ESXi host or hosts disconnecting and reconnecting to vCenter randomly and intermittently.

Products

• VMware vSphere ESXi
• VMware vCenter Server

Issue/Introduction

• One or more ESXi hosts show in 'Not Responding' state in the vCenter Server UI. 
• In the vCenter Server /var/log/vmware/vpxd/vpxd.log the same host(s) state has changed to NO_RESPONSE.
  
  2025-12-22T20:57:47.558Z warning vpxd[386913] [Originator@6876 sub=MoHost opID=HB-host-####@162289-a5e439e] host [vim.HostSystem:host-####,<host_fqdn>] connection state changed to NO_RESPONSE
  
  
• Some, or all, of the following log prints may also be seen in the vCenter Server vpxd logs. All reflect network connectivity problems between the vCenter Server and the ESXi host. 
  

  2025-12-22T20:57:12.589Z info vpxd[06905] [Originator@6876 sub=DAS opID=FdmMonitor-domain-c####-27b36072] Lost connection to master [vim.HostSystem:host-####,<host_fqdn>] - updating host fdm states to unknown
2025-12-22T20:57:12.594Z info vpxd[06905] [Originator@6876 sub=MoHost opID=FdmMonitor-domain-c####-27b36072] VC state for host host-#### (FDM unreachable (from VC) -> FDM unreachable (from VC)), FDM state (Master -> UNKNOWN_FDM_HSTATE), src of state (host-#### -> null)

  2025-12-22T20:57:47.532Z error vpxd[06767] [Originator@6876 sub=vmomi.soapStub[16209] opID=2a####65] Resetting stub adapter: service state request timed out; a: <</hgw/host-####>, /vpxa>, pa: <last binding: <<UNIX ''>, <UNIX '/var/run/envoy-hgw/hgw-pipe'>> >, timeout: 120sec

  2025-12-22T20:57:47.549Z warning vpxd[386913] [Originator@6876 sub=InvtHostCnx opID=HB-host-####@162289-a5e439e] Exception occurred during host sync; Host communication failed; [vim.HostSystem:host-####,<host_fqdn>], e: N5Vmomi5Fault17HostCommunication9ExceptionE(Fault cause: vmodl.fault.HostCommunication

• May see indications of the same from the ESXi host logs, like this example from the /var/run/log/vpxa.log. 
  
  2025-12-22T20:57:38.323Z In(166) Vpxa[2099484]: [Originator@6876 sub=HTTP.HTTPService.HttpConnection opID=TaskLoop-host-####-5e] HTTP Connection has timed out while waiting for further requests; <io_obj p:0x000000abf3e1cac0, h:-1, <TCP '127.0.0.1 : 8089'>, <TCP '127.0.0.1 : 58364'>>, N7Vmacore16TimeoutExceptionE(Operation timed out: Stream: <io_obj p:0x000000abf3e1cac0, h:-1, <TCP '127.0.0.1 : 8089'>, <TCP '127.0.0.1 : 58364'>>, duration: 00:20:04.830186 (hh:mm:ss.us))
  
  
• Once the physical network interruption is fixed, or ends, this log print from vCenter vpxd.log indicates the host has once again connected to the vCenter Server properly. 
  
  2025-12-22T20:59:34.764Z info vpxd[06315] [Originator@6876 sub=MoHost opID=HostSync-host-####-7d6de18c] host [vim.HostSystem:host-####,<host_fqdn>] connection state changed to CONNECTED

Environment

• VMware vSphere ESXi
• VMware vCenter Server

Cause

This issue will occur when there is an interruption in the physical network responsible for carrying and delivering the network communication between the vCenter Server and the ESXi host that is in "Not Responding" state. 

Resolution

1. This issue should be investigated and remediated on the physical network between the vCenter Server and ESXi hosts.
2. Investigate the physical network infrastructure (switches, routers, and firewalls) for port flapping or packet drops.
3. Verify bi-directional communication for UDP port 902 and TCP port 443 between vCenter and ESXi management IPs.
4. Consult the Broadcom TechDocs "vCenter Deployment and Setup" and "ESX Installation and Setup".

Additional Information

Additional troubleshooting steps for this issue can be found in KB Intermittent network connectivity on ESXi or host disconnect issues.

Contact Broadcom support