Inconsistent Subscription Ratio Displayed on Horizon Dashboard for NetApp Backend (OpenStack Cinder)
Applies to
- Cinder NetApp Backend (ONTAP driver)
- Cinder with multiple controller nodes
Issue
When using OpenStack Cinder with NetApp ONTAP as the backend, the subscription ratio (overprovisioning ratio) is displayed inconsistently across different compute/controller nodes on the Horizon dashboard. Each node shows a different ratio for the same NetApp backend volume pool, resulting in confusion and risk of overprovisioning.
Example Log Output:
DEBUG cinder.scheduler.filters.capacity_filter [...] Space information for volume creation on host hyd03ctrl01@cinder-netapp#OS_G01_NA01_POOL1 (requested/avail): 600/972.8 backend_passes DEBUG cinder.scheduler.host_manager [...] Consumed 600GB from backend: host 'hyd03ctrl02@cinder-netapp#OS_G01_NA01_POOL1': free_capacity_gb: 372.79, total_capacity_gb: 972.8, allocated_capacity_gb: 600, max_over_subscription_ratio: 1.0, provisioned_capacity_gb: 600 WARNING cinder.scheduler.filters.capacity_filter [...] Insufficient free space for thin provisioning. The ratio of provisioned capacity over total capacity 1.13 has exceeded the maximum oversubscription ratio 1.00 on host hyd03ctrl03@cinder-netapp#OS_G01_NA01_POOL1. ERROR cinder.scheduler.flows.create_volume [...] No valid backend was found. No weighed backends available.
Symptoms:
- Horizon dashboard shows different subscription ratios per backend per node.
- Volume creation may succeed even when total backend capacity is exceeded (if reporting is inconsistent).
- Overprovisioning risk when using
max_over_subscription_ratio=1.0.
