HI, Env: RKE2 managed downstream cluster, with the local UCM having CAPI controllers. Rancher/UCM version 2.12.3, downstream Kubernetes version 1.33.9, using the vSphere provider for VM provisioning.
Couple of questions:
• Is it possible to use multiple cloud credentials within the same cluster? My use case is to have existing node pools using one cloud credential to provision VMs in one vCenter/zone, and then create additional node pools using a different cloud credential to provision VMs in another vCenter/zone.
• Also, regarding
Kubernetes upgrades: we have Rancher-managed RKE2 downstream clusters with 300+ VMs. The clusters are provisioned via CAPI and managed through Helm charts from the Rancher local cluster. Upgrades are triggered by updating the Kubernetes version in the Helm chart, and the rollout does not appear to be controlled by System Upgrade Controller.
Is there a supported way to pause an in-progress upgrade and resume it later? Our use case is to stop the upgrade at the end of a maintenance window and continue during the next non-business-hours window, rather than allowing it to run through business hours.
rkecontrolplane.rke.cattle.io. ---> holds the kubernetes version