fix: final tweaks for 1.21 upgrade

containers/admin/v1.21/README.md

@@ -0,0 +1,25 @@
+! Ensure your Kube context points to the correct cluster !
+
+# Trigger the cluster upgrade
+`./upgrade_121.sh`
+
+# Upgrade CFN stacks for the control plane and all worker groups
+
+# Reboot controller(s) one by one
+
+# Patch current deployments, blocking ArgoCD otherwise
+`./kubezero_121.sh`
+
+# Migrate ArgoCD config for the cluster
+`./migrate_argo.sh "cluster/env/kubezero/application.yaml"`
+
+Adjust as needed, eg. ensure eck-operator is enabled if needed.
+git add / commit / push
+
+Watch ArgoCD do its work.
+
+# Replace worker nodes 
+
+## Known issues
+- pods seem stuck, eg. fluent-bit on workers shows NotReady *after* control nodes have been ugpraded  
+  -> restart `kube-proxy` on the affected workers

containers/admin/v1.21/kubezero_121.sh

@@ -6,8 +6,9 @@ set -x
 kubectl patch application aws-ebs-csi-driver -n argocd --type=json -p='[{"op": "remove", "path": "/metadata/finalizers"}]'
 kubectl patch application aws-efs-csi-driver -n argocd --type=json -p='[{"op": "remove", "path": "/metadata/finalizers"}]'
 
-# Migrate ZeroDownTime helm repo
+# Migrate ZeroDownTime helm repo and fix project permissions
 kubectl patch appproject kubezero -n argocd --type=json -p='[{"op": "replace", "path": "/spec/sourceRepos/0", "value": "https://cdn.zero-downtime.net/charts" }]'
+kubectl patch appproject kubezero -n argocd --type=json -p='[{"op": "replace", "path": "/spec/destinations", "value": [{"namespace": "*", "server": "https://kubernetes.default.svc"}] }]'
 
 # Delete EBS and EFS Deployments and Daemonsets as we cannot change the lables while moving them to storage.
 # This will NOT affect provisioned volumes