RDMA connection issue causes vSAN object becoming inaccessible

Products

• VMware vSAN

Issue/Introduction

The vSAN is RDMA-enabled.

One or more objects become inaccessible caused by  the RDMA connection errors 

You may find such messages on the vSAN nodes

vmkernel: cpu44:2099908)RDT: RDTRDMAConnectionSendMbufTo:5391: connection in error

vmkernel: cpu44:2099908)RDT: RDTRDMAConnectionSendMbufTo:5391: connection in error

vmkernel: cpu44:2099908)RDT: RDTRDMAConnectionRecvInt:5617: Connection in error: 0x433c69fffa80

There are similar error message for the object

 vmkernel: cpu18:2099908)DOM: DOMOwnerUnsubscribeClusterEncrState:5925: DOM Owner on <OBJECT_UUID> received premature cluster encryption state unsubscription

 

Environment

vSAN 8.x

Resolution

Quick workaround:

Disable RDMA from Cluster - Configure - vSAN - Services - Network

Make sure FW/Driver of the network card is upgraded to the latest version.