Skip to content

Latest commit

 

History

History
53 lines (39 loc) · 2.5 KB

upgrading_to_v14.0.md

File metadata and controls

53 lines (39 loc) · 2.5 KB

Upgrading to v14.0

The v14.0 release of kubernetes-engine is a backwards incompatible release for some versions of Anthos Service Mesh (ASM) and includes some variable changes.

registry_project_id removed

The registry_project_id variable has been replaced with a registry_project_ids list.

 module "gke" {
   source                  = "terraform-google-modules/kubernetes-engine/google"
-  version                 = "~> 13.0"
+  version                 = "~> 14.0"

-  registry_project_id  = "my-project-id"
+  registry_project_ids = ["my-project-id"]
}

network_policy disabled by default

The network_policy variable is now false by default (instead of true). If you want to keep using the network policy addon for your cluster, make sure that the network_policy variable is set to true:

module "gke" {
   source                  = "terraform-google-modules/kubernetes-engine/google"
-  version                 = "~> 13.0"
+  version                 = "~> 14.0"

+  network_policy = true
}

ASM default version changed to 1.8

ASM submodule has been changed to use ASM v1.8 as default.

The module now uses the new ASM installation script which:

  • Does not support installation and upgrades for ASM versions older than 1.7.3.
  • Supports upgrades only from versions 1.7.3+ or a 1.8 patch release.
  • Supports migrations from open source Istio 1.7 or 1.8 to ASM

Please see the script page for up to date details.

GKE Hub Register & Unregister behaviour has changed

The Hub submodule now supports registering a cluster to a Hub that is in a separate project. This is via the introduction of the hub_project_id. variable. If you specify this variable, the cluster will be registered to this project and the GKE cluster will be deployed in the project specified in the project_id variable.

To upgrade to the latest version, you will need to remove the state for the run_destroy_command[0] resource because, as of this release we register / unregister clusters using the --gke-uri option.

If you run into errors during upgrade, you can remove the state for the run_destroy_command resource by running: terraform state rm module.hub.module.gke_hub_registration.null_resource.run_destroy_command[0]