diff --git a/proxmox/terraform/staging/cluster-graphql.tf b/proxmox/terraform/staging/cluster-graphql.tf
index e91c327..f506d8c 100644
--- a/proxmox/terraform/staging/cluster-graphql.tf
+++ b/proxmox/terraform/staging/cluster-graphql.tf
@@ -1,159 +1,119 @@
 # This declares terraform manifests to provision vms and register containers within
 # those to a rancher (clusters management service) instance.
 
 # Each software has the following responsibilities:
 # - proxmox: provision vms (with docker dependency)
 # - rancher: installs kube cluster within containers (running on vms)
 
 # Requires (RANCHER_ACCESS_KEY and RANCHER_SECRET_KEY) in your shell environment
 # $ cat ~/.config/terraform/swh/setup.sh
 # ...
 # key_entry=operations/rancher/azure/elastic-loader-lister-keys
 # export RANCHER_ACCESS_KEY=$(swhpass ls $key_entry | head -1 | cut -d: -f1)
 # export RANCHER_SECRET_KEY=$(swhpass ls $key_entry | head -1 | cut -d: -f2)
 
 # Plan:
 # - create cluster with terraform
 # - Create nodes as usual through terraform
 # - Retrieve the registration command (out of the cluster creation step) to provide new
 #   node
 
 resource "rancher2_cluster" "cluster-graphql" {
   name = "cluster-graphql"
   description = "graphql staging cluster"
   rke_config {
     network {
       plugin = "canal"
     }
   }
 }
 
 output "rancher2_cluster_graphql_summary" {
   sensitive = true
   value = rancher2_cluster.cluster-graphql.kube_config
 }
 
 output "rancher2_cluster_graphql_command" {
   sensitive = true
   value = rancher2_cluster.cluster-graphql.cluster_registration_token[0].node_command
 }
 
 module "graphql-worker0" {
   source      = "../modules/node"
   vmid        = 162
   template    = var.templates["stable-zfs"]
   config      = local.config
   hostname    = "graphql-worker0"
   description = "elastic worker running in rancher cluster"
   hypervisor  = "uffizi"
   sockets     = "1"
   cores       = "4"
   onboot      = true
   memory      = "8192"
   balloon     = "4096"
 
   networks = [{
     id      = 0
     ip      = "192.168.130.150"
     gateway = local.config["gateway_ip"]
     bridge  = "vmbr443"
   }]
 
   storages = [{
     storage = "proxmox"
     size    = "20G"
     }, {
     storage = "proxmox"
     size    = "50G"
     }
   ]
 
   post_provision_steps = [
     "systemctl restart docker",  # workaround
     "${rancher2_cluster.cluster-graphql.cluster_registration_token[0].node_command} --etcd --controlplane --worker"
   ]
 }
 
 output "graphql-worker0_summary" {
   value = module.graphql-worker0.summary
 }
 
 module "graphql-worker1" {
   source      = "../modules/node"
   vmid        = 163
   template    = var.templates["stable-zfs"]
   config      = local.config
   hostname    = "graphql-worker1"
   description = "graphql worker running in rancher cluster"
   hypervisor  = "uffizi"
   sockets     = "1"
   cores       = "4"
   onboot      = true
   memory      = "8192"
   balloon     = "4096"
 
   networks = [{
     id      = 0
     ip      = "192.168.130.151"
     gateway = local.config["gateway_ip"]
     bridge  = "vmbr443"
   }]
 
   storages = [{
     storage = "proxmox"
     size    = "20G"
     }, {
     storage = "proxmox"
     size    = "50G"
     }
   ]
 
   post_provision_steps = [
     "systemctl restart docker",  # workaround
     "${rancher2_cluster.cluster-graphql.cluster_registration_token[0].node_command} --etcd --controlplane --worker"
   ]
 }
 
 output "graphql-worker1_summary" {
   value = module.graphql-worker1.summary
 }
-
-module "graphql-worker2" {
-  source      = "../modules/node"
-  vmid        = 164
-  template    = var.templates["stable-zfs"]
-  config      = local.config
-  hostname    = "graphql-worker2"
-  description = "graphql worker running in rancher cluster"
-  hypervisor  = "uffizi"
-  sockets     = "1"
-  cores       = "4"
-  onboot      = true
-  memory      = "8192"
-  balloon     = "4096"
-
-  networks = [{
-    id      = 0
-    ip      = "192.168.130.152"
-    gateway = local.config["gateway_ip"]
-    bridge  = "vmbr443"
-  }]
-
-  storages = [{
-    storage = "proxmox"
-    size    = "20G"
-    }, {
-    storage = "proxmox"
-    size    = "50G"
-    }
-  ]
-
-  post_provision_steps = [
-    "systemctl restart docker",  # workaround
-    "${rancher2_cluster.cluster-graphql.cluster_registration_token[0].node_command} --etcd --controlplane --worker"
-  ]
-}
-
-output "graphql-worker2_summary" {
-  value = module.graphql-worker2.summary
-}
diff --git a/proxmox/terraform/staging/cluster-graphql2.tf b/proxmox/terraform/staging/cluster-graphql3.tf
similarity index 56%
rename from proxmox/terraform/staging/cluster-graphql2.tf
rename to proxmox/terraform/staging/cluster-graphql3.tf
index 9004c4b..f571ccb 100644
--- a/proxmox/terraform/staging/cluster-graphql2.tf
+++ b/proxmox/terraform/staging/cluster-graphql3.tf
@@ -1,80 +1,120 @@
 # This declares terraform manifests to provision vms and register containers within
 # those to a rancher (clusters management service) instance.
 
 # Each software has the following responsibilities:
 # - proxmox: provision vms (with docker dependency)
 # - rancher: installs kube cluster within containers (running on vms)
 
 # Requires (RANCHER_ACCESS_KEY and RANCHER_SECRET_KEY) in your shell environment
 # $ cat ~/.config/terraform/swh/setup.sh
 # ...
 # key_entry=operations/rancher/azure/elastic-loader-lister-keys
 # export RANCHER_ACCESS_KEY=$(swhpass ls $key_entry | head -1 | cut -d: -f1)
 # export RANCHER_SECRET_KEY=$(swhpass ls $key_entry | head -1 | cut -d: -f2)
 
 # Plan:
 # - create cluster with terraform
 # - Create nodes as usual through terraform
 # - Retrieve the registration command (out of the cluster creation step) to provide new
 #   node
 
-resource "rancher2_cluster" "cluster-graphql2" {
-  name = "cluster-graphql2"
-  description = "2nd graphql staging cluster"
+resource "rancher2_cluster" "cluster-graphql3" {
+  name = "cluster-graphql3"
+  description = "3rd (tryout) graphql staging cluster"
   rke_config {
-    kubernetes_version = "v1.22.10-rancher1-1"
+    kubernetes_version = "v1.21.12-rancher1-1"
     network {
       plugin = "canal"
     }
   }
 }
 
-output "rancher2_cluster_graphql2_summary" {
+output "rancher2_cluster_graphql3_summary" {
   sensitive = true
-  value = rancher2_cluster.cluster-graphql2.kube_config
+  value = rancher2_cluster.cluster-graphql3.kube_config
 }
 
-output "rancher2_cluster_graphql2_command" {
+output "rancher2_cluster_graphql3_command" {
   sensitive = true
-  value = rancher2_cluster.cluster-graphql2.cluster_registration_token[0].node_command
+  value = rancher2_cluster.cluster-graphql3.cluster_registration_token[0].node_command
 }
 
 module "graphql-worker3" {
   source      = "../modules/node"
   vmid        = 165
   template    = var.templates["stable-zfs"]
   config      = local.config
   hostname    = "graphql-worker3"
   description = "graphql worker running in rancher cluster"
   hypervisor  = "uffizi"
   sockets     = "1"
   cores       = "4"
   onboot      = true
   memory      = "16384"
   balloon     = "8192"
 
   networks = [{
     id      = 0
     ip      = "192.168.130.153"
     gateway = local.config["gateway_ip"]
     bridge  = "vmbr443"
   }]
 
   storages = [{
     storage = "proxmox"
     size    = "20G"
     }, {
     storage = "proxmox"
     size    = "50G"
     }
   ]
 
   post_provision_steps = [
     "systemctl restart docker",  # workaround
-    "${rancher2_cluster.cluster-graphql2.cluster_registration_token[0].node_command} --etcd --controlplane --worker"
+    "${rancher2_cluster.cluster-graphql3.cluster_registration_token[0].node_command} --etcd --controlplane --worker"
   ]
 }
 
 output "graphql-worker3_summary" {
   value = module.graphql-worker3.summary
 }
+
+# module "graphql-worker2" {
+#   source      = "../modules/node"
+#   vmid        = 164
+#   template    = var.templates["stable-zfs"]
+#   config      = local.config
+#   hostname    = "graphql-worker2"
+#   description = "graphql worker running in rancher cluster"
+#   hypervisor  = "uffizi"
+#   sockets     = "1"
+#   cores       = "4"
+#   onboot      = true
+#   memory      = "8192"
+#   balloon     = "4096"
+
+#   networks = [{
+#     id      = 0
+#     ip      = "192.168.130.152"
+#     gateway = local.config["gateway_ip"]
+#     bridge  = "vmbr443"
+#   }]
+
+#   storages = [{
+#     storage = "proxmox"
+#     size    = "20G"
+#     }, {
+#     storage = "proxmox"
+#     size    = "50G"
+#     }
+#   ]
+
+#   post_provision_steps = [
+#     "systemctl restart docker",  # workaround
+#     "${rancher2_cluster.cluster-graphql3.cluster_registration_token[0].node_command} --controlplane"
+#   ]
+# }
+
+# output "graphql-worker2_summary" {
+#   value = module.graphql-worker2.summary
+# }