode, depending on the currently set license. This makes it possible for every transaction to see its own changes before commit and unlocks some new use cases, such as using the Sequences plugin, that were not possible with the old cluster. The new cluster also handles transactions differently, in a manner that is very similar to how transactions are handled in a non-cluster environment. While the old cluster needed to be defined at the repository level, the new cluster is defined per entire GraphDB instance, which means that once a repository is created in the cluster, it automatically becomes part of the cluster. This ensures a high uptime, zero data loss, fault tolerance, and smooth recovery from unexpected situations and scenarios. The Raft consensus algorithm uses majority voting to determine the current leader, and consensus to confirm every write operation. Unlike the old cluster, each node in the new cluster contains a replica of the data. The leader is akin to the master node in the old cluster, while a follower is similar to a worker. In the new cluster, any node can be either a leader or a follower. This GraphDB 10 release introduces the new high-availability cluster based on the Raft consensus algorithm. GDB-6755 Rapid cluster topology changes may lead to OverlappingFileLockException GDB-6952 Restoring a backup more than once may lead to stale cache and some operations not working GDB-7033 Cluster node not responsive after all cluster nodes were stopped and some nodes were restarted GDB-7048 Cannot add a node that was previously unsuccessfully added without a license GDB-7059 GraphDB no longer allows variables over grouped value as a valid query projection GDB-7074 Removed cluster nodes cannot be added back in some cases ![]() GDB-7088 Cannot add a node that was previously removed and has more than one entry in its transaction log GDB-7094 Cluster report shows an error in the log file GDB-7104 GraphDB fails to restore from snapshot upon initialization GDB-7148 Starting a new transaction from the transactions API returns wrong Location header GDB-7158 Cluster follower does not replicate after out-of-sync GDB-7164 The external proxy fails to read the cluster status when GraphDB basic security is enabled GDB-7165 The external proxy fails when you try using it for a PATCH request to the /rest/cluster/config api GDB-7171 Named graph not shown by getContexts() when it contains an owl:sameAs statement GDB-7172 Cluster failed quorum transaction does not always release transaction log lock GDB-7184 Out-of-sync node will become follower if verification of entry is synchronous GDB-7207 Error in implicit cluster transactions leaves stale transactions
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |