ORM Apply fails with Error: Waiting for rollout to finish: 3 replicas wanted; 0 replicas Ready
andytael opened this issue · 4 comments
I tried deploying MuShop using "Deploy to Oracle Cloud" button (https://cloud.oracle.com/resourcemanager/stacks/create?zipUrl=https://github.com/oracle-quickstart/oci-cloudnative/releases/latest/download/mushop-stack-latest.zip) and the Apply job failed with the following error:
2021/12/09 16:31:27[TERRAFORM_CONSOLE] [INFO] kubernetes_deployment.cluster_autoscaler_deployment[0]: Still creating... [9m40s elapsed] 2021/12/09 16:31:37[TERRAFORM_CONSOLE] [INFO] kubernetes_deployment.cluster_autoscaler_deployment[0]: Still creating... [9m50s elapsed] 2021/12/09 16:31:47[TERRAFORM_CONSOLE] [INFO] 2021/12/09 16:31:47[TERRAFORM_CONSOLE] [INFO] Error: Waiting for rollout to finish: 3 replicas wanted; 0 replicas Ready 2021/12/09 16:31:47[TERRAFORM_CONSOLE] [INFO] 2021/12/09 16:31:47[TERRAFORM_CONSOLE] [INFO] with kubernetes_deployment.cluster_autoscaler_deployment[0], 2021/12/09 16:31:47[TERRAFORM_CONSOLE] [INFO] on oke-autoscaler.tf line 190, in resource "kubernetes_deployment" "cluster_autoscaler_deployment": 2021/12/09 16:31:47[TERRAFORM_CONSOLE] [INFO] 190: resource "kubernetes_deployment" "cluster_autoscaler_deployment" { 2021/12/09 16:31:47[TERRAFORM_CONSOLE] [INFO] 2021/12/09 16:31:47[TERRAFORM_CONSOLE] [INFO]
That's a known issue generated because of the change on the autoscaler images tags. On the current stack, I recommend not to use Cluster Autoscaler if using version Kubernetes 1.20+ or latest.
A fix will be created by this EOW.
Fixes delayed for next week
The work around is to disable the cluster autoscaler if want to use 1.20 or 1.21 (latest) or extract the zip, change the file ole-autoscaler.tf line 7 and change the cluster_autoscaler_image_version to 1-1 (if 1.21) or 0-4 (if 1.20) and change the line 8 (col 132) to use "." instead of "-" for the version separator (will looks like this: .../oci-cluster-autoscaler:${local.k8s_major_minor_version}.${local.cluster_autoscaler_image_version})
This issue is stale because it has been open 14 days with no activity. Remove stale label or comment or this will be closed in 10 days.