System Ingestion Jobs Failing to Download acryl-datahub Wheel for a Specific Version – DataHub Customer Support Portal

Area: Ingestion Issues

Sub-Area: Remote Executor / Virtual Environment Setup

Issue

All system ingestion jobs fail immediately during the virtual environment setup phase, before any source-specific ingestion logic runs. The remote executor attempts to download and install a specific version of the acryl-datahub CLI wheel to create an isolated Python environment for the ingestion subprocess. When the requested wheel version is unavailable on the configured package index — whether due to a version mismatch in the server-side configuration, a release coordination gap, or the package index being unreachable — every ingestion job on that executor instance hard-fails with no fallback behavior. Because the defaultCliVersion is resolved server-side and applied globally, no individual ingestion source configuration can work around this failure.

Error Messages

RuntimeError: Failed to download acryl-datahub wheel for version '<version>'
acryl.executor.execution.task.TaskError: Failed to set up virtual environment: Failed to download acryl-datahub wheel for version '<version>'

You Might Be Asking

Why are all of my ingestion sources failing at the same time with a venv setup error?
Why does the executor fail before any ingestion logic even starts?
What does "Failed to download acryl-datahub wheel" mean and how do I fix it?
Can I configure a different CLI version at the individual source level to work around this?
Is my executor version (1!0.x.y.z+docker) causing the wheel download failure?

Solution

The following steps identify the root cause and restore normal ingestion operation.

Confirm the failure scope.
Check whether all ingestion sources are failing with the same TaskError: Failed to set up virtual environment error. If so, the problem is in the global defaultCliVersion setting, not in any individual source recipe. The executor version (e.g., 1!0.x.y.z+docker) shown in the logs is the executor process version and is a separate value — it is not the cause of the wheel download failure.
Identify the CLI version being requested.
In the execution summary logs, locate the version string referenced in the error:
```
RuntimeError: Failed to download acryl-datahub wheel for version '<version>'
```
This is the value that was resolved from the server-side defaultCliVersion configuration and passed as the pip requirement acryl-datahub==<version>.
Verify wheel availability on the configured package index.
Depending on your deployment, the executor resolves wheels from either PyPI or a private/internal package registry. Confirm that the requested version is present and not yanked on whichever index your environment uses:
- For PyPI: https://pypi.org/project/acryl-datahub/#history
- For a private registry (e.g., an internal ECR or Artifactory mirror): check with your platform or infrastructure team.
If the version does not exist on the index your executor uses, that is the direct cause of the failure.
Update defaultCliVersion to a valid, published version.
The defaultCliVersion is set server-side in the managed ingestion configuration. Update it to a version that is confirmed to be available on your package index. For DataHub Cloud deployments, contact DataHub Support to apply this change. For self-hosted deployments, update the value in your Helm values file:
```
# Example Helm values snippet (self-hosted deployments)
datahub:
  ingestion:
    defaultCliVersion: "<latest-stable-version>"  # e.g., "1.1.0.4" or newer
```
Replace <latest-stable-version> with the latest confirmed-available release. After updating, redeploy the affected services so the new value takes effect.
Verify system jobs resume successfully.
After the configuration update and redeployment, trigger or wait for the next scheduled ingestion run and confirm that the execution summary no longer contains TaskError: Failed to set up virtual environment. All sources using the default CLI version should recover automatically.

Additional Notes

There is no client-side workaround for this failure. The defaultCliVersion is resolved globally on the server and applied to all ingestion jobs using the default version. Individual source recipes cannot override it to bypass a missing wheel.
The executor has no fallback behavior if the requested wheel version is unavailable — the venv setup fails hard, which causes every ingestion job on that executor instance to fail immediately.
For DataHub Cloud (managed) deployments, do not attempt to modify executor infrastructure or package registries directly. Open a support ticket so the DataHub team can update the server-side defaultCliVersion to a valid release and coordinate any required redeployment.
This failure mode can surface as a result of a release coordination gap where the platform configuration is updated to reference a new CLI version before the corresponding wheel has been published to (or made available on) the configured package index.
The executor process version (displayed with a PEP 440 epoch prefix, e.g., 1!0.x.y.z+docker) is independent of the CLI wheel version and does not contribute to this error.
If your environment uses a private/internal package mirror rather than PyPI, ensure the mirror has been updated to include the target wheel version before updating defaultCliVersion.

Issue

Error Messages

You Might Be Asking

Solution

Additional Notes

Related Documentation