Skip to main content
Replace the management plane, and nothing it manages

Synopsis

The Management Plane is cattle; managed clusters survive its replacement. What goes is the plane’s own cluster — its nodes, its Kubernetes, the Kubox workloads and the network they sit in — and its own Pulumi stacks. Every managed cluster keeps running. Every CloudAccount, with the External ID its customer’s role trusts, every cluster record and every agent credential is exported from the plane’s store before it is destroyed and restored byte-identical into the new one, so no customer reconnects and no cluster is rebuilt. The order is the whole point and cannot be run otherwise: export, destroy the plane, empty its stacks, create, restore, verify customer trust — the new plane assumes every restored role, and the rebuild is complete only when every one succeeds. Refused while a cluster is being deleted, and refused when the plane’s store cannot be read: what cannot be exported cannot be restored, and a rebuild that went on would break every customer’s trust. --plan classifies everything it touches and changes nothing. Requires —yes.

Options

Global options

See global options.

SEE ALSO

  • kubox plane - Describe and build a management plane from one document