-
Bug
-
Resolution: Unresolved
-
Normal
-
None
-
Assisted Installer Docs Latest
-
None
-
False
-
-
False
-
-
Description of problem:
The original set of worker nodes that are created via the Assisted Installer are part of a worker MachineSet. We do not document this MachineSet or how it should be managed for node additions or node deletations.
1) If you add new worker nodes day-2, they are not part of the MachineSet.
2) If you lose or remove one of the original worker nodes, the MachineSet will have a Machine stuck in the Provisioning state and Desired will not be equal to Ready/Available.
For the control plane, we document how to properly create a BMH and Machine object and link it to a newly provisioned node.
If we default to creating a MachineSet, we should document how to properly use it to scale the cluster up or down and link newly created worker nodes to the Provisioning Machine.
Our current documentation
MachineSets: NAME DESIRED CURRENT READY AVAILABLE os-qwertyui-zxcvb-worker-0 4 4 4 4
Compared to the MCP with new nodes added:
MachineConfigPools: NAME CONFIG PAUSED UPDATED UPDATING DEGRADED MACHINECOUNT READYMACHINECOUNT UPDATEDMACHINECOUNT DEGRADEDMACHINECOUNT master rendered-master-qaz False True False False 3 3 3 0 worker rendered-worker-qaz False True False False 20 20 20 0
Version-Release number of selected component (if applicable):
4.17/4.18
How reproducible:
Always
Steps to Reproduce:
1. Create a new assisted installer cluster on bare metal
2. Add new nodes day 2
3. No docs on removing nodes via the MachineSet or otherwise
Actual results:
Customer confusion
Expected results:
Properly document how this should be achieved.
Additional info:
Doc: https://docs.redhat.com/en/documentation/assisted_installer_for_openshift_container_platform/2025/html-single/installing_openshift_container_platform_with_the_assisted_installer/index#additional_resources_6
