Closed utkuozdemir closed 4 months ago
I think we should have an integration test for it.
Also we need to skip deleting nodes which are Ready?
already doing the Ready
check here:https://github.com/siderolabs/omni/pull/150/files#diff-f856c47e45cecba4018f26b99f4529186e842968b8a8a6acf6201f6bbaec9d61R230-R234
and is tested here: https://github.com/siderolabs/omni/pull/150/files#diff-181a0e188b1644c2cabab30214f26546f614735afbd6f6bf47aa0616b24d1991R84-R95 & https://github.com/siderolabs/omni/pull/150/files#diff-181a0e188b1644c2cabab30214f26546f614735afbd6f6bf47aa0616b24d1991R108-R111
Will add an integration test.
Just remembered: we also need to drop MachineSet
dependency in the cluster machine teardown controller.
But please run all e2e tests after that: especially e2e scaling.
/m
Clean up invalid nodes (from the Kubernetes perspective) from the cluster if they are no longer part of a cluster from the perspective of Omni.
Introduce a new intermediate resource,
ClusterKubernetesNodes
, containing the ordered list of nodes in a cluster.Create the resource,
KubernetesNodeAuditResult
, to provide a summary on the latest deleted nodes from the Kubernetes cluster.Closes siderolabs/omni#9.