Introducing VoltDB v6.9
VoltDB v6.9 is shipping. Let’s look at what’s new.
VoltDB partitioning optimized to increase availability
VoltDB has made changes to partitioning that improve K-safety.
K-safety still guarantees the cluster will remain viable if K nodes fail. Under our new distribution algorithm, the cluster is substantially more likely to survive additional node failures, beyond the K-safety guarantee. The new algorithm is also applied as nodes rejoin or are added to the cluster. The new method of distributing partitions, called partition groups, is automatic and does not require any user action. The @SystemInformation system procedure has been enhanced to report the alignment of partition groups and placement groups within the cluster.
With no changes at all to their deployments, customers who upgrade will have a chance at greater uptime.
New SQL window functions
VoltDB v6.9 adds support for four new window functions: COUNT(), MAX(), MIN(), and SUM(), improving analytics capability in VoltDB. See the description of the statements in the Using VoltDB manual for details.
VoltDB v6.9 supports sending SNMP traps for selected server events, such as hosts joining and leaving the cluster, DR transitions, and excessive system resource usage. See the chapter on Monitoring VoltDB Databases in the VoltDB Administrator's Guide for details.
New @Statistics selector, DRROLE
The @Statistics system procedure has a new selector, DRROLE, that provides information about the role and status of database replication (DR) for the current cluster. The DRROLE selector reports whether the cluster is actively participating in replication and the role of the cluster (master, replica, or XDCR).
C++ query timeout
The C++ API now supports a configurable query timeout similar to the JAVA API timeout. See the README file included with the C++ client library for details.Download VoltDB