Skip to content

Orchestrator

Orchestrator manages MySQL replication topology and performs automatic failover. This operator bundles Orchestrator as a sidecar container in the mysql-operator StatefulSet.

What Orchestrator does

  • Discovers cluster members and their roles (primary, replica)
  • Detects primary failure and promotes a replica
  • Coordinates with the operator during planned shutdowns (--failover-before-shutdown)
  • Surfaces topology in a web UI

Each MysqlCluster registers with Orchestrator using the alias {cluster-name}.{namespace}.

Accessing the UI

Port-forward the operator Service (Orchestrator HTTP is exposed on port 80):

kubectl port-forward -n mysql-operator svc/mysql-operator 8080:80

Open http://localhost:8080 in your browser.

The Service name is always mysql-operator in the mysql-operator namespace when using default manifests — unlike Helm releases, there is no release-name prefix.

Cluster conditions

MysqlCluster status reflects Orchestrator state:

Condition Meaning
PendingFailoverAck A failover event needs acknowledgment in Orchestrator
FailoverInProgress Orchestrator is actively failing over

Check with:

kubectl describe mysql my-cluster

MySQL 8.4 compatibility

This fork ships a custom Orchestrator image built from Percona Orchestrator with MySQL 8.4 replication SQL support. See Version profiles.

Topology credentials

Orchestrator connects to MySQL using the topology user configured in mysql-operator-orc and propagated to each cluster's operated secret. Keep credentials consistent across operator upgrades — see Install the operator.