Understanding the messages: Lost connection to server nfs-server mount point and Restored connection to server nfs-server mount point

Products

• VMware vCenter Server
• VMware vSphere ESXi

Issue/Introduction

ESXi hosts lose or intermittently cycle connectivity to an NFS datastore. This behavior presents through system log events indicating that the connection to the network file system mount point has been dropped or subsequently restored.

 

This article provides information about the messages:

• Connectivity to NFS Server Lost
• Connectivity to NFS Server Restored

Environment

VMware vSphere ESXi 7.x

VMware vSphere ESXi 8.x

Cause

If the VMkernel interface used for the NFS connection is configured with MTU 9000/1500 (jumbo frames), and there is a mismatch in MTU settings along the network path or on the NAS storage end, the connection between the NFS server and the ESXi host can be disrupted.

Resolution

When an ESXi host experiences network layer interruption with an NFS volume, the following symptoms are recorded in /var/log/vmkernel.log

1. Connectivity to NFS Server Lost
    
   An event indicates that the ESXi/ESX host has lost connectivity to the NFS server. For example, if an NFS volume nfs-share is mounted from host nfs-server:/share, the events in the ESXi/ESX logs (located at /var/run/log/vmkernel on ESX and /var/run/log/messages on ESXi) if connectivity to the NFS server is lost:
   
   Lost connection to server nfs-server mount point /share, mounted as #### ("nfs-share")
   
   As a result, the NFS volume (and virtual machines residing in the affected volume) are unavailable to the ESXi/ESX host.
   
   This event is normally received in an environment where NFS was already working and for some reason the NFS connectivity was lost.
   
   The first troubleshooting step is to verify that the NFS server is up and running without issues. Check if other hosts using the same NFS mount are experiencing the same problem and check NFS server status and share points.
   
   If the issue points to a loss of connectivity between the NFS server and the ESXi/ESX host, see Troubleshooting connectivity issues to an NFS datastore on ESX / ESXi hosts.
   
   
2. Connectivity to NFS Server Restored
    
   If an NFS volume nfs-share is mounted from the host nfs-server:/share, the events in the ESX logs (located in /var/log/vmkernel on ESX and /var/log/messages on ESXi) when connectivity to the NFS server is restored:
   
   Restored connection to server nfs-server mount point /share, mounted as #### ("nfs-share")
   
   ESXi/ESX host has recovered from the loss of connectivity to the reported NFS server. No actions are required.
   
   
   
3. To ensure optimal performance and reliable connectivity for NFS datastores, MTU settings must be consistently configured across the entire network path. This includes:             

   The ESXi VMkernel interface used for the NFS connection

   All intermediate physical network switches and routers

   The network interface of the NFS storage array
   
   

4. Any mismatch in MTU values along this path can lead to packet fragmentation, dropped connections, degraded performance, or failure to mount the NFS datastore.
   
   MTU Validation
   
   vmkping -I vmk# -d -s 8972 <Destination_IP> (to test jumbo frames)
   esxcfg-nics -l (to check physical NIC MTU)
   esxcfg-vmknic -l (to check VMkernel MTU)