vCenter HA is new feature that provides a new availability option for vCenter Server. By deploying a passive node to vCenter Server with synchronous replication, vCenter HA provides a 5-minute RTO for the vCenter Server application. In addition to the passive node there is also a witness node that provides mediation and prevents split brain. vCenter HA can be deployed within a site or even across sites. The passive and witness nodes are created by cloning the active node which prevents long initial synchronization times and gets the HA cluster up and running quickly. vCenter HA is completely independent of another high availability feature that is being worked on called PSC HA (Platform Services Controller High Availability). However, when used together, vCenter HA and PSC HA provide resiliency to vSphere that will provide value to many VMware customers.

vCenter HA can be deployed using two methods:

• Automated Deployment - a more basic wizard that deploys all components within a single site
• Manual Deployment - relies on manual cloning of vCenter so as to allow placing the passive and witness nodes at remote sites and potentially across different networks

1. Click on Hosts and Clusters.
2. Click the Manage tab.
3. Click vCenter HA.

1. Click the New Configuration button.
2. Review the two deployment options. Note that we'll be following the Automated deployment. This is useful when deploying all vCenter HA nodes within the same datacenter. The Manual deployment can be chosen when deploying the Passive and Witness nodes in different physical locations or where there is network segmentation between the nodes.
3. Click Next to continue.

An additional ethernet adapter will be added to the Active vCenter Server node. We will need to provide some network settings for this adapter. This adapter will be used for communicating to the Passive and Witness nodes on a private network. The network provided in this interactive simulation is just used as an example. There are no specific requirements for this private cluster network other than all three nodes need to be able to communicate to each other and it is recommended that this network be a dedicated network just for VCHA.

1. Hit the Browse button to select the network that will be used as the private cluster network between the three vCenter HA nodes.
2. Click the VCHA Network and click OK.
3. Select the IPv4 address box and type in 192.168.0.1 for the private cluster IP address of the Active vCenter Server node.
4. Select the IPv4 subnet mask box and type 255.255.255.248 as the subnet mask.
5. Click Next to continue.

Here we'll need to provide two additional IP addresses on the private cluster network for the Passive and Witness nodes. These IP addresses need to be on the same network as the IP address from the prior step (192.168.0.1/29).

1. Select the vCenter HA IP address (nic1) box under the Passive Node (new) heading.
2. Enter the IP Address 192.168.0.2.
3. Select the vCenter HA IP address (nic1) box under the Witness Node (new) heading.
4. Enter the IP Address 192.168.0.3.
5. Click Next to continue.

In this particular lab environment DRS and Storage DRS are not enabled. If they were, we would likely just accept the default placement of the Passive and Witness VMs. DRS, by default, will automatically separate all three nodes and place them on different hosts and datastores if available. However, in this step, we will show you how to manually place the nodes if DRS is not available. The Active vCenter Server node is currently running on esx-01a.sddc.local and we will place the Passive node on esx-02a.sddc.local and the Witness onto esx-03a.sddc.local to ensure we are protected from a host failure.

1. Click the Change Location... button for the Passive Node.
2. Note that you may change the name of the Passive Node on this screen and also change the folder to where the passive is deployed. For this simulation we will leave the defaults. Please click Next.
3. Choose esx-02a.sddc.local as the target host for the Passive Node and then click Next.
4. This lab environment only has a single, local datastore per host so we'll accept the local datastore for this host. Click Next.
5. We've already chosen the appropriate networks so we can click Next.
6. Review the settings and click Finish.

Now we will repeat the previous process for the Witness Node.

1. Click the Change Location... button for the Witness Node.
2. Note that you may change the name of the Witness Node on this screen and also change the folder to where the passive is deployed. For this simulation we will leave the defaults. Please click Next.
3. Choose esx-03a.sddc.local as the target host for the Witness Node and then click Next.
4. This lab environment only has a single, local datastore per host so we'll accept the local datastore for this host. Click Next.
5. We've already chosen the appropriate networks so we can click Next.
6. Review the settings and click Finish.
7. Click Next to continue the wizard.

This final screen is to review all the settings we've previously set. Upon clicking Finish we'll start the process of cloning the vCenter Server to create the Passive and Witness nodes.

1. Click Finish.
2. Observe as the Passive (vcsa-01a-peer) and Witness (vcsa-01a-witness) nodes are created.
3. Once the operations are complete the browser refreshes and you can see information regarding the cluster. Also note that the status will briefly be Orange due to a final synchronization between the Active (vcsa-01a) and Passive (vcsa-01a-peer) nodes.
4. Click the vCenter HA Monitoring link to be taken to a page where we can view more detailed status of the cluster.

This screen shows us more detailed information about the vPostgres database and appliance configuration. Notice that now the health of the cluster is Green as it takes a very short amount of time, typically, for the delta sync to complete.

1. Click the vCenter HA Settings link to go back to the vCenter HA Configuration screen.

One of the operations we can perform against the cluster is a planned failover. We'll initiate a planned failover and see how the cluster behaves. You might do this during planned maintenance or period testing to ensure failover is working properly.

1. Click the Initiate Failover button.
2. Review the information on the screen. Note the Force option which can be used if the cluster is not in a healthy state.
3. Click Yes to initiate the failover.
4. Shortly after the failover is initiated we'll encounter an error stating that the vSphere Web Client is unreachable. This is expected as vCenter Server is down while the Passive Node is in the process of becoming the Active Node. Click OK to clear the error.
5. In order to make sure the browser is in a clean state, click your username (administrator@vsphere.local) and then click Logout.

Once the failover completes we'll be able to log back into the vSphere Web Client. At this point, vcsa-01a-peer is now the Active Node while vcsa-01a is now demoted to the Passive Node. At this point we can perform troubleshooting on the Passive Node or even replace it. Let's log back into vCenter Server and see what other operations we have available now that the failover operation has completed.

1. In the User name: field type administrator@vsphere.local
2. In the Password: field type VMware1!.
3. Click Login.

First, we'll log back in and look at the Monitoring are to make sure the failover was successful and we're showing all Green.

1. Click on Hosts and Clusters.
2. Click vCenter HA Monitoring.
3. Note all health indications are Green.
4. Click the vCenter HA Settings link.

There are a few different operating modes for vCenter HA. Let's review those now.

1. Click the Edit... button.

• We can see that we have a mode called Maintenance Mode. This mode can be used during patch cycles or planned maintenance. As its description states, all nodes remain online but automatic failover is disabled. This might be useful, for example, if making network changes and we want to prevent those changes from triggering a failover event.
• Next we can Turn OFF vCenter HA. This mode diables replication and automatic failover. This might be useful to change the location of the Passive and Witness nodes. When re-enabling vCenter HA the nodes will sync back up and the configuration is kept intact.
• The final option is to Remove vCenter HA cluster. This option destroys the Passive and Witness VMs and removes the all of the vCenter HA configuration. If you perform this action you will need to completely setup vCenter HA if you want to re-enable it.

Click the Cancel button.

This concludes the Hands-on Labs Interactive Simulation: vCenter HA

1. Click the Return to the lab link or close the windows to continue with this lab.