Red Hat Advanced Cluster Management (RHACM) hub upgrades traditionally carry high risk due to extended maintenance windows and potential failures affecting hundreds of managed clusters. The multicluster global hub introduces a managed cluster migration feature that enables parallel hub deployment: deploy a new RHACM hub at the target version, gradually migrate managed clusters in batches, validate each batch, and decommission the old hub only after full validation. The migration architecture transfers secrets, ConfigMaps, BareMetalHost resources, and cluster registration objects in a specific order to prevent race conditions. Built-in automatic rollback restores clusters to the source hub if any phase fails. Real-world testing with 300 ZTP-managed SNO clusters showed 100% success rate, ~9 minutes total migration time for RHACM 2.13→2.15, and zero cluster downtime.
Table of contents
The solution: Parallel hub deployment with cluster migrationHow it works: The migration architectureAutomatic migrationPrerequisitesThe migration processPerformance at scale: Real-world resultsBuilt-in safety: Automatic rollbackWrap upSort: