Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

I am building a new vSphere cluster from scratch. I have installed ESXi on the first host, and built a vCenter server on a VM residing on that host (storage is on the local hard drive, although we have iSCSI targets which I can reach from the host). The cluster is configured for HA. When I try and add the host to the cluster, I get an error at the point where HA is configured - Cannot complete the .

I have stripped the network configuration of the host down to the most basic - a single NIC attached to a single vSwitch - this is running the VMKernel Port on VLAN 8 - that is our Management VLAN. The vCenter server will have a network address on this VLAN, so I also set the initial Virtual Machine Port Group to this VLAN, and connected the vCenter server NIC to this port group. I understand I can't connect the vCenter server to the VMkernel port group, but shouldn't I be able to connect the vCenter server to a Port Group in the same VLAN? If not, do I need to create a VLAN specifically for VMKernel Port Group? I plan to set up another port group for vMotion with a dedicated and isolated VLAN (i.e. VLAN isn't routed) so this wouldn't allow vCenter to communicate.

Does anyone have any suggestions, or other ideas for what might be causing the problem. I've read through the documentation, but it isn't giving me any pointers, and the error message isn't helping me beyond telling me something is wrong with my network config.

share|improve this question

2 Answers 2

up vote 0 down vote accepted

Ok - tracked it down. It wasn't a misconfiguration of my network. It was a misconfiguration of the HA settings for the cluster.

The VMware hosts are connected to two Dell 6248 core switches with default routing to a VRRP address. This address isn't pingable, and VMware uses the gateway as the isolation address. I discovered this through the error messages for the host - the specific error (can't contact isolation address: ip address) isn't displayed at the cluster level.

I fixed this by adding the following values to the advanced options for HA:

  • das.isolationaddress1 = home ip address of core switch 1
  • das.isolationaddress2 = home ip address of core switch 2
  • das.usedefaultisolationaddress = false
  • das.failuredetectiontime = 20000

VMware recommend increasing the failuredetectiontime value when using more than one isolation address.

After adding these settings I disabled and renabled HA, and the error is no longer displayed for my host.

share|improve this answer

In a non-iSCSI setup there's nothing to stop every host having just a single vSwitch and single port group - it'd be carrying both non-iSCSI vmkernel traffic as well as guest traffic (including the VC of course) but that's just a design issue not an implementation one.

I actually suspect this is more likely to be a DNS issue, HA is very sensitive to having 100% reliable forward AND reverse DNS entries for just about everything, especially hosts and VC - I'd check those first as it's a quick check, other than that can you screen grab your setup for us please.

edit - damn, didn't see your own response - good job :)

share|improve this answer
    
+1 because that usually is the cause of this problem. –  dunxd Feb 11 '11 at 15:07
    
I have seen errors with HA directly related to DNS before as well. –  Chadddada Mar 11 '11 at 20:38

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.