VMs crash and Nutanix NFS Datastore enters All Paths Down (APD) state following ESXi upgrade to 8.0 U3k

Products

• VMware vSphere ESXi

Issue/Introduction

Following an upgrade to VMware vSphere ESXi 8.0 U3k, virtual machines (VMs) running on Nutanix clusters may crash or become unresponsive. This is typically accompanied by the Nutanix NFS datastore entering an All Paths Down (APD) state.

• Multiple VMs crash within 6 hours of an ESXi upgrade.
• Nutanix NFS datastores show as "Not Accessible" or "Disconnected" in vCenter Server.
• /var/log/vmkernel.log contains APD start and timeout events: StorageApdHandlerEv: Device or filesystem with identifier [####-####] has entered the All Paths Down state. StorageApdHandlerEv: Device or filesystem with identifier [####-####] has entered the All Paths Down Timeout state after 140 seconds. I/Os will now be fast failed.
• /var/log/hostd.log reports storage accessibility changes: UpdateStorageAccessibilityStatusInt: Vm's storage accessibility status changed to false VM config backing gone - try to mark VM invalid.

 

Environment

• VMware vSphere ESXi 8.0 U3k
• Nutanix AOS

Cause

Nutanix AOS is not at a version compatible with ESXi. 

Resolution

Contact Nutanix to determine a compatible Nutanix AOS version for the ESXi build you are running.