Hosts not responding [vCenter 4 U1]

[piercelynch]

I am currently having an issue with a completely clean install of vCenter 4.0 U1 with a number of vSphere 4 U1 hosts, whereby after 2 minutes of being added to the vCenter inventory the hosts are then marked as 'not responding' and disconnected.

Having noticed that there are a number of threads with regards to similar issues, none of the advice provided has resolved the issue. I have of course tried using both the IP address and the FQDN of each host but to no avail.

The vSphere hosts and the vCenter server reside on the same subnet. I have also configured the managed virtual center IP to be the same as the interface IP address that is in use.

Looking through the vpxd logs - I can see that the hosts are regularly marked as dirty, before being marked as not responding.

Any advice as to cause and resolution of this issue would be greatly appreciated.

[a_p_]

Some thoughts. You may already have tried some of these.

The vSphere hosts and the vCenter server reside on the same subnet. I have also configured the managed virtual center IP to be the same as the interface IP address that is in use.

After setting the managed IP address, did you remove and re-add the hosts to vCenter?

see Verifying the VirtualCenter Managed IP Address

Please double check:

- DNS resolution for all systems (hosts and vCenter) Ping all systems using the FQDN with a dot at the end, like "esx01.mydomain.local."

- Network settings (IP, subnet and gateway addresses) as well as the DNS domain names on the hosts

- Do you use VLANs? Make sure the VLAN-ID for the VMkernel does not match the Native VLAN configured on the physical switch port

André

[piercelynch]

Hi Andre,

I do indeed use VLANs, with no native VLAN configured on the switch or vmnics.

To elaborate a little more, I am using HP BladeSystem kit with Flex10 virtual connect interconnect modules. I have installed the customer NIC driver supplied by VMware.

I am about to upgrade the hosts and vCenter to 4.1 to see if the issue is resolved, as everything that I have tried has not resolved the issue. So I would expect this could potentially be a bug within 4.0 with Flex10s....

P.