Performance issues and occurrences of Lost Access to Volume

Products

• VMware vSAN

Issue/Introduction

• VSAN hybrid OSA cluster having performance issues on a VM  and host level, as well as observed the following messages "Lost access to volume" messages on the ESXi hosts in this cluster.
• Experiencing high latency, in vSAN (OSA),where VM performance impact is observed as well.  When impacted customer's may observe intermittent performance drops, slowness at the application layer, and cluster wide performance impact.

Environment

VMware vSAN 8.0.x

Cause

• If the NIC teaming and failover settings on the vSAN distributed port group is set to “Route based on physical nic load” with both NICs being in the Active Uplinks section. 
• Each vmnic is going through a different physical switch upstream.

Resolution

Change  the  NIC teaming and failover settings on the vSAN distributed port group from “Route based on physical nic load” with both NICs being in the Active Uplinks section to the vSAN distributed port group load balancing configuration to “Use explicit failover order” with Uplink 1 in the Active uplink section and the Uplink 2 in the Standby uplink section.