Skip to content

Add MachineCreationTimeout to IroncoreMetalMachine #90

@defo89

Description

@defo89

Summary

In case a node fails to provision or kubeadm fails to init, a Machine will be stuck in Provisioned forever until this is manually mitigated.

› kubectl get machine kcp-cluster-6k7ws
NAME                   CLUSTER      NODENAME   PROVIDERID                                  PHASE         AGE   VERSION
kcp-cluster-6k7ws      cluster                 metal://capi-runtime/kcp-cluster-6k7ws      Provisioned   17h   v1.31.5

Gardener MCM provides a good guideline we could use in Cluster-API case (others are good, too):

  • MachineCreationTimeout: Amount of time after which a machine creation is declared Failed and the machine is replaced by the MachineSet controller.

We could explore adding and handling Machine deletion that have reached MachineCreationTimeout.

Metadata

Metadata

Assignees

No one assigned

    Labels

    area/metal-automationAutomation processes within the Metal project.enhancementNew feature or request

    Projects

    Status

    Backlog

    Status

    No status

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions