Some times, for some reason (like for rearrangement in my case) system administrator wants to change the host location and move one (or more) host from one cluster to the other, to do so, the host should be remove from first cluster and join to the other.
Well, it’s a very easy process but there is a little point I find out today, during doing this on my home lab, lets check these steps.
Condition:
Have three cluster and at least two hosts under each one.
Goal:
Moving one host from one cluster to the other cluster.
Process:
well, this is the screen of two hosts which I wanted to move (both of them are virtualized).
Both of these hosts are same in the spec and configuration except the dvSwitch they connected to.
In first step, I have put the first host (vesxi41-2) in the “Maintenance Mode” so if any VMs is assign to this host (powered on or off), would migrate to the other host on the same cluster.
After that, I have right clicked on it and select the “Remove” to remove it from current cluster.
Normal Warning shows up which indicated that all logs and performance data would be removed as well as VMs, Resource pools and vApps, so as long as I didn’t have anything on this host, I selected the “Yes” option.
A little latter, the unexpected warning shows up!
“The resource ‘vesxi41-2.kasraeian.com’ is in use.”
“See the error stack for details on the cause of this problem.”
As the picture below shows, the cause of this error was: “… it’s part of vDS dvSwitch”
So I went to the “Networking” section of vCenter to check the config of dvSwith for this host.
As the above picture shows, I selected the dvSwith for vDC1 cluster and there is 2 hosts in detail of this dvSwitch; after this, I have selected “Manage Hosts” from “Summary” tab, then I selected the proper host (vesxi41-2) in the new window.
After pressing “Next” in previous page, the list of pNICs on this host appeared with a check mark on the pNIC which connected to the dvSwitch.
I have removed this check mark so the connection of the pNIC and the dvSwitch would be broken.
Just after removing the check mark, following warning message shows up to warn about network problem and …
After pressing two “Next” button, preview of dvSwitch state after pressing “Finish” shows up.
One pNIC from one host connected to one dvUplink, thats what I was looked for.
I pressed the “Finish” button and I directed back to the main window of vCenter and this time in the “Networking” Section – “Configuration” tab.
I opened the “Hosts” tab to check the host state, yet the vesxi41-2 was there.
So I right clicked in this host and select the “Remove from vNetwork Distributed Switch…“.
It took a while (about 10~15 sec) for the system to reconfigure the dvSwitch.
After the task is finished successfully, the list of “Hosts” changed.
I browse back to the summary page for this host to check the available network and config, it seems everything would be OK this time.
I did the “Remove” action again, for removing the host from cluster and the summary page shows me something good this time and it was “State: Disconnected (maintenance…“.
After completion of this task, I joined the vesxi41-2 to the new cluster (vCluster2 under vDC2 data center) and did the same steps for second host (vesxi41-3).
Summary:
- If there is host which is connected to the dvSwitch(es) and wants to be moved, first try to disconnect this host from dvSwitch(es) and then try to remove it from host.
- During my test on the second host, I find out the steps of disconnecting the pNIC from dvSwitch can be removed which means, the host can be removed from dvSwitch directly.
