eBook

The Complete Guide to

VMware Clustering
 eBook

Contents
Overview ............................................................................................................................................... 3

Cluster: Definition and Types .......................................................................................................... 3

High Availability (HA) Cluster ........................................................................................................ 4

Distributed Resource Scheduler (DRS) Cluster .......................................................................... 4

How to Сreate a VMware Сluster .................................................................................................. 5

Installing VMware ESXi Server ...................................................................................................... 6

Installing Active Directory Domain Controller ........................................................................... 18

Installing and Setting Up vCenter Server ................................................................................... 27

Setting Up a Shared Datastore ..................................................................................................... 35

Connecting Hosts in Clusters ........................................................................................................ 37

Configuring Network for the Cluster ........................................................................................... 45

How to Create a DRS Cluster ........................................................................................................ 50

How to Create a HA Cluster ........................................................................................................... 53

Fault Tolerance: Purpose and Setup ............................................................................................. 58

Requirements for Virtual Machines with FT (v.6.0) ................................................................... 58

How to Enable Fault Tolerance ..................................................................................................... 58

Removing Hosts from the Cluster ................................................................................................. 62

More protection of your cluster with NAKIVO Backup & Replication ................................ 65

Conclusion ............................................................................................................................................ 67

2
 eBook

Overview
Modern information technology evolves rapidly, and the resulting innovations are used in
the growing number of industries. While these IT solutions provide automation, as well as
ensuring rational usage of natural and human resources, their hardware requirements are
gradually increasing. Even a powerful server can be overloaded with multiple computing
tasks. For better performance and reliability, servers can now be connected with each other
over networks. For this purpose, clustering technologies are widely used. This eBook explains
what clusters are, what issues you can resolve with clustering, and how to deploy clusters in
your VMware environment.

Cluster: Definition and Types

A cluster is a group of independent servers that communicate with each other over the
network and can act as a single system. The servers forming a cluster are called “nodes”
or “members” and are fine-tuned to perform the same tasks under the control of special
software. Any cluster consists of at least two nodes.

There are three commonly known types of clusters:

 High-Performance Computing clusters
 High Availability clusters
 Load Balancing clusters

High-Performance Computing (HPC) clusters are also called “parallel clusters”. They provide
a single system image. This means that an application can be executed on any of the servers
within the cluster. HPC clusters are used to execute computation-intensive and data-intensive
tasks by running a job on multiple nodes simultaneously, thus enhancing application
performance.

High Availability (HA) clusters are also referred to as “failover clusters” and deliver robust
operation with the minimal amount of downtime. Redundant storage, software instances,
and networking provide continued service when system components fail. HA clusters usually
use a heartbeat private network connection to monitor the health and status of each node in
the cluster.

Load Balancing (LB) clusters ensure better performance. In LB clusters, tasks are distributed
between nodes to load hardware more rationally and avoid overloading each server if there
are enough computing resources available.

In VMware vSphere, you can deploy two of the above types of clusters working on the virtual
machine layer: HA clusters and LB clusters (the latter of which are called Distributed Resource
Scheduler (DRS) clusters in the context of VMware vSphere).

3
 eBook

High Availability (HA) Cluster

A High Availability (HA) cluster supports the migration of virtual machines from one ESXi host
to another in case of failure. Two or more ESXi servers on the same network with shared
storage unite into a logical group called the “pool”. When one ESXi servers fails, the virtual
machines that were running on this host are started on another ESXi server within the
cluster. Powering on and loading these virtual machines may take some time (resulting in
short periods of idle time). After an ESXi server is added to a cluster, a special agent called the
Fault Domain Manager (FDM) is automatically installed. This utility monitors signals called
heartbeats from other ESXi hosts in the cluster and communicates with the vCenter Server
(once per second by default). If only one virtual machine fails, this VM is restarted on the
same ESXi server. The type of action taken depends on the type of failure detected, and you
can set rules in Preferences. The first five hosts added to a cluster are considered primary,
and all subsequent hosts are secondary. If one of the primary hosts is removed from the
cluster, a secondary takes up the primary role.

The main features of HA clusters are:

 Host monitoring – a feature that helps monitor each ESXi host in the cluster and make
sure that this server is running. ESXi hosts exchange network heartbeats in the cluster. If
an ESXi host fails, that host’s virtual machines can be restarted on another host.
 Admission control – a feature that controls the policy used by an HA cluster for reserving
resources to ensure failover capacity within the cluster.
 VM monitoring – a feature that helps monitor each virtual machine in the cluster with
VMware Tools heartbeats to ensure this VM is running. Any virtual machines that fail are
restarted.
 Datastore heartbeating – a feature that uses datastores to monitor hosts and virtual
machines when the management network fails. Moreover, this feature reduces the
probability of false restarts and false migration.
 Fault tolerance – a feature that allows you to avoid downtime of virtual machines by
running a VM replica on another ESXi host within the cluster.

Distributed Resource Scheduler (DRS) Cluster

A Distributed Resource Scheduler (DRS) cluster supports distributing computing resources
between hosts based on the performance required by virtual machines and the availability
of free ESXi host resources. The DRS checks the performance of your virtual machines
and makes placement decisions – that is, the scheduler determines to which host within
the cluster a particular VM should be migrated automatically or manually after receiving a
notification. Some virtual machines can be idle and “wake up” when they are required to
execute some important tasks, using CPU, memory, along with network bandwidth.

4
 eBook

This might prompt the DRS to move these VMs to another host with more free resources
available. The feature helps save management time that would otherwise be spent on
monitoring and maintaining the infrastructure.

The main features of DRS clusters are:

 Load balancing – a feature that allows performing or recommending migrations of virtual
machines between ESXi hosts to balance the load according to the settings you select.
 Power management – a feature that supports migration of virtual machines from one
ESXi host to another, if there are enough free resources, and sets the standby power
mode for the source ESXi server.
 Affinity rules – a feature that allows you to control the placement of virtual machines
among hosts by establishing rules.

How to Сreate a VMware Сluster

The hardware and software requirements for a VMware cluster are as follows:
 Availability of at least two ESXi servers with unique host names and static IP addresses.
For full compatibility, they should have processors of the same family, produced by the
same vendor.
 All hosts within the cluster should be attached to some form of shared storage, such as
network-attached storage (NAS) device or storage area network (SAN), via Fibre Channel,
Internet Small Computer System Interface (iSCSI), or Network File System (NFS) protocols.
Virtual Machine File System (VMFS) volumes must be used. There must be sufficient free
space on the storage medium.

NOTE: When choosing NAS or SAN solutions, use an authorized vendor that meet your
requirements for your production environment. Set up the storage according to the
manufacturer documentation. Multiple NAS or SAN devices can be used to create a VMware
cluster.
 All volumes on the ESXi hosts must use the same volume names.
 At least one VM must have the Active Directory Domain Controller has to be installed.
 At least one VM must have vCenter installed.

To create a VMware cluster, the following steps must be performed (each is explored
with an in-depth walkthrough in this section):

5
 eBook

1. Install a VMware ESXi Server.

2. Install Active Directory Domain Controller.
3. Install and set up vCenter Server.
4. Set up a shared datastore.
5. Connect hosts in clusters.
6. Configure the network for the cluster.
7. Configure HA and/or DRS clusters.

Installing VMware ESXi Server

VMware ESXi Server is an enterprise-class type-1 hypervisor with a proprietary kernel
(VMkernel) that runs directly on server hardware and does not require the installation of
an additional underlying operating system (OS). ESXi is highly reliable and includes an ultra-
thin architecture that is not dependent on a general-purpose OS. The smaller code-base
contributes to a low-vulnerability environment and makes for quick installation as well as
booting. As a virtualization server, ESXi is the most important component of the vSphere
environment. While the VMs themselves run on ESXi Server, the virtual disks of the VMs can
be stored either on internal data storage located directly on ESXi Server (such as hardware
SAS RAID) or a on a shared external datastore. You can access ESXi Server remotely via the
Secure Shell (SSH) client, the VMware Command Line Interface (CLI), or vSphere Client. SSH
access is disabled by default.

The minimum hardware requirements for ESXi are as follows:

 64-bit processor with at least two cores.
 At least 8 GB of RAM to take full advantage of ESXi 5.1+ features and run virtual machines
in typical production environments.
 Support for hardware virtualization (Intel VT-x or AMD RVI).
 Two or more Ethernet controllers (1Gb or 10Gb) to provide network redundancy for the
cluster.
 Small Computer System Interface (SCSI) disk, Serial Attached SCSI (SAS) disk, or a local,
non-network, Redundant Array of Independent Disks (RAID) Logical Unit Number (LUN)
with unpartitioned space.
 Serial ATA (SATA) disks connected through supported SAS controllers, or supported on-
board SATA controllers that are considered remote, not local; these disks are not used as a
scratch partition by default, because they are seen as remote.

6
 eBook

To install ESXi Server, do the following:

1. Insert the installation disc into the optical drive, select this disc as the first boot device in
BIOS, and boot from this disc. Alternatively, you can write the installation image to a USB
flash drive.

Select the Installer from the boot menu (see Figure 1.1).

Figure 1.1

NOTE: If your system hangs at the “user loaded successfully” stage (see Figure 1.2), this may
be due to insufficient RAM. Check the amount of memory available. Press Alt+F12 to view
details.

7
 eBook

Figure 1.2

If everything is okay, the “welcome” installation screen appears (see Figure 1.3).

Figure 1.3

8
 eBook

2. Select a disk on which to install ESXi (see Figure 1.4).

Figure 1.4

3. Select your keyboard layout.

4. Enter a root password (see Figure 1.5).

Figure 1.5

Then wait for system scanning to complete (see Figure 1.6).

9
 eBook

Figure 1.6

NOTE: If there are less than 2 CPU cores, an error message appears (see Figure 1.7).

Figure 1.7

If everything is okay, the confirmation message is displayed (see Figure 1.8).

10
 eBook

Figure 1.8

8. Press Install and wait for the installation progress to complete (see Figure 1.9).

Figure 1.9

7. You can see the “Installation Complete” message. Press Enter (see Figure 1.10).

11
 eBook

Figure 1.10

Wait for the server to reboot (see Figure 1.11).

Figure 1.11

8. After rebooting the server, log in ESXi (see Figure 1.12).

9. Press F2.

12
 eBook

Figure 1.12

10. Select Configure Management Network from the menu, and set up the host name along
with the IP address manually – for example, [Link] (see Figure 1.13).

Figure 1.13

11. Press Yes to confirm the network changes (see Figure 1.14).

13
 eBook

Figure 1.14

12. Now, you can download VMware vSphere Client. Open your web browser and enter the
URL of your ESXi server (see Figure 1.15).

14
 eBook

Figure 1.15

Follow the Download vSphere Client for Windows link and download the installer from the
official VMware website.

13. Install VMware vSphere Client by following the steps the installation wizard guides you
through (see Figure 1.16).

15
 eBook

Figure 1.16

14. Log in to ESXi Server via vSphere Client by entering the user name and password you
specified during the ESXi installation (see Figure 1.17).

NOTE: Ignore the certificate warning.

16
 eBook

Figure 1.17

You now have one ESXi server installed. The second one can be installed in the same way.
This is how the vSphere Client interface looks (see Figure 1.18):

Figure 1.18

17
 eBook

Installing Active Directory Domain Controller

vCenter needs a domain controller with a Domain Name System (DNS) server installed for
centralized management and resolving DNS names to IP addresses. The Domain Controller is
the main component of the Active Directory Domain Services server role, which needs to be
installed in this case.

NOTE: Consult the table on the VMware website displaying the compatibility of the various
versions of Microsoft Windows Server with the version of vCenter you are using to make sure
that the installation can work properly. In the walkthrough that follows, Windows Server 2008
R2 x64 is used, as this version of Windows Server is compatible with most versions of vCenter.

NOTE: The instance of Windows Server you use as the domain controller can be installed on a
physical server or a virtual machine.

To install Active Directory Domain Services, take the following steps:

1. Press Win+R (or go to Start -> Run) and type “dcpromo” in the Run window (see Figure
2.1).

Figure 2.1

Wait for the installation wizard to launch (see Figure 2.2).

18
 eBook

Figure 2.2

2. Tick the checkbox near Use advanced mode installation and click Next (see Figure 2.3).

Figure 2.3

3. Read the information on OS compatibility and click Next (see Figure 2.4).

19
 eBook

Figure 2.4

4. For the first Domain Controller setup in your infrastructure, select Create a new domain
in a new forest (see Figure 2.5).

Figure 2.5

20
 eBook

5. Enter your domain name (see Figure 2.6).

Figure 2.6

6. Enter Accept the Domain NetBIOS name, or alter the one provided, if necessary.
(see Figure 2.7).

Figure 2.7

21
 eBook

7. Set the forest functional level (see Figure 2.8).

Figure 2.8

NOTE: If there are no Windows Server operating systems older than Windows Server 2008
that you want added to the domain as domain controllers, set the forest functional level to
Windows Server 2008.

8. Similarly, set the domain functional level (see Figure 2.9).

22
 eBook

Figure 2.9

9. Under Additional Domain Controller Options, tick the checkbox near DNS server
(see Figure 2.10).

23
 eBook

Figure 2.10

10. As this is an internal domain, click Yes to continue (see Figure 2.11).

Figure 2.11

11. Set the location for the Domain Controller files. You can leave the default location
(see Figure 2.12).

Figure 2.12

24
 eBook

12. Set the password for the Domain Controller Administrator account. The password must
meet the complexity requirements set in the default Domain Controller security policy
(see Figure 2.13).

Figure 2.13

13. View the summary and click Next to start the installation (see Figure 2.14).

25
 eBook

Figure 2.14

Wait for the installation to complete (see Figure 2.15).

Figure 2.15

14. Click Finish and reboot the server (see Figure 2.16).

Figure 2.16

26
 eBook

Installing and Setting Up vCenter Server

vCenter is a VMware centralized management server application that controls all virtual
infrastructure, providing resource provisioning and performance evaluation of virtual
machines in the vSphere virtual environment. vCenter Server can be installed manually on
Windows or deployed as a virtual appliance. The vCenter Virtual Appliance is a preconfigured
Linux-based virtual machine image with all the necessary software already installed.

NOTE: Do not install vCenter Server on the machine with Active Directory Domain Controller
installed.

vCenter can be installed either on a physical machine or a virtual machine. Installing vCenter
on a virtual machine offers a number of advantages, including the following:

 no need to dedicate a separate server.

 snapshots usage and ease of backup.
 easy migration of VMs from one host to another.
 provision of high availability for the vCenter Server system by using vSphere HA.

The following are the minimum requirements for vCenter Server installation:
 64-bit operating system, such as Windows Server 2008R2 x64 (this is required for some
versions of vSphere, but not all)
 CPU: 2 GHz or faster Dual Core 64-bit processor
 RAM: 8 GB minimum (requirements can increase with higher numbers of virtual machines
in vSphere)
 Disk storage: 40 GB minimum (this varies depending on the database type and the
number of VMs)

To install and set up vCenter Server, take the following steps:

1. Insert the installation media for the machine on which you want to install vCenter Server
and run [Link] (see Figure 3.1).

27
 eBook

Figure 3.1

2. In the VMware vCenter Installer window, select vCenter Server for Windows and click
Install (see Figure 3.2).

Figure 3.2

3. Read and accept the terms of the End User License Agreement displayed. Click Next (see
Figure 3.3).

28
 eBook

Figure 3.3

4. Select the deployment type. In this installation walkthrough, the Embedded Deployment
was used. Click Next (see Figure 3.4).

Figure 3.4

5. Enter the System Network Name (see Figure 3.5).

29
 eBook

Figure 3.5

5. Set vCenter Single Sign-On domain and click Next (see Figure 3.6).

Figure 3.6

7. Specify a vCenter Server Service Account for logging into vCenter with the vSphere Client.
Click Next (see Figure 3.7).

30
 eBook

Figure 3.7

8. Select a database for deployment (see Figure 3.8).

Figure 3.8

31
 eBook

NOTE: The embedded PostgreSQL database can be used in environments of up to 20

ESXi hosts and 200 virtual machines. Larger infrastructures require a supported external
database, such as Oracle or MS SQL.

9. Configure ports. You may leave the default settings (see Figure 3.9).

Figure 3.9

10. Select the destination directory (see Figure 3.10).

32
 eBook

Figure 3.10

11. Tick the checkbox near Join the VMware Customer Experience Improvement Program,
if you want to participate (see Figure 3.11).

Figure 3.11

33
 eBook

12. View the installation summary and click Install (see Figure 3.12).

Figure 3.12

Wait for the installation process to complete (see Figure 3.13).

Figure 3.13

34
 eBook

You can now launch vSphere Web Client (see Figure 3.14).

Figure 3.14

Setting up a Shared Datastore

For high availability, ESXi hosts need access to shared data storage resources.

To set up a shared datastore, take the following steps:

1. Browse hosts and clusters in the vSphere Web Client, select a host, then click the Manage
tab. Select the Storage subtab.
2. Click Storage Adapters to view the list of available adapters (see Figure 4.1).

35
 eBook

Figure 4.1

3. Click Storage Devices to view the list of available devices (see Figure 4.2).

Figure 4.2

4. Go to Home -> vCenter -> Datastores. Click the Create a new datastore icon to add the
shared datastore you want (see Figure 4.3).

Figure 4.3

5. Select the IP address of the datastore you want to connect via network using the NFS, iSCSI,
or Fibre Channel protocols (see Figure 4.4).

36
 eBook

Figure 4.4

NOTE: The recommended redundant storage network scheme is as follows: two ESXi hosts
connected via redundant network to a SAN with two storage processors. However, you can
use a NAS server for this purpose (see Figure 4.5).

Figure 4.5

Connecting Hosts in Clusters

Now that the ESXi server, Domain Controller, and vCenter Server are installed, with a Shared
Datastore (NAS or SAN) set up, you can create a cluster.

There are several ways to connect hosts in clusters:

NOTE: The icons in the diagrams below signify the following:

37
 eBook

1. In this scenario, the Domain Controller and vCenter Server are installed on physical
machines (see Figure 5.1). Bear in mind that if you use this setup initially, you can always use
VMware vCenter Converter to convert a physical machine into a virtual machine at any time.
You could thus migrate vCenter Server or Domain Controller to the vSphere environment
later.

Figure 5.1

2. In this scenario, vCenter Server is a virtual machine that is installed on an ESXi server, using
the CPU, RAM, and storage of the ESXi server (see Figure 5.2).

38
 eBook

Figure 5.2

3. vCenter Server is a virtual machine running on an ESXi Server that uses CPU and RAM
of the ESXi server, but the virtual disk is stored on a shared datastore (see Figure 5.3). This
method of connecting hosts in a cluster allows you to use clustering features, such as High
Availability, the Distributed Resource Scheduler, and Fault Tolerance.

Figure 5.3

4. Domain Controller as well as vCenter Server are both installed and running on an ESXi
server. They are using CPU, RAM, and storage resources of the ESXi server (Figure 5.4).

39
 eBook

Figure 5.4

5. Domain Controller and vCenter Server are running on ESXi Server. They are using CPU,
RAM, and storage of the ESXi server, but virtual disks of these VMs are stored on the shared
datastore (Figure 5.5). This connection method offers similar advantages to Scenario 3.

Figure 5.5

40
 eBook

To create a cluster, take the following steps:

1. Log in to vCenter with vSphere Client or vSphere Web Client. In order to log in, enter
vCenter IP and port 9443 in the address bar.

NOTE: vSphere Client is a C#-based locally installed application for Windows only. vSphere
Web Client is a cross-platform web application. Both vSphere Client and vSphere Web Client
can connect to vCenter, or directly with ESXi hosts with the full range of administrative
functionality. Since vCenter Server 5.1, VMware recommends using the vSphere Web Client
to administer virtual environments. For older versions of vSphere Web Client, you may need
to install a Flash player plugin. The newest versions of vSphere Web Client use HTML5 rather
than Flash.

This is how the main screen of vSphere Web Client interface looks (see Figure 5.6):

Figure 5.6

2. In the left pane of the vSphere Web Client interface, select vCenter -> Datacenters –>
New Datacenter to create a new datacenter.

NOTE: A datacenter is a container for all the inventory objects required for a fully functional
environment for virtual machine operation. You can create multiple datacenters for each
department in your enterprise, or for different purposes, such as low- and high-performance
tasks. Your virtual machines can hot-migrate from one ESXi host to another ESXi host within
the same datacenter. However, they cannot migrate from a host in one datacenter to a host in
a different datacenter.

41
 eBook

3. Type the name for your new datacenter (in our case – “Datacenter03”; (see Figure 5.7).

Figure 5.7

4. Add your ESXi host(s) to the Datacenter. Right-click on the newly created datacenter
and select Add Host. Enter the IP address and root credentials of your ESXi host on the
Connection Settings page of the Add Host Wizard. Do the same for each ESXi host you want
added to the datacenter.

Now, right-click on the newly created datacenter and select the Create New Cluster option
(see Figure 5.8).

42
 eBook

Figure 5.8

5. Set a name for your cluster. Leave the checkboxes for the DRS and vSphere HA options
unticked (you can add these functionalities later; see Figure 5.9).

Figure 5.9

6. Add ESXi hosts to the cluster. Note that your ESXi hosts must belong to the same
datacenter. Click “Add Host…” (see Figure 5.10).

Figure 5.10

7. Enter the name or IP address of the ESXi host you want added to your cluster (see Figure
5.11).

43
 eBook

Figure 5.11

8. Enter the username and password for the administrative account of the ESXi host (see
Figure 5.12). The root user is the administrator by default.

Figure 5.12

9. Assign a license, if needed. You may try a 60-day Evaluation license.

10. Leave the lockdown mode disabled and click Next (see Figure 5.13).

Figure 5.13

44
 eBook

11. Set the resource pool. This option defines what to do with existing virtual machines and
resource pools on the ESXi host. If there are no virtual machines on the ESXi host, accept the
default option. Click Next (see Figure 5.14).

Figure 5.14

12. At the final screen, click Finish.

NOTE: Repeat steps 6 through 12 for each ESXi host that you want added in the cluster.

Configuring Network for the Cluster

To configure the network for your cluster, take the following steps:

1. Go to ESXi host -> Manage -> Networking.

NOTE: You can view or change the settings of physical network controllers, virtual network
controllers, and virtual switches from your vSphere Web Client (see Figure 6.1).

45
 eBook

Figure 6.1

Physical adapters are hardware network adapters of ESXi servers (see Figure 6.2).

Figure 6.2

The standard Maximum Transmission Unit (MTU) of a Frame is 1500 Bytes. In the Properties
tab, you can set the MTU value as high as 9000 Bytes to enable using Jumbo Frames, if
needed (see Figure 6.3).

46
 eBook

Figure 6.3

The VMkernel network adapter is used to provide network connectivity for hosts as well as
handling system traffic for vSphere vMotion, IP storage, Fault Tolerance logging, and vSAN
(see Figure 6.4).

Figure 6.4

2. Add a Virtual Switch

If there are more than two Network Interface Controllers (NICs) available in the ESXi server,
create two virtual switches. One of them should host the Service Console and VMkernel
(including iSCSI as well as vMotion traffic). The other should be dedicated to virtual machine
traffic. The NICs carrying the iSCSI traffic should be connected to redundant Ethernet
switches (see Figure 4.5 above).

There are two types of virtual switches (vSwitches) in vSphere: standard and distributed.
The standard virtual switch is configured manually on each host and is used for small
environments. The Distributed vSwitch allows you to manage networks for multiple hosts
from a single vCenter interface. Use the standard switch for the purposes of this walkthrough.

To edit virtual switches, go to Manage –> Networking –> Virtual switches (Figure 4.6 above).

47
 eBook

Select a virtual switch and click the Edit Settings icon. In the Teaming and Failover tab, you
can set active adapters, standby adapters, and unused adapters.

Active adapters use the uplink if the network adapter connectivity is up and active.

Standby adapters use this uplink if one of the active physical adapters is down.

3. Click the Add host networking icon (see Figure 6.5).

Figure 6.5

Here you can select VMkernel Network Adapter (see Figure 6.6).

Figure 6.6

4. Select New standard switch as a target device (see Figure 6.7).

48
 eBook

Figure 6.7

5. Configure an IP address for the new VMkernel port and select the appropriate services
to be used for the cluster. These could include vMotion traffic, provisioning traffic, Fault
Tolerance logging, management traffic, vSphere replication traffic, vSphere replication NFC
traffic, and/or Virtual SAN traffic.

NOTE: If you already have a virtual adapter created, you can select this adapter and edit
settings by clicking the Edit Settings icon (see Figure 6.8).

Figure 6.8

Below, you can see an example of a vMotion virtual network scheme (see Figure 6.9):

49
 eBook

Figure 6.9

vMotion is a feature that lets you hot-migrate powered-on virtual machines from one ESXi
host to another. Enable vMotion if you want to create a DRS or HA cluster. Separate the
vMotion network, the storage network, and the production network. This can help you
prevent overloading and reduce network bandwidth usage (see Figure 6.10).

NOTE: The production network is the network to which your physical computers
(workstations), routers, printers, etc. are connected.

Figure 6.10

Now that you have created a cluster, you can set up the DRS and/or HA features.

How to Create a DRS Cluster

In order to create a Distributed Resource Scheduler cluster, take the following steps:

50
 eBook

1. Go to vCenter -> Cluster and select your cluster (in the example, our cluster is named
“temp-cluster”)
2. Click Settings, select vSphere DRS and click Edit (see Figure 7.1).

Figure 7.1

3. Mark the Turn on vSphere DRS checkbox.

4. Configure the Automation Level, Migration Threshold, and Virtual Machine Automation
settings. There are useful tips displayed in the vSphere client interface that can help you
understand what settings could best meet your needs (see Figure 7.2).
 Under Automation Level, you can select Manual, Partially Automated, or Fully Automated
for the level of Load Balancing automation.
 Migration Threshold is the option that controls how conservatively or aggressively the
DRS runs. You can set the threshold at a value from 1 to 5. Migration Threshold 1 means
that only migration recommendations of the highest priority are executed. The only such
instances are affinity rules that you set yourself, or a host entering Maintenance Mode.
Cluster imbalance or VM demand do not result in migrations. Aside from Migration
Threshold 1, Migration Threshold 2 is the most conservative and Migration Threshold
5 is the most aggressive. The higher the Migration Threshold, the more migrations are
allowed. Thus, at Migration Threshold 5, the DRS executes recommended migrations even
if they can only yield very slight benefits to performance or cluster balance, whereas at
Migration Threshold 2, the potential benefits must be very significant.
 Virtual Machine Automation is the option that allows setting custom automation levels
for VMs.

51
 eBook

Figure 7.2

5. Set the Power Management settings (see Figure 7.3).

Figure 7.3

VMware Distributed Power Management (DPM) supports three power management protocols
to bring a host out of standby mode:
 Intelligent Platform Management Interface (IPMI);
 Hewlett Packard Enterprise Integrated Lights-Out (iLO);
 Wake-on-LAN (WOL).

52
 eBook

Each of these protocols requires separate hardware support and configuration. If a host does
not support any of these protocols, this host cannot be put into a standby mode by the DPM.
If a host supports multiple protocols, they are used in the following order: IPMI, iLO, WOL.

6. Add Affinity Rules, if necessary.

NOTE: Affinity Rules allow you to control the placement of virtual machines that interact in
particular ways with each other. For example, if you run a database server, a web server, and
an application server on different virtual machines, and they interact closely, you can create
an affinity rule to place ensure they reside on the same ESXi host. This would reduce network
load and can increase performance.

Go to vCenter -> Hosts and clusters. Select your cluster. Then click the Manage tab ->
Settings -> DRS Rules (see Figure 7.4).

Figure 7.4

Choose from three types of affinity rules: Keep Virtual Machines Together (affinity), Separate
Virtual Machines (anti-affinity), and Virtual Machines to Hosts (affinity or anti-affinity)

How to Create a HA Cluster

In order to create a High Availability cluster, take the following steps:
1. Go to vCenter -> Cluster and select your cluster (temp-cluster in this walkthrough’s case).
2. Click Settings.
3. Select vSphere HA and click Edit (see Figure 7.5).

53
 eBook

Figure 7.5

4. Tick the Turn on vSphere HA checkbox and set the following options (see Figure 7.6).

Figure 7.6

 Host Monitoring: If this option is enabled, frequent checks are performed to ensure each
ESXi host in the cluster is running. If a host failure occurs, the relevant virtual machines
are restarted on another host. Host Monitoring is also required for the VMware Fault
Tolerance recovery process to function properly. Remember to disable Host Monitoring
when performing network maintenance.

54
 eBook

 Virtual machine options. These options define how High Availability should react to host
failures and isolations:

• VM Restart Priority – determines the relative order in which virtual machines are
restarted after a host failure. You can specify the priority level for each VM: Disabled,
Low, Medium, or High. You can set your main VMs, such as those that run the Domain
Controller, the database server, and/or the email server, to restart with high priority.
• Host Isolation Response – three options are available here:
–– Leave powered on – when a network isolation occurs for the ESXi host, the state of
the virtual machines on the host remains unchanged. The virtual machines on the
isolated host continue to run, even if the host can no longer communicate with other
hosts in the cluster. This setting reduces the chances of a false positive.
–– Power off, then failover – when a network isolation occurs, all virtual machines
are powered off and restarted on another ESXi host. This is a hard stop. A power-
off response is initiated on the fourteenth second, and a restart is initiated on the
fifteenth second.
–– Shutdown, then failover – when a network isolation occurs, all virtual machines
running on that host are shut down via VMware Tools and restarted on another ESXi
host. This approach allows the services and programs that are running on the virtual
machines to be stopped correctly. If shutdown is not successful within 5 minutes, a
power-off response type is executed.
 Admission control (see Figure 7.7). Admission control is used by vCenter to ensure that
sufficient resources are available in a cluster for failover protection. The cluster reserves
resources to allow failover for all running virtual machines on the specified number of
hosts.

Figure 7.7

55
 eBook

Each Admission Control Policy has a separate Admission Control mechanism. Slots dictate
how many virtual machines can be powered on before vCenter triggers the “Out of
resources” notification. The Admission Control process is a function of vCenter, and not of
the ESXi host.


The percentage of Cluster Resources Reserved is the least restrictive and most flexible
Admission Control policy. 25% is the default reserved percentage, meaning that 25% of the
total CPU and total memory resource across the entire cluster is reserved for the cluster.


Failover hosts are the ESXi hosts that are reserved for a failover situation. Failover hosts
don’t factor into DRS recommendations or migrations, and virtual machines can’t run on
these hosts in the regular mode.


NOTE: Remember to enable Admission Control, because this option guarantees the ability
of virtual machines restart after a failure.
 VM Monitoring. This service evaluates whether each virtual machine in the cluster is
running by checking for regular heartbeats and input/output activity from the VMware
Tools process running inside the guest. VM Monitoring is different from the host
monitoring in that the item being watched is an individual virtual machine, rather than an
ESXi host. If vSphere can’t detect VM heartbeats, the VM reboot happens. You can select
the level of sensitivity using a preset, or set the failure interval, the minimum uptime, and
the maximum per-VM resets manually (see Figure 7.8).

Figure 7.8

 Datastore Heartbeating (see Figure 7.9). If the management network of an ESXi host
becomes isolated but the virtual machines are running, a restart signal is sent. Datastore
Heartbeating is used to determine more accurately the state of the ESXi host, even if the
management network fails. Thus, the feature reduces the probability of falsely triggering
the virtual machine reboot mechanism. There are locking mechanisms to prevent

56
 eBook

concurrent usage of open files located on shared storage and avoid file corruption. HA
manages the existing Virtual Machine File System (VMFS) locking mechanism, which is also
called a “Heartbeat Region”; this is updated as long as the lock file exists. HA determines
that at least one file is opened on the VMFS volume by checking files specially created for
Datastore Heartbeating. These files are named in the VMname-hb format: WindowsVM-
hb, LinuxTest-hb, host1tst-hb, etc. You can find them in the .vSphere-HA directory which
is located on the shared datastore with vSphere Client. Go to Home -> Datastores ->
DatastoreName -> Manage -> Files. Don’t delete or modify these files.

Figure 7.9

5. Click OK to finish the High Availability setup and wait while vSphere HA is configured (see
Figure 7.10).

Figure 7.10

57
 eBook

Fault Tolerance: Purpose and Setup

Fault Tolerance (FT) provides continuous availability for virtual machines and enables them
to survive a physical server failure. With this great feature, you can have an exact and
continuously available replica VM on another ESXi host that can take over from the virtual
machine affected at the time of failure.

In High Availability mode, virtual machines need some time to load on another ESXi host after
the ESXi host on which they were running fails. With Fault Tolerance, a virtual machine has a
copy running on another ESXi host with disabled VM network connection. If the ESXi host with
the primary copy of a VM fails, the secondary copy on another ESXi host just needs to have
networking enabled; this is why the migration process looks seamless. If there are more than
two ESXi servers included in the cluster, vSphere HA runs the replica of the VM on the second
ESXi server at the moment of failure, then creates a new VM replica on a third ESXi server.

Requirements for Virtual Machines with FT (v.6.0)

 10-Gigabit network adapters on ESXi hosts;
 Maximum 4 vCPU;
 Maximum 64 GB RAM;
 Maximum 4 VMs to run simultaneously on one ESXi host;
 Disabled CPU Hot-Add and RAM Hot-Add;

NOTE: The following restrictions apply:

 If more than one vCPU is used, Storage vMotion is not supported.
 Snapshots and thin disks are supported.
 HA cluster needs to be configured.

How to Enable Fault Tolerance

To enable FT for virtual machines, do the following:
1. Select your cluster and go to the list of virtual machines.
2. Right click on the virtual machine you want to make fault-tolerant: All vCenter Actions –>
Fault Tolerance –> Turn On Fault Tolerance (see Figure 8.1).

58
 eBook

Figure 8.1

A confirmation message appears for vSphere version 5.5, informing you of disk conversion
(see Figure 8.2). If you turn on Fault Tolerance, thin-provisioned disks, as well as disks that zero
out blocks as they were written to (lazy-zeroed thick-provisioned disks), become disks with
all blocks zeroed out (eager-zeroed, thick-provisioned disks). This conversion means that the
virtual machine uses more disk space and needs some processing time. This warning does
not appear in vSphere 6, as thin-provisioned disks are supported.

Figure 8.2

59
 eBook

3. Click Yes to continue.

The virtual machine Fault Tolerance status is displayed in VM -> Summary tab.
4. Now you can set up Latency Sensitivity (a high performance mode for the virtual
machine). Go to VM –> Manage –> Settings –> Advanced settings (see Figure 8.3).

Figure 8.3

With “High” Latency Sensitivity, the ESXi host provides vCPU access to physical CPU while
calculating the actual CPU load. With this option enabled, a virtual machine processor can
interact directly with the physical processor, without using the VMkernel scheduler. Thus, the
Latency Sensitivity mode is useful for virtual machines demanding high performance.

Here is an example of how High Availability and Fault Tolerance features work.
Both of ESXi servers are running in a High Availability cluster. The virtual machine VM2 is
running on ESXi Server 1 with the Fault Tolerance option enabled, and has an exact replica
with disabled networking running on ESXi Server2. VM1 is also running on ESXi Server 1, but
the Fault Tolerance option is disabled for this virtual machine (see Figure 8.4):

60
 eBook

Figure 8.4

Now, a failure occurs for ESXi Server 1. VM2, which was running on ESXi Server 1, also fails,
but the replica of VM2 that is still running on ESXi Server 2 becomes reachable in an instant;
networking is enabled for the replica by the automatic failover protection of the VMware Fault
Tolerance feature. VM2’s failover is seamless and instant. VM1 becomes unreachable because
of the same Server 1 failure, but since there is no replica of VM1, this virtual machine must be
migrated to ESXi Server 2. Loading VM1’s operating system and other services may take some
time (see Figure 8.5):

Figure 8.5

If you decide to turn off Fault Tolerance, click on the respective virtual machine and select: All
vCenter Actions –> Fault Tolerance –> Turn Off Fault Tolerance.

61
 eBook

NOTE: There is a difference between disabling Fault Tolerance and turning off Fault Tolerance.
If you disable FT, the secondary virtual machines are preserved with their configuration and
history. Using this option allows you to re-enable FT in the future. Turning off VMware FT
deletes all secondary virtual machines, their configurations, and their history. Use this option
if you do not plan on re-enabling VMware FT (see Figure 8.6).

Figure 8.6

Removing Hosts from the Cluster

If, for any reason, you decide to remove an ESXi host from the cluster, take the following
steps:
1. Power off all virtual machines that are running on the host.
2. Right-click the ESXi host and select Enter Maintenance Mode (see Figure 9.1).

Figure 9.1

62
 eBook

A confirmation prompt appears (see Figure 9.2).

Figure 9.2

3. Right-click the ESXi host you want to remove from the cluster and select Move To... (see
Figure 9.3).

Figure 9.3

4. Select a new location for the host (see Figure 9.4).

63
 eBook

Figure 9.4

5. Right-click the ESXi host and select Exit Maintenance Mode (see Figure 9.5).

Figure 9.5

64
 eBook

More protection of your cluster with NAKIVO Backup &

Replication
NAKIVO Backup & Replication is a fast, reliable, and affordable solution for protecting VMware
environments. The product offers image-based, application-aware backup and replication, as
well as a wide variety of recovery options for DR scenarios, all from an intuitive web UI.
The solution ensures that backups are small by using VMware’s Changed Block Tracking (CBT)
technology to identify and copy only the data that has changed since the last increment.
Repository-wide deduplication and several compression options mean that your backups will
occupy minimal storage space and can be transferred quickly.
LAN-free data transfer and Network Acceleration features can boost your VM backup speed,
making the process even shorter.
The product also offers backup-to-cloud functionality and automated failover to prevent data
loss or downtime in the event of a site-wide disaster. Screenshot verification for completed
jobs allows you to check that your VMs are recoverable. When disaster strikes, you can use
Flash VM Boot to instantly recover an entire VM or restore files or application objects to their
source location granularly.
Easy scheduling and management of all types of jobs (backup, replication, backup copy,
or failover) from the Calendar Dashboard, as well as the option to chain them, makes the
solution suitable for businesses of all sizes.
NAKIVO Backup & Replication provides the following cluster-friendly features for the
protection of VMs residing on ESXi hosts:

Adding VMware containers to inventory

Entire VMware containers, such as standalone ESXi hosts or vCenter servers, can be added to
the inventory of NAKIVO Backup & Replication. When an ESXi host or vCenter server is added,
the VMs, VM folders, datacenters, clusters, and resource pools that belong to it are added
automatically. Thus, you can add all the needed objects to your inventory quickly and easily.

Backup of VMware containers, including clusters

Clustering can provide high availability for your VMs, but you can increase your protection
level by creating backups and replicas of your VMs. This approach drastically increases your
chances of successful failover and disaster recovery. By using NAKIVO Backup & Replication,
you can select VMware containers with all the objects inside them for backup. When you
select an ESXi host to be backed up, all VMs residing on the host are selected as well. When
you select a cluster, all VMs that belong to the cluster are selected; thus, you can back up the
entire cluster in a few clicks. All new VMs added to a container that is protected with NAKIVO
Backup & Replication are automatically included in subsequent backup jobs. (You are free to
manually select additional VMs or deselect automatically included VMs and other objects in
the job options.)

65
 eBook

Automatic tracking of VMs migrating within a cluster

VMs can migrate from one ESXi host to another within a cluster during failover or load
balancing events. NAKIVO Backup & Replication tracks the migrating VMs within a cluster
automatically. As a result, such VMs are always protected by the product and can be backed
up or replicated without any added effort.

Replication of VMs from one cluster to another

NAKIVO Backup & Replication can replicate VMs to a cluster. Thus, you can replicate the
business-critical VMs from a local cluster to a remote cluster and use the resulting VM replicas
for failover during disaster recovery. NAKIVO Backup & Replication includes a feature that
called Automated VM Failover, with which you can perform failover using a VM replica. With
Network Mapping and Re-IP features, your replicas are automatically adapted to the new
virtual environment; the network settings of the VMs are changed automatically during
failover. All you need to do is create a rule when configuring the replication or failover job.

66
 eBook

Conclusion
VMware vSphere is a virtual environment with a long list of features that help manage virtual
machines, provide great capability, reliability, and scalability. Clustering technologies are
widely used in vSphere to connect servers over the network and achieve better performance
in executing resource-intensive tasks. VMware supports creating Distributed Resource
Scheduler (DRS) clusters and High Availability (HA) clusters. Creating a DRS cluster helps
improve performance by rational usage of computing resources. A HA cluster reduces the
downtime of virtual machines in the event of failure by restarting VMs on another host via
redundant network. Fault Tolerance feature of HA clusters ensures avoiding the downtime
and provides for the seamless migration of virtual machines from the failed host to the
running host. That is vital for business critical processes. Using vSphere HA and DRS together
combines automatic failover with load balancing. This helps provide a more balanced cluster
after vSphere HA moves virtual machines to different hosts.

High Availability and Fault Tolerance do not replace the need for data backup. In the event of
cluster usage, virtual machines are stored on a shared datastore and should be backed up to
another storage. A combination of VMware cluster features with backup ensures the efficient
resource management, increased reliability, and protection.

67
 eBook

NAKIVO Backup & Replication at a Glance

NAKIVO Backup & Replication is a fast, reliable, and affordable VM backup solution.
The product protects VMware, Hyper-V, and AWS EC2 environments. NAKIVO Backup &
Replication offers advanced features that increase backup performance, improve reliability,
and speed up recovery. As a result, you can save time and money.

Deploy in under 1 minute Reduce backup size

Pre-configured VMware VA and AWS Forever-incremental backups with CBT/
AMI; 1-click deployment on ASUSTOR, RCT, LAN-free data transfer, network
QNAP, Synology, or WD NAS; 1-click acceleration; up to 2X performance
Windows installer, 1-command Linux when installed on NAS
installer

Protect VMs Ensure recoverability

Native, agentless, image-based, Instant backup verification with
application-aware backup and screenshots of test-recovered VMs;
replication for VMware, Hyper-V VMs, as backup copy offsite/to the cloud
well as AWS EC2 instances

Increase backup speed Decrease recovery time

Exclusion of SWAP files and partitions, Instant recovery of VMs, files, Exchange
global backup deduplication, adjustable objects, SQL objects, Active Directory
backup compression objects; DR with VM replicas

About NAKIVO
The winner of a “Best of VMworld 2018” and the Gold Award for Data Protection, NAKIVO is a
US corporation dedicated to developing the ultimate VM backup and site recovery solution.
With 20 consecutive quarters of double-digit growth, 5-star online community reviews,
97.3% customer satisfaction with support, and more than 10,000 deployments worldwide,
NAKIVO delivers an unprecedented level of protection for VMware, Hyper-V, and Amazon EC2
environments.

As a unique feature, NAKIVO Backup & Replication runs natively on leading storage systems
including QNAP, Synology, ASUSTOR, Western Digital, and NETGEAR to deliver up to 2X
performance advantage. The product also offers support for high-end deduplication
appliances including Dell/EMC Data Domain and NEC HYDRAstor. Being one of the fastest-
growing data protection software vendors in the industry, NAKIVO provides a data protection
solution for major companies such as Coca-Cola, Honda, and China Airlines, as well as works
with over 3,000 channel partners in 137 countries worldwide. Learn more at [Link]

© 2018 NAKIVO, Inc. All rights reserved. All trademarks are the property of their respective owners.