Skip to main content

Process to Take Out Hosts from NSX Cluster


We recently implemented NSX in one of the VDI environment.  The environment had vShield and trend Micro Deep Security to protect the desktops. Once we migrated from vShield to NSX, we observed that taking out the host from the NSX cluster needs to follow specific steps. If these steps are not followed in the specific order there may be possibility that the host may not be cleaned up thoroughly with NSX vibs. Here is the process that will make sure the hosts are cleaned up properly once removed from NSX cluster.


  • Put the ESXi host in maintenance mode. The NSX agent VMs that are specific and residing on that host for NSX management, will be powered down once host is in maintenance mode.
  • Take out the host from the cluster by dragging it out of the cluster (in the data center or in some other cluster)
  • Wait till the NSX agent/VIB gets uninstalled from the host (you will see that in the task manager of vCenter)

  • Once VIBs are removed by vCenter – Reboot the host


  • Once reboot is complete, remove the host out of maintenance mode – it will delete the 2 VMs that were residing on it (Guest Introspection and trend Micro Deep Security)


  • Once this is complete the host will be all cleared from the NSX vibs. 

Comments

Popular posts from this blog

Guest Introspection(GI) NSX Appliance gets Deleted Automatically after vCenter Disconnect

I was working in an environment where NSX was configured with Trend micro deep security in VDI environment to protect VDI desktops. We had 2 vCenters setup in that way and one of the vCenter is showing some odd symptoms. We are seeing some weird actions with NSX manager in one vCenter. So, I lose the connection to the vCenter from fat client, web client is still fine and doesn’t kick me off. In the process, the Guest Introspection appliance or GI VMs (appliance managed by NSX) from some  random hosts were getting deleted automatically -  from the cluster that was protected with NSX. When I checked the windows logs (as vCenter was on windows server) this is what I saw -  I opened a case with VMware support - They reviewed the NSX logs and came to conclusion that this is not the NSX issue but it's vCenter that is deleting those VMs. vCenter gets disconnected for few seconds, then connects back. It doesn't recognizes those GI VMs and starts de...

How to Check Up Time of ESXi Host or Virtual Machine from vCenter Server

To check up time of ESXi host from vCenter server - Logon to vCenter Server -  If using C# (fat) client - click on cluster - on right side pane click on Hosts - select the up time option by right clicking on the header. The column will be added at the end. You can drag and drop column to place it in your view. If using HTML client - click on cluster - on right side pane click on Hosts - select the up time attribute from show/hide column by clicking on the header  If using web (Flash)  client - click on cluster - on right side pane click on related objects - hosts - right click on the header to show/hide column - click on up time attribute Similarly,  you can check the up time for a virtual machine by clicking on the virtual machine tab instead of host tab, but only caveat  is vCenter gives you up time of a virtual machine according to reboots performed from the vCenter. So, rather than rebooting server fr...

VMWare ESXi Host showing Up Time as 0 second

I have seen this more on 5.5 ESXi hosts but the host shows the up time as 0 seconds and CPU and Memory percentage as 0% Easiest way to resolve this is to put the host in maintenance mode and then restart the management agents on the ESXi hosts.  #esxcli services.sh restart But to be more cleaner I prefer to put host in maintenance mode and reboot the host to get the clean state.