Terraform Remote Executor Module Not Updated After dockerize Removal from Executor Images – DataHub Customer Support Portal

Area: Deployment Issues

Sub-Area: Remote Executor / Terraform Module Configuration

Issue

When DataHub removed the dockerize binary from remote executor Docker images (starting with image tag v1.1.0-cloud and later), the official Terraform module for remote executor deployment was not updated to reflect this change. The remote-ingestion-executor/main.tf file in the datahub-terraform-modules repository continued to specify dockerize as the container entrypoint command. As a result, any user deploying or upgrading the remote executor using that Terraform module with a modern executor image (v1.1.0-cloud or later) will encounter a startup failure, because dockerize no longer exists in the container image. The equivalent fix had already been applied to the CloudFormation and Helm deployment templates but was inadvertently missed in the Terraform module.

Error Messages

executable file not found in $PATH: dockerize
container failed to start: unknown command: dockerize

You Might Be Asking

Why does my remote executor container fail to start after upgrading to a newer DataHub executor image?
Why does the Terraform module still reference dockerize when it was removed from the Docker image?
How do I pin my remote executor to a specific DataHub Cloud image version using the Terraform module?
Why are there no recent release tags on the datahub-terraform-modules repository that correspond to DataHub Cloud versions like v1.1.4?

Solution

Identify the breaking change. The dockerize binary was intentionally removed from the remote executor Docker images beginning with image tag v1.1.0-cloud. The Terraform module's container command definition was not updated at the same time, leaving a stale reference to dockerize as the entrypoint.
Apply the fix to your Terraform module. In remote-ingestion-executor/main.tf, locate the container command definition (previously referencing dockerize) and update it to use the image's default entrypoint script directly. The corrected command should be:
```
command = ["/start_datahub_executor.sh"]
```
Remove any preceding reference to dockerize, so the block no longer reads:
```
# OLD — broken with images v1.1.0-cloud and later
command = ["dockerize", "/start_datahub_executor.sh"]
```
And instead reads:
```
# CORRECT — works with images v1.1.0-cloud and later
command = ["/start_datahub_executor.sh"]
```
This is the same one-line change that was already applied to the CloudFormation template.
Pull the latest main branch of the Terraform module. The official fix has been merged into the main branch of the datahub-terraform-modules repository. If you reference the module from GitHub, update your module source to pull from main:
```
module "remote_executor" {
  source = "github.com/acryldata/datahub-terraform-modules//remote-ingestion-executor?ref=main"
  # ... other variables
}
```
Pin the executor image version if needed. The Terraform module exposes an image_tag variable (defaulting to the latest supported cloud image). If you need to target a specific DataHub Cloud executor version, override this variable explicitly in your Terraform configuration:
```
module "remote_executor" {
  source = "github.com/acryldata/datahub-terraform-modules//remote-ingestion-executor?ref=main"

  datahub = {
    image_tag = "v1.1.4-cloud"   # Replace with your target DataHub Cloud version
    # ... other required fields
  }
}
```
Re-apply your Terraform configuration. After making the above changes, run:
```
terraform init -upgrade
terraform plan
terraform apply
```
Verify that the executor container starts successfully and that ingestion runs can be triggered from the DataHub UI.

Additional Notes

This issue is a confirmed regression introduced when executor images were updated to remove dockerize (starting at server version 1.1.0) without a corresponding update to the Terraform module. The CloudFormation and Helm templates were patched at the time of image change; the Terraform module was patched subsequently. If you are on an executor image older than v1.1.0-cloud, you will not encounter this issue, as those older images still included the dockerize binary.

Regarding Terraform module release tags: the datahub-terraform-modules repository does not maintain versioned release tags that correspond 1:1 to DataHub Cloud image versions (e.g., v1.1.4). The main branch is kept up to date and is the recommended reference for all deployments. The image_tag variable is the correct mechanism to control which executor image version is deployed; the module version and the image version are managed independently.

Users who deploy the remote executor via Helm charts or CloudFormation are not affected by this specific issue, as those templates were already updated.

Issue

Error Messages

You Might Be Asking

Solution

Additional Notes

Related Documentation