Slow FabricPool on StorageGRID VMs with io wait due to Pointer Block Cache
Applies to
- StorageGRID 11.x
- Virtualized StorageGRID storage nodes
- ONTAP 9
- FabricPool
Issue
- ONTAP 9 has slowness in FabricPool response times
- StorageGRID is configured to use virtual appliances on VMWare ESXI
- There are high i/o waits in the StorageGRID dashboard with 100% CPU on one node of StorageGRID
- No ONTAP, backend, or network latency or errors are seen
- Example: From the StorageGRID Support Node Grafana dashboard, high CPU and i/o wait are seen