I’m using the native/Google-managed Backup and DR Service (Cloud console backup plans and backup vaults, not the appliance-based setup). I want to back up complete Compute Engine instances (VM config + all attached disks), not just individual disks.
My instances use Hyperdisk Balanced disks. Backing up the disks standalone via backup vaults works fine, and that’s what we’re using today as a workaround. But full-instance backup fails with:
Invalid resource state for "projects/<PROJECT>/zones/europe-west4-b/instances/<ID>": VM has disk(persistent-disk-0) with disk type ``https://www.googleapis.com/compute/v1/projects/``<PROJECT>/zones/europe-west4-b/diskTypes/hyperdisk-balanced
This matches the constraint listed here: Back up Compute Engine instances – Constraints and limitations, which still says (as of today):
Instances using any hyperdisk-* disk types [are not supported]
However, a different page (About Backup and DR Service backup plans) lists a narrower restriction, only excluding Hyperdisk Extreme, implying Hyperdisk Balanced should already work for full-instance backup.
These two official pages currently contradict each other, and my real-world test confirms the more restrictive behavior (all hyperdisk-* types blocked) is what’s actually enforced.
What I need clarified:
- Which page is correct? Is full-instance Google-managed backup for Hyperdisk Balanced-attached VMs supported today, or not?
- If not supported yet, is there a roadmap or timeline for adding Hyperdisk-* (espacially hyperdisk-balanced) support for full-instance backups?
- Is disk-only backup (our current workaround) expected to remain the only option for Hyperdisk-backed VMs indefinitely, or is full-instance support actively being worked on?
This is blocking us from consolidating onto one backup method (VM-level) for our fleet, since we have a mix of Persistent Disk and Hyperdisk Balanced instances today. Any confirmation from Google (official roadmap, timeline, or even just “not yet planned”) would help us decide whether to keep the disk-only workaround long-term or wait.
Thanks!