Red Hat turns the OpenShift node-OS jump into a pool-by-pool migration
RHEL CoreOS 10 is a Technology Preview in OpenShift 4.22; Red Hat’s planned mixed-stream model would let administrators validate and move nodes gradually.
Red Hat is previewing Red Hat Enterprise Linux CoreOS 10 in OpenShift 4.22 and has laid out a larger change to how OpenShift clusters will cross major node-operating-system generations. Instead of coupling a platform update to an all-at-once CoreOS migration, a future OpenShift release is expected to support RHEL CoreOS 9 and 10 nodes in the same cluster during the transition.
That distinction matters: the RHEL CoreOS 10 image is available now as a Technology Preview, while the supported mixed-stream operating model is a future capability. Platform teams should test the destination today without treating the migration machinery as generally available.
What changes
Red Hat describes a new cluster-level OSStreams API that will let the Machine Config Operator manage more than one RHEL CoreOS stream. Administrators will be able to assign a custom MachineConfigPool—down to a single validation node—to the RHEL CoreOS 10 stream, run workloads there, and then expand the migration pool by pool.
Existing nodes upgraded from OpenShift 4.22 are expected to remain on RHEL CoreOS 9 rather than being forced onto the new operating system immediately. Red Hat says both streams will be supported during the transition, giving organizations time to complete hardware certification and workload validation.
The operational model deliberately reuses OpenShift’s existing node-configuration mechanics: cordon, drain, reboot and reconcile. Red Hat also says a node can return to the earlier stream if validation fails, avoiding a separate rollback system.
Why teams may move
RHEL CoreOS 10 inherits the RHEL 10 kernel and its newer hardware enablement for GPUs, network adapters, NVMe and CXL devices. Red Hat also points to sealed image mode with fs-verity-based continuous integrity verification and access to post-quantum cryptography work from the RHEL 10 base.
Those are useful incentives, but the migration design is the more consequential platform-engineering change. It separates the decision to update OpenShift from the decision to certify a new major host operating system.
What to do now
Administrators can evaluate the Technology Preview on a non-production OpenShift 4.22 cluster by creating a custom MachineConfigPool with one node and pointing it at the RHEL CoreOS 10 stream. Teams using on-cluster layering should test their MachineOSConfig builds against both RHEL CoreOS 9 and 10 bases, because those configurations are scoped to machine pools.
Production plans should wait for general availability and final support documentation. In the meantime, the preview gives platform teams a concrete way to discover driver, layered-image and workload assumptions before the mixed-stream migration path becomes a supported operating model.
sources
comments · 0