Moving ESX Server from Site to Another - Still within the same vCenter

[MarcusOsb]

Hi,

Just a quick question, I have a ESX 5.0 host set up at one site that I need to physically move to another site, it will still be within the same vCenter environment. All that's changing is its physical location, and that the IP addresses between the sites will be changed accordingly.

I set the server in to maintenance mode and shut it down, and before I plugged it in to the new location I changed its IP Addresses, eg vmKernel and vmManagement addresses to be within the sites address range.

But when I plugged it in to the physical switch at this site all existing ESX hosts became unavailable and it still listed ESX host i had just moved at the previous site.

Have I overlooked something? I do see an option (when right clicking on the host) to remove it but advises that it will remove all VM's on it.

If i did this will all hosted VM's be removed? or should I simply drag and drop the host to the new site?

Thanks,

Mark

[suj27]

This should be also a move from one cluster to another, unless it's a stretched cluster. Hope the VMs are on shared storage, so when setting to maintenance mode the VMs would be moved/re-registered to other hosts. You should remove the host from one cluster and add to another, the message is about VMs getting removed from vCenter inventory, if they are on shared storage they are already moved to another host and you can remove the host safely.

[Kedar_D]

I assume you have required network connectivity from esxi host to vcenter. Is your host (recently moved) connected to vcenter or it is in disconnected state?

Are you using fqdn to register hosts to vcenter?

If you remove host from vcenter it will remove all related information from vcenter.. is your host part of cluster? If not., You can simply drag and drop the host to new site folder or datacenter. It will retain the host and vms data..

[SureshKumarMuth]

First thing, you cannot remove a running ESXi host from vcenter directly, you have to disconnect it first then you have to remove it else it will ask to put the host in maintenance mode which will in turn required to move or shutdown all the VMs on the host. Hence, try disconnecting the host from VC then remove it.

I don't understand this

But when I plugged it in to the physical switch at this site all existing ESX hosts became unavailable and it still listed ESX host i had just moved at the previous site.

How plugging a new machine in a switch affects other ESXi hosts?? could you please explain this part?

[MarcusOsb]

Thanks for your reply.

Yes, we have network connectivity from the existing esx hosts to the vCenter.

The host was connected to vCenter, as it is currently grayed out in a "not responding" status. And still listed at the old site.

Yes, we are using fqdn.

The host is not part of a cluster, it is a stand alone host.

I assume that I can simply drag and drop it and keep all hosted VM's intact? They are stored locally on the host and not on shared storage.

[Kedar_D]

First of all update the DNS records of the host to reflect the new IP address for name resolution. Ensure that the vcenter is able to resolve the host names to new IP address and required ports are allowed between vcenter and esxi host. Then connect the host to vcenter. Since VMs are stored locally, drag and drop should work fine.

[MarcusOsb]

Thanks for your reply.

What I meant was when I connected the host up to the physical switches, the other ESX hosts where connected to them all had a red warning ! on them when viewed in vCenter.

I don’t fully understand it myself why this occurred, I thought that there may have been a clash in IP addresses and/or datastores. Only when I disconnected it from the switch did this clear the warning.

Thanks,

Mark

[Kedar_D]

What was that alert related to? Maybe a alert description or error message can help.

[MarcusOsb]

Having a look at the messages indicate that the vSphere server was disconnected from the master HA agent and that the vSphere HA availability was changed to unreachable.

The site that I connected the ESX host to does fave a failover cluster set up, could this have been the cause? As i'm only adding the ESX server to the environment as a standalone server eg testing purposes that will not affect the current environment.