High availability#

Starburst control plane supports two kinds of high availability:

  • Control plane high availability keeps the control plane itself running. You run several control plane instances against one backend database, with a load balancer in front of them. See How control plane high availability works.

  • Coordinator high availability keeps your Starburst Enterprise platform (SEP) clusters running. The control plane monitors the coordinators of a cluster, and it promotes a standby coordinator when the active coordinator fails. See How coordinator high availability works.

Known issues#

The following issues affect high availability:

  • Helm and managed mode. The SEP Helm chart sets its own discovery.uri, and managed mode needs discovery.uri to equal deployment.manager.uri. Set discovery.uri yourself, to the same value, in additionalProperties. See Configure coordinator high availability with Helm.

  • Helm upgrades stop both coordinators. The coordinator Deployment of the Helm chart uses the Recreate strategy, so coordinator high availability does not keep the cluster available during a Helm upgrade.

  • A stale coordinator address after a restart. After a coordinator restarts, the control plane can keep calling the old address of the coordinator for several minutes. Restart the control plane pods to fix it.

  • Internal TLS and managed mode. Do not use internalTls: true with a managed cluster.