CKS cluster goes into alert state when worker node is unmanaged #12783
Description
problem
CKS cluster goes into alert state
versions
ACS 4.22 , KVM hypervisor
The steps to reproduce the bug
-
Created a K8s issue with 1 controller, and 3 workers
-
Make sure the CKS cluster is in running state
-
Navigate to one of the worker instances and unmanaged it
- The CKS cluster went into Alert state
- Delete the CKS Cluster, following error is issued
"com.cloud.vm.VMInstanceVO.getBackupOfferingId()" because "vm" is null"
logs
2026-03-10 11:39:56,869 ERROR [c.c.a.ApiAsyncJobDispatcher] (API-Job-Executor-55:[ctx-53050b87, job-73]) (logid:755e27ea) Unexpected exception while executing org.apache.cloudstack.api.command.user.kubernetes.cluster.DeleteKubernetesClusterCmd java.lang.NullPointerException: Cannot invoke "com.cloud.vm.VMInstanceVO.getBackupOfferingId()" because "vm" is null
at com.cloud.kubernetes.cluster.KubernetesClusterManagerImpl.checkIfVmsAssociatedWithBackupOffering(KubernetesClusterManagerImpl.java:2018)
- Stop the CKS cluster if the following error is issued
Failed to find all VMs in Kubernetes cluster :
logs
2026-03-10 11:40:47,436 ERROR [c.c.k.c.a.KubernetesClusterStopWorker] (API-Job-Executor-56:[ctx-37106f8f, job-74, ctx-f1370fcc]) (logid:eeffc52f) Failed to find all VMs in Kubernetes cluster : test
- Import back the unmanged instance using the api > select the same cks network
https://cloudstack.apache.org/api/apidocs-4.22/apis/importUnmanagedInstance.html
- Add the imported worker node instance back to the CKS cluster
logs
2026-03-10 12:52:22,913 DEBUG [o.a.c.f.j.i.AsyncJobManagerImpl] (API-Job-Executor-61:[ctx-3eab1e04, job-81]) (logid:fba281a3) Complete async job-81, jobStatus: FAILED, resultCode: 530, result: org.apache.cloudstack.api.response.ExceptionResponse/null/{"uuidList":[],"errorcode":"530","errortext":"Failed to add nodes to cluster ID 1 due to: No valid nodes found to be added to the Kubernetes cluster"}
Other related CKS Alert issues
What to do about it?
CloudStack should not allow to unmanage an instance if its a part of a cks cluster
or Cloudstack CKS cluster should support addition of an unmanaged instance