Continuing to my last blog post series, this blog will cover exam objective 1.2 i.e. identify requirement for vCenter server high availability requirements. VCHA is not covered in any other objective so I will be covering brief overview, deployment scenario and how to configure it in this part of blog. So let’s get started.
Overview of vCenter HA Deployment
Before we can start with configuring vCenter server high availability we need to understand different requirements as well as what can be its deployment models. As vCenter server is available both as appliance and Windows base solution, we will look at how to achieve high availability in both scenarios. First we will talk about vCenter high availability in vCenter Server Appliance (VCSA) and later on we will look at how to achieve vCenter high availability using Microsoft failover clustering in Windows. I will cover vCHA in depth for appliance base solution as this is main focus of exam.
vCenter High Availability in VCSA
We can deploy vCenter server HA either using embedded platform services controller or with external platform services controller. If you are using external platform services controller, you can put it behind load balancer for added protection against PSC failure.
vCenter HA with Embedded PSC
When deploying vCenter HA with embedded PSC, deployment is similar to below:
Workflow of deployment:
- First we deploy a vCenter server with an embedded PSC. Perform initial configuration creating datastores, virtual network etc.
- Enable vCenter HA using Setup vCenter HA wizard. Here setup can be done in two ways either automatic or manual.
- If Automatic setup is selected, then cloning and configuration of network is taken care by vCenter server itself.
- If manual setup is selected, we need to create and configure clones.
- There will be total of three VMs which will be deployed –
- Active vCenter: It runs the active vCenter instance, provide connectivity to management interface and uses vCenter HA network to replicate data to passive node and communication with witness node.
- Passive vCenter: It is a clone of vCenter server, it receives updates from active vCenter and uses vCenter HA network to communicate with witness node.
- Witness Node: It is lightweight clone of vCenter and used as a quorum to protect against split brain scenarios.
- When cloning will start, it includes PSC services and data as well.
- vCenter server will perform replication over vCenter HA network after cloning is complete to make sure that passive node is in sync with active node. In embedded deployment, it will send PSC data as well to passive node.
- Once configuration is complete, vCenter appliance will be protected by vCenter HA and in case of host failure, passive node will take over and all services will be available from passive node.
vCenter HA with External PSC
Deployment of vCenter HA with External PSC will be similar to below diagram:
If you are using external PSC, then load balancer is required to protect against failure of Platform services controller. If one PSC goes down, then load balancer will redirect all request to second PSC.
Workflow of deployment:
- First we need to setup two external PSC and configure load balancer to redirect request coming to load balancer VIP to PSCs. These PSC will replicate SSO domain information and other services with each other.
- When deploying a vCenter server, we select external PSC and point it to load balancer VIP.
- Enable vCenter HA using Setup vCenter HA wizard. Here setup can be done in two ways either automatic or manual.
- If Automatic setup is selected, then cloning and configuration of network is taken care by vCenter server itself.
- If manual setup is selected, we need to create and configure clones.
- There will be total of three VMs deployed which will be –
- Active vCenter: It runs the active vCenter instance, provide connectivity to management interface and uses vCenter HA network to replicate data to passive node and communication with witness node.
- Passive vCenter: It is a clone of vCenter server, it receives updates from active vCenter and uses vCenter HA network to communicate with witness node.
- Witness Node: It is lightweight clone of vCenter and used as a quorum to protect against split brain scenarios.
- When cloning will start, it includes only information about load balancer VIP and external PSC and not PSC data.
- vCenter server will perform replication over vCenter HA network after cloning is complete to make sure that passive node is in sync with active node.
- Once configuration is complete, vCenter appliance will be protected by vCenter HA and in case of host failure, passive node will take over and all vCenter services will be available from passive node.
vCenter High Availability Requirements in VCSA:
- ESXi 6.0 or later
- Three hosts are recommended, if using VMware DRS to protect then minimum three hosts are required.
- vCenter server 6.5 or later
- Deployment size small or bigger, tiny is not recommended for production environment
- vCHA network latency should be less than 10ms
- vCHA network must be different from management network
- vCenter server Single standard licence
Configuring vCHA – Step by Step Guide
Step 1 – Login in vSphere client and navigate to vCenter server -> Config -> Settings -> vCenter HA
Step 2 – Click on Setup vCenter HA. A new window will open to start configuration of vCenter HA.
Step 3 – If you want to create and configure clones manually then un-select checkbox Automatically create and leave it checked.
Step 4 – Select the port group created for vCenter HA network by clicking on browse.
Step 5 – Click on Edit to setup Host, Storage and Network for Passive node as well as Witness node. Complete configuration should look similar to below:
Step 6 – Scroll down to click Next and configure IP settings.
Step 7 – Click on Finish to start deployment of vCHA. In Automated Deployment, vCenter will start configuring Active node and start cloning as well.
Step 8 – It will take some time and vCenter High Availability will configured.
vCenter server for Windows High Availability using Microsoft Clustering
If you are using vCenter server for windows and like to deploy high availability then Microsoft clustering can be used to deploy a highly available solution. We would explore this option under this configuration.
You can configure multiple instances in a cluster solution but only one can be active at a time. MSCS solution can be configured only for windows based vCenter server and doesn’t support appliance base vCenter server. Also, it is supported only for VMs and doesn’t support physical server based vCenter server.
Workflow to deploy vCenter server high availability using Microsoft Clustering
- Create a VM with OS 2008 R2 or 2012 R2.
- Attach RDM disk with separate disk controller in physical bus sharing option
- Install PSC before vCenter installation can be initiated.
- Install vCenter server on one of RDM disk and set all services startup type to manual.
- Power off VM and detach RDM disks.
- Clone the VM and do not select any customization specifications and do not use sysprep.
- Attach RDM disk to both VMs and power them on.
- Change the hostname and IP address on VM1 and install failover clustering feature on both nodes.
- Create a failover cluster and include both nodes.
- Create a new role as generic service, assign hostname and IP address which was used to install vCenter server, add RDM disk on which vCenter installation was done.
- For services, add VMware Service Lifecycle Manager and registry value as
HKLM\system\CurrentControlSet\Services\VMwareDirectoryService
- Add VMware AFD and VMware vCenter Configuration services to the role.
- Restart role.
High Availability for vCenter server for windows should be done now.
Further Reading and Reference Material:
https://docs.vmware.com/en/VMware-vSphere/6.7/vsphere-esxi-vcenter-server-671-availability-guide.pdf
I hope you enjoy the post. If you have any questions or feedback, please let me know know in comments.
VMvTips 