v.19.15Improvement

Improvements for Failover of Distributed Queries

Official CH changelog text
Improvements for failover of Distributed queries. Shorten recovery time, also it is now configurable and can be seen in system.clusters. #6399 (Vasily Nemkov)
Improved failover mechanism for Distributed queries that shortens recovery time and makes failover behavior configurable.

Why it matters

This feature addresses the need for faster and more controllable recovery in Distributed queries when a node or replica fails. By reducing the failover recovery time and exposing configuration options, it enhances query reliability and gives users better insight and control over cluster failover behavior via system.clusters.

How to use it

Users can configure the failover parameters for Distributed queries in the cluster configuration files. The current failover status and configuration can be monitored through the system.clusters table. Adjusting these settings allows tuning the failover recovery time according to workload needs.

Related resources

PreviousNext