Local unused RAID controller showing high latency spikes in performance graphs

Products

• VMware vSphere ESXi

Issue/Introduction

ESXi hosts have a local RAID controller with virtual disk configured and visible to host, but not used

ESXi performance graphs in vCenter show large latency spikes occurring either every 30 or every 60 minutes

Hosts are under light or no workload

Environment

ESXi (Any version)

Local RAID controller

Cause

The cause is under investigation from hardware vendor.

Resolution

1. Ensure the virtual disk (RAID 0, 1, 5 etc) presented by the local controller is not used in ESXI
2. Detach the virtual disk from ESXI
3. Delete the virtual disk from your RAID controller
4. Disable the controller from the out of band management (iDRAC, iLO, etc) or BIOS if desired and possible

Additional Information

Once the disk is no longer attached to ESXi the system no longer attempts to monitor and will not display the latency spikes. Removing the disk from the controller ensures it is not picked up on subsequent boot or rescan operations. Disabling the unused controller ensures no new storage is inadvertently presented to the host to be picked up.

This only applies to situations where a RAID volume is configured on a local controller and presented to the host, but remains unused (no data on it).