Announcement: Analytics Environment Dataproc Cluster Upgrade Requirement (2026-07-31)
Google Cloud has announced upcoming changes to Dataproc cluster image support that will affect some Safe Haven Analytics Environment customers.
What Is Changing
Beginning August 25, 2026:
Image End-of-Life: Compute Engine image versions 1.x and 2.0 will reach end-of-life, and new clusters can no longer be created with those images.
Deletion Risk for Existing Clusters: Existing clusters running on 1.x or 2.0 image versions that are not upgraded to a supported sub-minor version (2.1 or higher) may stop working and could be automatically deleted by Google Cloud as early as September 15, 2026.
Based on our review of GCP projects, some Dataproc clusters are currently running on older image versions, including 1.3, 1.4, and 2.0, and will be impacted by this change.
Required Actions Before August 25, 2026
Review your Dataproc clusters and scheduled jobs in Job Management:
Dataproc Clusters in Use
Clusters running on image versions 1.x and 2.0 still in use: Contact your LiveRamp representative to create a new cluster using a supported image version (2.1 or later), then migrate and validate your workflows on the new cluster before decommissioning the old one.
Active Scheduled Jobs (currently enabled in Job Management)
Python jobs: Validate your code and migrate active scheduled jobs to Python 3.10 or 3.11.
PySpark jobs: Validate your code and migrate active scheduled jobs to Dataproc image version 2.1 or later.
New scheduled jobs: Ensure all newly created Python jobs specify Python 3.10 or later, and all newly created PySpark jobs specify Dataproc image version 2.1 or later.
Inactive Scheduled Jobs (currently disabled in Job Management)
Potential reactivation issues: Inactive jobs configured with image versions 1.x, 2.0, or Python versions earlier than 3.10 may fail if re-enabled without changes. Validate and update any disabled scheduled jobs before re-enabling them.