ESXi host not responding in vCenter is one of those problems that looks scary and is usually boring: the host and its virtual machines keep running, but vCenter loses its management connection and marks the host as Not Responding. Here are seven proven steps, in the order that saves the most time, to find the cause and bring the host back without a reboot.
What an ESXi Host Not Responding Error Really Means
When vCenter can no longer talk to the management agents on a host, it flips the connection state to NO_RESPONSE. Broadcom’s knowledge base article on ESXi hosts disconnecting and reconnecting describes the same pattern: hosts show Not Responding, then reconnect on their own once the network recovers. Running VMs are not affected, because the VMkernel keeps them going. You just lose management.
The cause that Broadcom names is an interruption on the physical network carrying traffic between vCenter Server and the host. A second, less common family of causes lives on the host itself: a hung hostd or vpxa agent, or a failed reconnect after an upgrade. That is why the order of the steps below matters.

ESXi Host Not Responding: Symptoms and Log Messages
Before you change anything, confirm that you really have an ESXi host not responding case and not a powered-off server. Typical signs:
- The host shows Not Responding or Disconnected in the vSphere Client, sometimes flapping at random.
- In vpxd.log on vCenter you see
connection state changed to NO_RESPONSE,Resetting stub adapter: service state request timed outandException occurred during host sync; Host communication failed. - In vpxa.log on the host you see
HTTP Connection has timed out while waiting for further requests. - VMs on the host still answer pings and run normally.
On a vCenter Server Appliance, vpxd.log is at /var/log/vmware/vpxd/vpxd.log. The quickest way to find the event is a simple search from the appliance shell:
grep -i "NO_RESPONSE" /var/log/vmware/vpxd/vpxd.log | tail -n 20Code language: JavaScript (javascript)Step 1: Test the Network Path First
Start with the network, because that is where an ESXi host not responding problem usually begins. Broadcom lists UDP 902 and TCP 443 as the ports that must work in both directions between vCenter and the host management IP. From the ESXi shell, check that the management VMkernel interface can reach vCenter and that port 443 answers:
vmkping -I vmk0 <vcenter-ip>
nc -z <vcenter-ip> 443
esxcli network ip interface ipv4 getCode language: HTML, XML (xml)Then test from the other side with PowerShell on an admin workstation or the vCenter-side network, replacing the name with your host:
Test-NetConnection esx02.lab.local -Port 443Code language: CSS (css)If pings drop or time out intermittently, stop here and look at the switch ports, routers and firewalls for flapping links and packet loss. Restarting agents will not cure a bad cable.
Step 2: Check NICs, VLANs and Teaming for an ESXi Host Not Responding
If the path is flaky, work through the checklist in Broadcom’s article on intermittent network connectivity or host disconnect issues: confirm the physical link is up, the NIC driver and firmware are current, there are no duplicate IP addresses, the VLAN ID on the port group is correct, teaming and load balancing match the physical switch, and port security and STP delays (no portfast) are not blocking the port. These commands list the adapters and their link state:
esxcli network nic list
esxcli network vswitch standard listCode language: PHP (php)
Step 3: Read the Logs on Both Sides
If the network is clean, the logs tell you which side of the ESXi host not responding problem is stuck. On the host, check the agent logs. If vpxa.log shows the HTTP timeout message and vpxd.log shows the stub adapter reset, the management agents on the host are the next suspect.
tail -n 100 /var/log/vpxa.log
tail -n 100 /var/log/hostd.logCode language: JavaScript (javascript)Step 4: Restart hostd and vpxa
To restart hostd and vpxa, open an SSH session or the ESXi Shell as root. Broadcom’s article on restarting management agents in ESXi recommends restarting the two services individually:
/etc/init.d/hostd restart
/etc/init.d/vpxa restartThe same article warns against the restart-everything command, services.sh restart, in environments that use vSAN or NSX, so use the two commands above unless you have confirmed your setup is safe. You can also restart either service from the Host Client under Manage, Services. Wait a minute, then check whether the ESXi host not responding state clears and vCenter flips back to Connected.
Step 5: Use the DCUI When SSH Restarts Do Not Help
There is a documented ESXi host not responding case where the management network itself is unreachable and restarting hostd and vpxa over SSH does not help. In that case, Broadcom’s article on a server state changing to not responding points to the Restart Management Agents option in the DCUI on the host console. Notice that this conflicts with the cautions in the restart article above, so check your vSAN and NSX situation before using it.
Step 6: Reconnect an ESXi Host Not Responding in vCenter
If an ESXi host disconnected from vCenter stays that way, reconnect it manually. In the vSphere Client, right-click the host, choose Connection, then Connect. With PowerCLI the equivalent is:
Connect-VIServer vcenter.lab.local
Set-VMHost -VMHost esx02.lab.local -State ConnectedCode language: CSS (css)If the reconnect fails right after an upgrade with a message that begins Datastore conflicts with an existing datastore in the datacenter, you may be hitting a different problem. Broadcom’s article on reconnecting an ESXi host to vCenter traces it to a datastore that is force-mounted on some hosts and mounted normally on others. The fix there is to find the force-mounted datastores, unmount them, update the host, and remount them afterwards.
Step 7: Verify the ESXi Host Is Connected Again
Do not declare victory after the first green icon when an ESXi host not responding alert clears. Run a quick check across all hosts, then watch for an hour. A repeat of the ESXi host not responding state means the network problem is intermittent and you should capture packets with pktcap-uw on the VMkernel interface and the physical NIC, as Broadcom suggests.
Get-VMHost | Select-Object Name, ConnectionState, PowerStateCode language: JavaScript (javascript)
How to Stop an ESXi Host Not Responding Problem From Returning
- Monitor switch ports for flapping and enable portfast on host-facing ports.
- Keep NIC drivers and firmware aligned with the hardware compatibility list.
- Document and test the UDP 902 and TCP 443 rules whenever firewalls change.
- Alert on the NO_RESPONSE event in vCenter so you hear about it before your users do.
For other vSphere management problems, see our guide to the greyed out Migrate button in VMware vCenter and, if you are locked out of the host itself, how to reset a lost ESXi root password.
Read more: Greyed Out “Migrate” Button Issue in VMware vCenter and Its Solution
Read more: Reset a Lost ESXi Root Password
Read more: Exchange OWA Not Working After October 2026 Updates? Fix It Without Uninstalling
Why is my ESXi host not responding in vCenter?
Most often because the physical network between vCenter and the host is interrupted. Broadcom also lists a hung management agent or a failed reconnect after an upgrade as other possibilities.
Do running VMs stop when an ESXi host is not responding?
No. The VMs keep running because only the management connection to vCenter is lost, not the VMkernel. You lose management and monitoring until the host reconnects.
Which ports must be open between vCenter and an ESXi host?
Broadcom lists UDP 902 and TCP 443 for bidirectional communication between vCenter Server and the ESXi management IP.
How do I restart hostd and vpxa on ESXi?
Log in over SSH as root and run /etc/init.d/hostd restart followed by /etc/init.d/vpxa restart, or restart them from the Host Client under Manage, Services.
Is it safe to run services.sh restart?
Broadcom advises against it in environments that use vSAN or NSX. Restart hostd and vpxa individually instead.
References
Broadcom: ESXi host or hosts disconnecting and reconnecting
Broadcom: Restarting Management Agents in ESXi
Broadcom: The server’s state changed to not responding
Broadcom: Intermittent network connectivity on ESXi or host disconnect issues
Broadcom: Unable to reconnect ESXi host to vCenter






