Known issues

This page provides known issues that you might run into while using Manufacturing Data Engine (MDE), and potential workarounds until they are tackled in a future MDE release. Read them carefully as they might affect your MDE deployment.

Enable GKE cluster filestore driver

Status: Active.

MDE versions affected: 1.40 and earlier with existing GKE clusters.

Root cause: As of 1.4.0, the MDE GKE cluster uses a filestore instance to store JVM heap dumps which enables the engineering team to troubleshoot customer issues more effectively. Due to Terraform restrictions, this add-on might not be enabled for the autopilot clusters of earlier versions.

Workaround: If you experience that the pods are not starting up and they are in the ContainerCreating state, then you need to enable the CSI driver for the Filestore using the following command:

gcloud container clusters update CLUSTER_NAME \
   --update-addons=GcpFilestoreCsiDriver=ENABLED --region REGION_NAME

MDE deployment requires Object-level ACLs enabled in Cloud Storage

Status: Active. To be fixed in MDE 1.5.3

Issue 529102892: Terraform deployment fails with the error

   The GCP Organizational policy "constraints/storage.uniformBucketLevelAccess" prevented
   to proceed with Manufacturing Data Engine platform deployment with this error "│
   Error: googleapi: Error 412: Request violates constraint
   'constraints/storage.uniformBucketLevelAccess', conditionNotMet"

MDE versions affected: 1.3.x to 1.5.2

Root cause: Newly created Google Cloud orgs enforce storage.uniformBucketLevelAccess organization policy constraint, which clashes with the Cloud Storage fine grained access bucket requirements for MDE.

Workaround: Ask the org administrator to disable the storage.uniformBucketLevelAccess constraint for the MDE project in the https://console.cloud.google.com/iam-admin/orgpolicies/list.

Timeout during import of configuration packages

Status: Active. To be fixed in MDE 1.5.3

Issue 503724179: When importing >20 types in one MDE configuration package, the process fails.

MDE versions affected: 1.5.x

Root cause: Importing a big number of types causes MDE to hit a 3 min timeout.

Workaround: Split the configuration package in smaller packages and import them sequentially.

Grafana module not foundwhen running terraform init

Status: Fixed in MDE 1.5.1.

Issue 435653743: The terraform grafana module is disabled by default in MDE, however, during MDE deployment or upgrade, terraform init will attempt to load it, generating the "Grafana module not found" error.

MDE versions affected: 1.3.x to 1.5.0.

Root cause: The repository that contained the helm-grafana terraform module has been removed by the owner. We are looking into alternatives for those customers interested in the integration.

Workaround: Rename the suffix of the module file with the following command, so terraform init doesn't try to load it:

  mv modules/deployment/helm-grafana.tf modules/deployment/helm-grafana.disabled_tf

Alternatively, if you are not planning on using Grafana with MDE, you can remove helm-grafana.tf with the following command:

  rm modules/deployment/helm-grafana.tf