VMware Notes

VMware Notes



ESX/ESXi Log files – hostd.log and messages

 Contain entries made during bootup and while the system is running



ESX also has vmkernel, vmksummary.txt and vmkwarning log files

 Track service console availability, VMkernel alerts, warning, messages and ESX host

availability



Remote command prompt management

 VMware vSphere Command Line Interface (vCLI) application

 VMware vSphere Management Assistant (vMA) virtual appliance

o A platform for running a variety of toolkits such as vCLI, vSphere SDK for Perl and

vSphere API

 VMware vSphere PowerCLI

o Automation tool for administering a vSphere environment

o Distributed as a snap in to Windows Power Shell



vCenter Server Maximums

 1,000 hosts

 10,000 powered on VMs

 15,000 registered VMs



vCenter Services

 Core Services – management of resources and VMs, task scheduling, statistics logging,

management of alarms and events, VM provisioning and host and VM configuration

 Distributed Services – vMotion, DRS and HA



vCenter Hardware and Software Requirements

 Two 64 bit CPUs or one 64 bit dual core processor (2.0 GHz or higher)

 3GB RAM minimum

 3GB disk storage minimum

 Gigabit network recommended

 64 bit OS – XP Pro 64 bit SP2, 2003 Enterprise SP2, 2008 R2 64 bit



vCenter Supported Databases

 SQL 2005 and 2008

 Oracle 10g and 11g

 IBM DB2 9.5

 SQL 2005 Express

vCenter Ports

 443 - HTTPS

 80 - HTTP

 902 – UDP heartbeat

 8080 - Web Services HTTP

 8443 - Web Services HTTPS

 60099 – Web services change service notification

 389 – LDAP

 636 – SSL



vCenter Java Virtual Machine (JVM) Memory – VMware VirtualCenter Management

Webservices

 Requires 1-4GB of additional memory

 1GB for less than 100 hosts

 4GB for more than 400 hosts



vCenter Windows Services

 VMware Mount Service for Virtual Center – Used during guest OS customization such

as cloning or deploying from a template.

 VMware vCenter Orchestrator Configuration – Used for Orchestrator which is a

workflow engine that helps admins automate existing manual tasks.

 VMware VirtualCenter Management Webservices – Allows configuration of vCenter

management services.

 VMware VirtualCenter Server

 VMwareVCMSDS – Provides vCenter Server LDAP directory services.





vCenter Plugins

 vCenter Storage Monitoring – Allows vCenter Server to monitor and report on

storage and adds the Storage Views tab to the vSphere client

 vCenter Service Status (health status) – adds the vCenter Service Status icon to the

Administration panel in the vSphere client

 vCenter Hardware Status – Allows vCenter to display the hardware status of the hosts

and adds the Hardware Status tab to the vSphere client.



vCenter Server uses the root account to add hosts to the inventory and creates a special user

account named vpxuser for all future authentication



Lockdown Mode (ESXi only) – Disables remote access for the administrator account to ensure

the host is only managed by vCenter



You can add ESX 2.5.x and later as well as ESXi 3.5 and later hosts to the vCenter Inventory

VPXA Process – vCenter Server agent that provides access to ESX/ESXi hosts

 Resides on ESX/ESXI host - Installed when the host is added to vCenter

 Vpxa process communicates with the host agent known as the hostd process to relay

the tasks to perform on the host.

 It’s not used if logged in directly to the host but instead communications use hostd

directly



vCenter can also manage licenses for legacy hosts such as ESX 3.x and ESXi 3.5 using a separate

license server.



You can’t have 2 Virtual switches (Layer 2 devices) mapped to the same physical NIC.

You can have 2 or more physical NICs mapped to the same virtual switch.



Standard virtual switch

 Maximum of 4,088 virtual switch ports per switch

 Maximum of 4,096 virtual switch ports per host

 120 switch ports created by default

 Virtual switch ports used for VM connections and for uplinks to physical NICs

 Some ports used for internal purposes by the VMkernel



VLANs

 Can be configured at the port group level

 ESX/ESXi hosts provide VLAN support through virtual switch tagging (gives a port group

a VLAN ID)

o VMkernel then takes care of all the tagging

 A switch port on the physical host must be defined as a static trunk port

 No VLAN configuration is needed on the VM



Network Policies

 Security, Traffic shaping and NIC teaming

 Defined at the standard virtual switch level for the entire switch

 Can also be defined for a VMkernel port, VM port group and ESX service console

 Policies defined for an individual port or port group override the default policies defined

for the switch.

Network security policy exceptions

 Promiscuous Mode – when set to reject, placing a guest adapter in promiscuous mode

has no effect on which frames are received by the adapter (default is Reject)

o Set Promiscuous mode to Accept if you want to use an application in a VM that

analyzes of sniffs packets.

 MAC Address Change – When set to Reject, if the guest attempts to change the MAC

address assigned to the virtual NIC, it stops receiving the frames. (default is Accept)

 Forged Transmits – When set to Reject, the virtual NIC drops any frames that the guest

sends, where the source address field contains a MAC address other than the assigned

virtual NIC MAC address (default is Accept)

 Set MAC Address Changes and Forged Transmits to Reject to help protect against

certain attacks launched by a rouge guest operating system.

 Leave MAC Address Changes and Forged Transmits at their default values of Accept if

you applications change the mapped MAC address.



Traffic shaping shapes outbound network traffic only when used on a standard virtual switch

 Off by default



ESX/ESXi hosts shape outbound traffic only by establishing parameters for 3 traffic

characteristics: Average Bandwidth, Peak Bandwidth and Burst Size.

 Establish the policy at the virtual switch level or the port group level

 Settings at the port group level override the settings at the switch level

Average Bandwidth

 Establishes the number of kilobits per second to allow across a port, averaged over

time.

 The average bandwidth is the allowed average load.

Peak Bandwidth

 The maximum number of kilobits per second to allow across a port when it is sending a

burst of traffic.

 This tops the bandwidth used by a port whenever the ports is using it burst bonus.

Burst Size

 The maximum number of kilobytes to allow in a burst.

 If this parameter is set, a port might gain a burst bonus if it does not use all its allocated

bandwidth.



NIC Teaming

 Policies include load balancing and failover settings

 Default policies are set for the entire Standard Switch

 Policies can be overridden at the port group level

 Virtual Port ID load balancing – a VMs outbound traffic is mapped to a specific physical

NIC. This method is simple and fast and does not require the VMkernel to examine the

frame for necessary information.

 MAC Hash load balancing – Each VMs outbound traffic is mapped to a specific physical

NIC’s MAC address. This method has low overhead and is compatible with all switches

but may not spread traffic evenly across physical NICs.

 IP Hash load balancing – A NIC for each outbound packet is chosen based on its source

and destination IP address. This method has higher CPU overhead but a better

distribution of traffic across physical NICs. It also requires 802.3ad link aggregation

support or EtherChannel on the switch.



Network failure is detected by the VMkernel, which monitors:

 Link state only – Detects cable pulls and physical switch failures. Doesn’t detect

configuration errors.

 Link state plus beaconing – VMkernel sends out and listens for probe packets on all

NICs in the team

Switches can be notified whenever:

 There is a failover event

 A new virtual NIC is connected to the virtual switch

Failover implemented by the VMkernel based on configurable parameters:

 Failback – Determines how a physical adapter is returned to active duty after recovering

from a failure. If set to Yes, the failed adapter is returned to active duty immediately

after recovery, displacing the standby adapter that took its place. If set to No, the failed

adapter is left inactive after recovery until needed.

 Load balancing option: Use explicit failover order – Always use the highest order uplink

from the list of active adapters that pass failover detection criteria.





VMFS

 A clustered file system that allows multiple physical servers to read and write to the

same storage simultaneously.

 A VMFS datastore can be configured to use an 8MB block size to support virtual disk

files up to 2TB.

 A VMFS datastore uses subblock addressing to make efficient use of storage for small

files.



NFS

 File sharing protocol used to communicate with a NAS device

 NFS datastores are treated like VMFS datastores – can hold VM files, ISOs, templates

and use vMotion etc.

 ESX/ESXi supports NFS version 3 over TCP only

 ESX/ESXi hosts do not use the standard Network Lock Manager (NLM) protocol

 VMware uses its own locking protocol. NFS locks are implemented by creating lock files

on the NFS server. Lock files are named .lck-, where is the value of the

fileID field. The lock file generates small 84-byte WRITE requests to the NFS server.

RDM – Raw Device Mapping

 Acts as a proxy for a raw physical device

 Stores data directly on a raw LUN

 Can be used for data, VM clustering and storage array snapshots

 Allows you to use your existing SAN commands to manage storage for the disk

 Used when clustering VMs using Microsoft Clustering Service (MSCS)

 A VMFS datastore can be used to hold RDMs that point to raw iSCSI volumes.



Thin provisioning disks can reduce the cost of storage for virtual environments by up to 50%.



SCSI Storage Devices

 Use a SCSI ID – The unique address of a SCSI device

 Canonical name – The Network Address Authority ID. Globally unique identifiers that

are persistent across system reboots.

 The T10 identifier is another unique identifier. It can appear on any SCSI device.

They always begin with the string t10

 mpx is a VMware namespace that is used when no other valid namespaces can

be obtained from the LUN. It is not globally unique or persistent across reboots.

 Runtime name is the name of the first path to the device. It is created by the host. It is

not reliable or persistent.



ESX/ESXi support 2 types of IP storage

 iSCSI – Used to hold one or more VMFS datastores

 NFS – Used to hold one or more NFS datastores

 Both support vSphere features like vMotion, HA and DRS

ESX/ESXi supports:

 Up to 64 NFS volumes

 iSCSI or NFS over a 10GbE interface

 iSCSI or NFS in an IPv6 environment (experimental only)

ESX/ESXi supports booting from an iSCSI SAN

 ESX hosts: From independent hardware iSCSI

 ESXi hosts: From software iSCSI and dependent hardware iSCSI – The network adapter

must support only the iSCSI Boot Firmware Table (iBFT) format.



The ESX/ESXi host is configured with a software or hardware iSCSI initiator

 Hardware is an iSCSI HBA

 Software is an iSCSI Initiator

 Software initiator is VMware code built into the VMkernel.

 Hardware initiator is a 3rd party adapter capable of accessing iSCSI storage over TCP/IP

 The Dependent hardware initiator depends on VMware networking and on iSCSI

configuration and management interfaces provided by VMware. You need to bind the

adapter and an appropriate VMkernel iSCSI port.

 An independent hardware adapter handles all the iSCSI and network processing and

management for the ESX/ESXi host.



LUN Masking is available for iSCSI and works the same as in Fibre Channel.



Ethernet switches don’t use Zones like FC but rather use VLANs instead.



iSCSI Names

 iSCSI qualified name (IQN) or the Extended Unique Identifier (EUI)



IQN - iSCSI qualified name

 Can be up to 255 characters long

 Uses the prefix iqn

 Has a date code specifying the year and month in which the organization registered the

domain or subdomain uses as the naming authority string

 Has an organizational naming authority string which consists of a valid, reversed domain

or subdomain name

 May have a colon (:) followed by a string of the assigning organization’s choosing

 Example - iqn.2001-04.com.example or iqn.2001-04.com.example:storage.disk2.sys1.xyz



EUI - Extended Unique Identifier

 Uses the prefix eui followed by a 16 character name. The name includes 24 bits for a

company name that is assigned by the IEEE and 40 bits for a unique ID, such as a serial

number

 Example - eui.02004567A425678D



Configuring a iSCSI software initiator

 Create a VMkernel port on a virtual switch

 Enable the software iSCSI initiator

 Configure one or more target discovery addresses so that the iSCSI initiator can

determine which storage resources on the network is available for access. You cannot

change the IP address, iSCSI name, or port number of an existing target. To make

changes, remove the target and make a new one

 Configure Challenge Handshake Authentication Protocol (CHAP) if needed.

 ESX/ESXi supports per-target CHAP where you use different credentials for each target

(Software iSCSI only)

ESX/ESXi supports two iSCSI target discovery methods

 Static Discovery – The initiator does not need to perform discovery. It knows in advance

all the targets it will be contacting and uses their IP addresses and domain names to

communicate with them.

 Dynamic Discovery (SendTargets discovery) – Each time the initiator contacts a

specified iSCSI server; it sends the SendTargets request to the server. The server

responds by supplying a list of available targets to the initiator. The names and IP

addresses of these targets appear as static in the vSphere client.





CHAP

 Unidirectional (one way CHAP) – the target authenticates the initiator, but the initiator

does not authenticate the target. You specify the CHAP secret. (Hardware and software

iSCSI).

 Bidirectional (mutual CHAP) – The initiator is able to authenticate the target as well

(Software iSCSI only).

 Only Unidirectional CHAP is available for hardware initiators



Configuring the iSCSI initiator

 Install the iSCSI hardware adapter

 Modify the iSCSI name and configure the iSCSI alias

 Configure iSCSI target addresses

 Configure iSCSI security (CHAP)



NFS Privileges

 NFS privileges are assigned to the root user

 When root_squash is on, the NFS server treats access by the root user as access by any

unprivileged user and might refuse the ESX/ESXi host access to VM files stored on the

NFS volume.

 You must use the no_root_squash option instead to export an NFS volume. It allows the

root user to be recognized as root.

 The NFS administrator must allow read and write privileges to the NFS datastore with

no_root_squash if you are deploying VMs on the NFS datastore.



Configuring an NFS Datastore

 Separate it from the iSCSI network for better security and performance

 Provide the NFS server name or IP address

 Provide the folder on the NFS server

 Choose whether to mount the NFS file system read-only or not. Use read-only for ISO

libraries and read/write for VMs

 Choose the NFS datastore name



To see NFS datastores go to the Storage Views tab and display the Show all NAS Mounts

To unmount an NFS datastore right click the datastore and select Unmount or select the

datastore and click the Delete link.



Fibre Channel

 ESX/ESXi supports 8Gb Fibre Channel and Fibre Channel over Ethernet (FCoE)

 You can boot ESX/ESXi from a Fibre Channel SAN LUN – The BIOS of the Fibre Channel

adapter must be configured with the World Wide Name (WWN) and LUN number of the

boot device.



A Fibre Channel SAN consists of:

 Storage System – Physical hard disks (array) and one or more intelligent controllers. The

storage system supports the creation of LUNs (logical volumes)

 LUN – The address of a logical unit (LU). An LU can be a JBOD (Just a Bunch Of Disks), a

RAID set or part of a storage container

 Storage Processor – A storage processor can partition a JBOD or RAID set into one or

more LUNs. Each connection is referenced by the HBA’s WWN.

 HBA – Connects the ESX/ESXi host to the Fibre Channel network. A minimum of 2 HBA

adapters are used for FT

 Fibre Channel Switches – One or more Fibre Channel switches form the Fibre Channel

fabric. The Fibre Channel fabric interconnects multiple nodes.



Soft Zoning – Controls LUN visibility per WWN and is done at the Fibre Channel switch



Hard Zoning – The control of storage processor visibility per switch port



Fabric Zoning – Controls target presentation and tells an ESX/ESXi host whether a target exists



WWNs are assigned by the manufacturer of the SAN. HBAs and storage processors have WWNs.

They are used to identify equipment for zoning purposes.



LUN Masking – Controls LUN visibility per host. Can be done in the ESX/ESXi host or at the

storage processor level (more secure and better data integrity)



The VMkernel scans for LUNs 0-255 (256 total). You can’t have a LUN with an ID over 256.



The Storage Views tab allows you to review associations between all storage entities available

in vCenter and analyze storage usage. Reports are updated every 30 minutes.

VMFS

 Use VMFS 3 datastores whenever possible

 VMFS is optimized for storing and accessing large files

 A VMFS can have a maximum volume size of 64 TB (32 x 2TB -512k extents)

 Offers some functions that NFS doesn’t support

 Use RDMs if your VM is performing SAN snapshotting, is clustered using MSCS or has

large amounts of data that you don’t want to convert into a virtual disk



You cannot store an RDM on an NFS datastore but you can store an RDM on a VMFS datastore



You cannot use MSCS to cluster a VM that resides on a NFS datastore



Overcommitted datastore - When there are many thin provisioned virtual disks that use close

to their allotted disk space.



Increasing the size of a VMFS datastore

 Add an extent to the VMFS datastore. You can add any extent to any VMFS datastore up

to 32 extents

 Expand the VMFS datastore. Increase the size of the VMFS datastore within its extent if

it has free space



To expand a RDM’s underlying raw LUN on the array, you have to remove the RDM and re-

create it



Deleting a VMFS datastore permanently deletes the pointers to the files on the datastore, so

the files cannot be retrieved



Multipathing allows continued access to SAN LUNs in the event of hardware failure and also

provides load balancing



Hardware iSCSI Multipathing

 Use 2 or more hardware iSCSI adapters



Software or dependent hardware iSCSI Multipathing

 Use multiple NICs

 Connect each NIC to a separate VMkernel port

 Associate VMkernel ports with an iSCSI initiator so that each VMkernel port connected

to a separate NIC becomes a different path



Multiple paths can exist to a datastore on an ESX/ESXi host

 Click the host’s Configuration tab

 Click the Storage link

 Right click the datastore and select Properties

 Click Manage Paths



Path selection policies

 Fixed – The host always uses the preferred path to the disk when that path is available.

Fixed is the default policy for active-active storage devices.

 Most Recently Used – The host uses the most recent path to the disk until this path

becomes unavailable. The host does not revert back to the preferred path. Most

Recently Used is the default and required type for active-passive storage devices.

 Round Robin – The host uses a path-selection algorithm that rotates through all

available paths. RR supports load balancing across the paths.



Pluggable Storage Architecture (PSA)

 A VMkernel layer responsible for managing multiple storage paths

 A collection of VMkernel APIs that allow third party vendors to insert code directly into

the ESX/ESXi storage I/O path (multipathing plug-ins MPPs)

 VMware provides a generic MPP by default called Native Multipathing Plug-in (NMP)





When naming VMs, its best practice to avoid using special characters including spaces in the

name since the VM name is used to name the files that make up the VM.



Files that make up a Virtual Machine

 .vmx – Virtual machine configuration file

 .vmdk – File describing virtual disk characteristics

 .-flat.vmdk – Pre-allocated virtual disk file that contains the data

 .nvram – Virtual machine BIOS

 Vmware.log & vmware-#.log – Virtual machine log file and files containing old virtual

machine log entries

 .vswp – Virtual machine swap file

 .vmsd – File that describes the virtual machine’s snapshots

 .vmtx – Virtual machine template configuration file

 If a VM is converted to a template, a virtual machine template configuration file (.vmtx)

replaces the virtual machine configuration file (.vmx)

 If a VM has more than one disk file, the file pair for the second disk file and later is

named _#.vmdk and _#.-flat.vmdk where # is the next

number in the sequence starting with 1.

 6 of the archive log files are maintained at any one time. Name-1.log, name-2.log etc.



A virtual disk consists of 2 files

 The .vmdk files which describes the virtual disk’s characteristics

 A –flat.vmdk file which contains the virtual disk’s data

 The datastore browser only shows the .vmdk file

You can add multiple USB devices to a VM that resides on an ESX/ESXi host to which the device

is physically attached. The device is only available to VMS that reside on that host. A USB device

is available to only one VM at a time.



VMware virtual SNP allows you to take advantage of configuring a virtual machine with up to 8

virtual CPUs, allowing larger CPU-intensive workloads to run on ESX/ESXi hosts.



Adding the first virtual disk to a VM implicitly adds a virtual SCSI adapter for it to be connected.

ESX/ESXi offers a choice of adapters:

 BusLogic Parallel

 LSI Logic Parallel

 LSI Logic SAS

 VMware Paravirtual









Independent disk mode

 Persistent – Use if you want changes to be immediately and permanently written to disk

 Nonpersistent – Use if you want to discard changes when the VM is powered off or

reverted to a snapshot

Virtual Machine Network Adapters

 Flexible – Functions as a vlance adapter if VMware tools is not installed on the VM. It

functions as a vmxnet driver if VMware tools are installed on the VM.

 vlance – An emulated version of the AMD 79C970 PCnet32 LANCE NIC. Drivers are

available in most 32 bit operating systems.

 vmxnet – A virtual network adapter that has no physical counterpart or vendor drivers

and is optimized for performance in a VM. The VM needs to have VMware tools

installed.

 e100 – An emulated version of the Intel 8254EM Gigabit Ethernet NIC with drivers

available in most newer operating systems. It’s the default adapter for 64 bit guest

operating systems.

 vmxnet2 (Enhanced vmxnet) – Based on the vmxnet adapter but provides high

performance features commonly used on modern networks such as jumbo frames and

hardware off-loads.

 vmxnet3 – The next generation of paravirtualized NIC designed for performance. It’s not

related to vmxnet or vmxnet2. It offers all the features of vmxnet2 plus multiqueue

support (Receive-Side Scaling in Windows), IPv6 off-loads, MSI/MSI-X interrupt delivery,

fault tolerance and record/replay. Only supported by a limited set of guest operating

systems and only available on VMs with hardware version 7





The virtual CD/DVD drive can point to:

 The CD/DVD drive or floppy drive of the ESX/ESXi host

 A CD/DVD ISO image or floppy (.flp) image

 The CD/DVD or floppy on your local system



Features of VMware Tools

 Device Drivers

o SVGA display

o Bus Logic SCSI driver

o vmxnet/vmxnet3

o Balloon driver for memory management

o Sync driver for quiescing I/O

o VMware mouse driver

 Virtual Machine Heartbeat

 Time Synchronization

 Ability to shut down a virtual machine

 VMware Tools control panel

 Scripts to help automate guest operating system operations

 VMware user process – lets you copy/paste

Virtual Appliances

 Typically includes a preinstalled guest OS

 VAs are deployed as an Open Virtual Machine (OVF) template.

 To import VA go to File>Browse VA marketplace then complete the deploy OVF

template wizard to download it and add it to the vCenter Server inventory

 vSphere client allows you to import and export any file in OVF format

 Specify OVF filename or URL that points to the file

 Exporting VMs allows you to create virtual appliances that can be imported by other

users



VMs can be changed into templates without the need to make a full copy of the virtual machine

files and the creation of a new object



You can create a template by:

 Cloning a VM to a template

 Converting a VM to a template

 Cloning a template



When you clone a VM to a template, the original VM is maintained.

When you convert a VM to a template, the original VM is replaced by the template.

When you clone a template, you make a copy of a template.



Clone to Template offers you the choice of format in which to store the VMs virtual disks

 Same format as source

 Thin provisioned disk

 Thick format



Convert to Template does not offer a choice and leaves the VMs disk file intact.



View templates from the VMs and Templates inventory view or from Hosts and Clusters view by

selecting a container and clicking its Virtual Machines tab.



To deploy a VM from a template, right click the template and choose Deploy Virtual Machine

from this Template.



To convert a template to a VM, go to the VMS and Templates inventory view. Right click the

template and select Convert to Virtual Machine. You can also use the vCenter Update

Manager.



You can’t clone a VM if connected directly to an ESX/ESXi host.

When you clone a VM that is powered on, services and applications are not automatically

quiesced when the VM is cloned.

 When you clone a VM or deploy from template, you can customize its guest OS

beforehand.

 Use the Guest Customization wizard during cloning or deployment.

 Or create customization specifications and apply to the new VM

 vCenter must be configured for customizations

 For Windows 2000, XP and 2003 you must install the Sysprep tools on the vCenter

Server

 Sysprep tools are built into Vista and 2008





You can provision VMs across datacenters in vCenter. You can also create a template in one

datacenter and then deploy a VM from that template into a different datacenter.



vCenter Converter tasks

 Converts physical machines to VMSs

 Convert and import VMs created by VMware Workstation or Microsoft Virtual Server

2005

 Convert third party backup or disk images to vCenter VMs

 Restore VMware Consolidated Backup images to vCenter VMs

 Export vCenter VMs to other VMware VM formats

 Reconfigure vCenter VMs so they are bootable

 Customize vCenter VMS



vCenter Converter Components

 vCenter Converter Server – Enables the import and export of VMs . Install it on a

vCenter Server or an independent server with access to vCenter Server

 vCenter Converter agent – Prepares a powered on physical or virtual machine for

import

 vCenter Converter client – Plugin which provides access to the vCenter Converter

Import, Export and Reconfigure wizards from the vSphere Client.

 Converter and Converter Client only run on Windows

 Converter supports Windows and Linux for importing and exporting

 Installation file – 100MB

 vCenter Converter client – 25MB

 vCenter Converter server – 300MB

 vCenter Converter agent – 100MB

 When performing a hot clone\live clone, vCenter Converter requires 350MB on the

source machine

 vCenter Converter supports only pure IPv4 or IPv6 environments and the source,

destination and vCenter Converter server and agent must run the same version of IP

vCenter Converter uses cloning and the destination virtual disk might not be an exact copy of

the source disk



4 stages of hot cloning performed by vCenter Converter

 Preparing the source machine for conversion

o vCenter Converter installs the vCenter Converter agent on the source machine

o The agent then takes a snapshot of the source volume

o vCenter Converter creates the snapshot with Microsoft’s Volume Snapshot

Service (VSS)

 Preparing the VM on the destination machine

o vCenter Converter creates a new VM on the destination ESX/ESXi host

 Completing the conversion process

o vCenter Converter installs required drivers to allow the OS to boot in the virtual

machine

 Cleaning Up

o The agent removes all traces from the source machine. The VSS snapshot

created in stage one is deleted and the vCenter Converter agent is uninstalled

from the source





The Import Machine wizard allows you to import from the following sources

 Powered on machine (physical or virtual)

 VMware infrastructure VM

 VMware Workstation or other VMware VM

 Backup image or third party VM supported by vCenter Converter

 Hyper-V Server VM



Data is copied to the destination using volume-based or disk-based cloning during importing

 Volume-based cloning

o Used for hot cloning and importing existing VMs

o All volumes in the destination VM are basic volumes regardless of the source

volume

o Volume based cloning at the file level is when you specify a size smaller than the

original volume

o Volume based cloning at the block level is performed when you specify the same

or a larger volume size

o Supports all types of source volumes that Windows recognizes

 Disk-based cloning

o Transfers all sectors from all disks and preservers all volume metadata

o The destination VM receives the same partitions, of the same type, as the

partitions of the source VM

o All volumes on the source machine’s partitions are copied as they are

o Disk based cloning supports all types of basic and dynamic disks

VM importing supports basic and dynamic volumes except RAID, Windows NT 4 fault-tolerant

and GUID partition table volumes.



Importing services

 You can select which services to stop before vCenter Converter synchronizes the data

between the source and destination machine.



You can transfer data for the second time by copying only the changes made during the first

transfer of data. This process is called synchronization. Only available for Windows XP or later

source operating systems.



Settings that remain identical include operating systems configuration, computer name, SID,

user accounts, profiles, preferences, applications and data files, and the volume serial number

for each disk partition.



Modifying VM Settings

 CPU hot plug – add CPU and memory to a VM while its powered on (enabled by default)

 You must install VMware Tools and the VM must use hardware version 7 or later

 The guest OS in the VM must support CPU and memory hot plugging

 The hot plug option must be enabled in the Options tab of the VM’s properties



You can increase the size of a virtual disk that belongs to a VM that is powered on if it is a flat

virtual disk in persistent mode and the VM does not have snapshots.



Raw Device Mapping (RDM)

 When you create a raw device mapping, vCenter Server creates a file in the specified

VMFS volume that points to the raw LUN

 Encapsulating disk information in the file (the RDM) allows the VMkernel to lock the

LUN so that only one virtual machine can write to it.

 An RDM supports 2 compatibility modes:

o Physical Compatibility mode

 Allows the guest OS to access the hardware directly. Useful if you are

using SAN-aware applications in the VM

 Cannot be cloned, made into a template or migrated if the migrations

involves copying to the disk

o Virtual Compatibility mode

 Allows the VM to use VMware snapshots and other advanced

functionality.

 Allows the LUN to behave as if it were a virtual disk

 Can be cloned or made into a template (content of LUN copied to a

virtual disk file - .vmdk)

Virtual Machine Snapshots

 Organized in a linear process or as a process tree

 Linear Process – Each snapshot has one parent and one child, except for the last

snapshot which has no children

 Process Tree – Each snapshot has one parent, but one snapshot can have more than

one child



A snapshot captures the entire state of the VM at the time you take the snapshot including:

 Memory State – The contents of the VMs memory (if powered on)

 Settings State – The VMs settings

 Disk State – The state of all the VMs disks



In the Snapshot Manager you can do 3 things:

 Delete – Commits the snapshot data to the parent snapshot and then removes the

selected snapshot

 Delete All – Commits all the intermediate snapshots before the current state icon (You

are here) to the base disk and removes all snapshots for that VM

 Go to – Allows you to restore, or revert to, a particular snapshot. The snapshot you

restore to becomes the current snapshot





A virtual machine can have one or more snapshots. Each snapshot consists of:

 Memory state file - -Snapshot#.vmsn where # is the next number the

sequence starting with 1

 Snapshot description file - -00000.vmdk – This file is a small text file that

contains information about the snapshot

 Snapshot delta file - -00000#-delta.vmdk – This file contains changes to

the virtual disk’s data at the same time the snapshot was taken



.vmsd is the snapshot list file, created at the time the VM is created. It contains

information about all the snapshots that belong to the VM. This information includes the name

of the snapshot .vmsn file and the name of the virtual disk file



To create a vApp, use the New vApp wizard and then modify its settings

 Resource allocation – Determines how CPU and memory should be allocated for the

vApp

 IP allocation policy – Determines how IP addresses are allocated for the vApp

o Fixed (static)

o DHCP

o Transient – IP addresses are automatically allocated using IP pools from a

specified range



The distribution format for a vApp is OVF

When you delete a VM from a datastore, it is removed from vCenter Server and all VM files are

deleted from the datastore



Concurrent VM migrations

 A host can be involved in up to 2 migrations with vMotion or Storage vMotion at one

time

 A maximum of 8 simultaneous vMotion, cloning, deployment, or Storage vMotion

access to a single VMFS-3 datastore is supported

o Maximum of 4 for a NFS or VMFS-2 datastore





Comparison of Migration Types









Storage Tiering – Migrating VMs from Fibre Channel to iSCSI or NAS or within or between

enclosures with Storage vMotion



Upgrading datastores without VM downtime with Storage vMotion

 You can migrate running VMs from a vMFS-2 datastore to a VMFS-3 datastore and

upgrade the VMFS-2 datastore without affecting VMs

Storage vMotion limitations:

 VMs with snapshots cannot be migrated with Storage vMotion

 VM disk must be in persistent mode or be RDMs

 You can’t do a vMotion and Storage vMotion at the same time with the VM powered on





Access Control – Defined with the following concepts

 Privilege – The ability to perform a specific action or read a specific property

 Role – A collection of privileges

 Object – An entity upon which actions are performed

 User or Group – A user or group who can perform the action

 The combination of a role, a user or group and an object equals a permission



Users who are in the Active Directory group ESX Admins are automatically assigned the

Administrator role. On ESXi you can use the Direct Console User Interface (DCUI) and technical

support mode to log in with AD accounts.



vCenter Server and ESX/ESXi hosts manage their own set of roles. A role that is created on the



vCenter Server is not visible to an ESX/ESXi host if a user logs in directly to a host.



A role is assigned to a user or group

All roles are independent of one another



Objects are entities on which actions are performed

 Objects include datacenters, folders, resource pools, clusters, hosts, datastores,

networks and virtual machines

All objects have a Permissions tab

 This tab shows which user or group and role are associated with the selected object



To assign a permission:

 Select a user

 Select a role

 Propagate the permission to child objects (Optional)



You can view all of the objects to which a role was assigned and all of the users or groups who

were granted the roles (Home>Administration>Roles)



You can override permissions set at a higher level by explicitly setting different permissions for

a lower level object

When a user is a member of multiple groups, and these groups have permissions on the same

object in the inventory, the user is assigned the union of privileges assigned to the groups for

that object.



Permissions defined explicitly for the user on an object take precedence over a user’s group

permissions on that same object.



Mechanisms for optimizing virtual machine resource use (3 groups)

 Mechanisms that are managed by the VMkernel

 Mechanisms that are used at the discretion of each virtual machine’s owner

 Mechanisms that are used by the vSphere administrator to set policies for virtual

machines



Resource management is the allocation of resources from providers (hosts, clusters, and

resource pools) to consumers (virtual machines). Resources include CPU, memory, storage and

network



Resource allocation settings

 CPU and memory is controlled by using shares, limits and reservations

 Storage I/O is controlled by using shares and limits

A virtual machine has 3 user defined settings that affect its CPU resource allocation:

 CPU limit defines the maximum amount of CPU, measured in megahertz, that this VM is

allowed

 CPU reservation defines the amount of CPU, measured in megahertz, reserved for the

VM when CPU contention occurs

 Shares specify the relative priority or importance of a VM. If a VM has twice the CPU

shares as another virtual machine, it is entitled to consume twice as much CPU when

these VMs are competing for resources



The Proportional Share mechanism applies to CPU, memory, and storage I/O allocation. It

operates only when VMs are contending for the same resource



Shares guarantee that a VM is given a certain amount of a resource

 You can add shares to a VM while it is running, and it will get more access to that

resource (assuming competition for the resource)

A virtual machine has 4 user defined memory settings that affect its memory resource

allocation:

 Available memory is the amount of memory of given to the VM at the time it was

created

 Memory limit defines the maximum amount of virtual machine memory that can reside

in RAM, not to exceed available memory

 Memory reservation is the amount of RAM reserved for that VM. Unused memory

reservations, like CPU reservations, are not wasted

 Memory shares control how often it wins competition for RAM when RAM is scarce



Storage I/O Control provides quality of service capabilities for storage I/O in the form of I/O

shares and limits that are enforced across all virtual machines accessing a datastore, regardless

of which host they are running on



When you allocate storage I/O resources, you can limit the input/output operations per second

(IOPS) that are allowed for a virtual machine.



Configuring Storage I/O Control is a 2 step process:

 Enable Storage I/O Control for each datastore that you want to control

 Set the number of storage I/O shares and upper limit of IOPS for each VM



By default, all VM share are set to Normal (1000), with unlimited IOPS





A Resource Pool is a logical abstraction for hierarchically managing CPU and memory resources

 It is used on standalone hosts or clusters enabled for VMware Distributed Resource

Scheduler (DRS)



The topmost resource pool is called the root resource pool. Each standalone host and each DRS

cluster has an (invisible) root resource pool that groups the resources of that host or cluster.

 The root resource pool does not appear, because the resources of the host (or cluster)

and the root resource pool are always the same



A vApp is not only a container for VMs but also a resource pool for its virtual machines



Benefits of resource pools:

 Flexible hierarchical organization

 Isolation between pools and sharing within pools

 Access control and delegation

 Separation of resources from hardware

 Management of sets of virtual machines running a multitier service

Resource pool attributes:

 Shares – Low, normal, high and custom

 Reservations, in MHz and MB

 Limits in MHz and MB (unlimited by default)

 Expandable reservation?

o Yes – VMs and subpools can draw from this pool’s parent

o No – VMs and subpools can draw only from this pool, even if its parent has free

resources



You can create a resource pool on an ESX/ESXi standalone host, DRS cluster, or in another

resource pool.



Except for the root resource pool, every resource pool has a parent resource pool. A resource

pool might contain child resource pools or only VMs that are powered on within it



A child resource pool is used to allocate resources from the parent resource pool for the child’s

consumers. Administrative control can also be delegated to individuals or organizations. A child

resource pool cannot exceed the capacity of the parent resource pool. Creating a child pool

reserves resources from the parent pool, whether or not any virtual machines in the child pool

are powered on.



Expandable reservation allows a resource pool that cannot satisfy a reservation to request

through its hierarchy to find unreserved capacity to satisfy the reservation request.



Admission Control is used to ensure that you cannot allocate resources that are not available.

Certain operations must satisfy admission control

 Powering on a VM

 Creating a resource pool with its own reservations

 Increasing a resource pool’s reservation



The resource pool Summary tab displays information that applies to the host machine and its

resources:

 The General pane displays basic information about VMs in the resource pool, as well as

child resource pools

 The CPU pane displays host CPU usage

 The Memory pane displays host memory usage

 The Commands pane allows you to perform actions like creating a VM, creating a

resource pool, and editing a resource pool’s settings

 The Resource Allocation tab allows you to display information about a resource pool’s

CPU, memory and storage resources



You can schedule a task to change the resource settings for a resource pool or virtual machine

You can configure a VM with up to 8 virtual CPUs. The VMkernel includes a CPU scheduler that

dynamically schedules vCPUs on the physical processor of the host system.



Hyperthreading provides more logical CPUs on which vCPUs can be scheduled. It does not

double the power or a core. Hyperthreading is enabled by default. You can enable

hypertheading in the system BIOS.



Logical processors on the same core have adjacent CPU numbers. Logical processors 0 and 1 are

on the first core together; logical processors 2 and 3 are on the same core, and so on.



Every 2-40 milliseconds (depending on the socket-core-thread topology), the VMkernel looks to

migrate vCPUs from one logical processor to another to keep the load balanced. The VMkernel

does its best to schedule virtual machines with multiple vCPUs on 2 different cores rather than

on 2 logical processors on the same core.



For ESX hosts only, the service console always runs on the first logical processor and is never

migrated to another one.



If a logical processor has no work, it is put into a halted state. This action frees its execution

resources.





The VMkernel manages a machine’s entire memory

 Part of this memory is used by the VMkernel

 Some of this memory is used by the service console (ESX only)

 The rest is available for use by VMs (configured memory, plus overhead)



VMS can use more memory than the physical machine has available (Overcommitment)



Memory compression improves virtual machine performance when memory is overcommitted.

 When memory becomes overcommitted, virtual pages are compressed and stored in

memory

 Compressed memory is faster to access than memory swapped to disk

 Enabled by default

 When a host’s memory becomes overcommitted, ESX/ESXi compresses virtual pages

and stores them in memory

 Accessing compressed memory is faster than accessing memory that has been swapped

to disk



The Service Console typically uses 300MB



The VMkernel dynamically scans memory to look for duplicate pages. The VMkernel detects

when different virtual machines have memory pages with identical content and arranges for

those pages to be shared. That is, a single physical page is mapped into each VM’s address

space. If a VM tries to modify a page that is shared, the VMkernel creates a new, private copy

for that VM and then maps that page into the address space of that VM only. The other VMs

continue to share the original copy.



The Balloon Driver refers to the vmmemctl device driver

 Used to perform memory deallocation or reallocation

 Installed on the guest OS when you install VMware Tools

 It demands memory from the guest OS and later to relinquishes it under the control of

the VMkernel

 When a system is not under memory pressure, no VM’s balloon is inflated. But when

memory becomes scarce, the VMkernel chooses a VM and inflates it balloon telling the

balloon driver in the VM to demand memory from the guest OS



VMkernel Swap File

 Each VM has its own

 Created when the VM is powered on and deleted when it’s powered off

 Default location is the same VMware vStorage, VMFS volume as the VM’s boot disk.

 Size is equal to the difference between the memory guaranteed to it, if any, and the

maximum it can use

 Allows the VMkernel to swap out the VM’s machine entirely if memory is scarce

 Used as last resort since performance is slow

If a VM can’t get enough memory through ballooning, the VMkernel forcibly reclaims memory

from other VMs. The VMkernel copies the contents of the pages of these VMs to their

corresponding swap files before giving the pages to the VM that needs memory.



By default, up to 65% of a VM’s memory can be taken away in the ballooning process, subject

to the memory reservation settings.



VMware Tools includes a library of functions called the Perfmon DLL.

 Perfmon allows you to access key host statistics in a guest VM.

 The Perfmon performance objects (VM Processor and VM Memory) allow you to view

actual CPU and memory use alongside observed CPU and memory use of the guest OS.

 Click Overview to display charts for the most common data counter for CPU, disk,

memory, and network metrics.

 Click Advanced to view data counters not supported in the overview performance

charts, to export chart data, and to print charts.

 The key to interpreting performance data is to observe the range of data from the guest

operating system, the virtual machine, and the host’s perspective



Multiple Virtual Machines are constrained by CPU if:

 There is high CPU use in the guest OS

 There are relatively high CPU ready values for the VMs

Ready Time refers to the interval when a VM is ready to execute instructions but cannot,

because it cannot get scheduled onto a CPU



When a VM experiences ballooning activity, some of the guest operating system’s physical

memory is being reclaimed from the VM by the balloon driver. If a VM experiences high

ballooning values, this might not be a problem if the VM continues to have the memory that it

needs. But if a VM experiences high ballooning activity over time and its guest operating system

stars to page, the VM might be constrained for memory.



Monitoring for increases in active memory on the host

 Host active memory refers to active physical memory used by virtual machines and the

VMkernel

 If amount of physical memory is high, this could lead to VMs that are memory

constrained



Disk-intensive applications can saturate the storage or the path. If you suspect that a VM is

constrained by disk access:

 Measure the throughput and latency between the virtual machine and storage

 Use the advanced performance charts to monitor:

o Read rate and write rate

o Read latency and write latency

If you select a host object, you can view throughput and latency for a datastore, a storage

adapter, or a storage path. The storage adapter charts are only available for Fibre Channel

storage. The storage path charts are available for Fibre Channel and iSCSI storage, not NFS.



To monitor throughput, view the Read rate and Write rate counters. To monitor latency, view

the Read latency and Write latency counters



Find disk problems by monitoring disk latency and data counters

 Kernel Command Latency

o Measures the average amount of time, in milliseconds, that the VMkernel

spends processing each SCSI command

o For best performance, the value should be 0-1 milliseconds

 Physical Device Command Latency

o Measures the average amount of time, in milliseconds, for the physical device to

complete a SCSI command

o Depending on your hardware, a number greater than 15 milliseconds indicates

that the storage array might be slow or overworked

If you suspect a VM is constrained by the network:

 Confirm that VMware Tools is installed and that the enhanced network drivers are

available

 Measure the effective bandwidth between the VM and its peer system

 Check for dropped receive packets and dropped transmit packets

 To determine whether packets are being dropped, use the advanced performance

charges to examine the droppedTx and droppedRx network counter values of a VM



Alarms

 The predefined alarms are configurable

 To create an alarm, right click an object in the inventory and select Alarm>Add Alarm

 The Alarm Settings dialog box has 4 tabs: General, Triggers, Reporting and Actions

 In the General tab, you name the alarm, give it a description, enable or disable the

alarm, give it an alarm type and select what to monitor

o Monitor for specific conditions or state

o Create conditions based alarms for VMs, hosts and datastores

o Monitor for specific events occurring on this object

o Create event based alarms for VMs, hosts, clusters, datacenters, datastores,

networks, distributed virtual switches, and distributed virtual port groups

 Triggers tab

 Alarms have 2 types of Triggers: condition or state triggers and event triggers

 Condition or State Triggers

o Monitor the current condition or state of virtual machines, hosts and

datastores

o Conditions or states include power states, connection states, and

performance metrics such as CPU and disk usage

 Event Triggers

o Monitor events that occur in response to operations occurring with a

managed object in the inventory or the vCenter Server itself



If you add multiple triggers, you can choose to trigger the alarm if any one of the conditions is

satisfied or if all the conditions are satisfied



Reporting tab

 Used to define a tolerance range and trigger frequency for condition or state triggers

(not available for event triggers)

 Reporting further restricts when the condition or state trigger occurs. You can specify a

range or a frequency

o If using a range, the triggered alarm is repeated when the condition exceeds the

range

o If using a frequency, the triggered alarm is repeated every so often (in minutes)

Actions tab

 Every alarm can send a notification email, send a notification trap, or run a command

 You can set alarms to trigger when the state changes:

o From a green circle to a yellow triangle

o From a yellow triangle to a red diamond

o From a red diamond to a yellow triangle

o From a yellow triangle to a green circle

o For every action, you can specify an option for each color transition:

 Empty indicates no interest in the transaction

 Once tells vCenter to do the action only one time

 Repeat tells vCenter to repeat the action until another color change

occurs. The default if 5 minutes and the maximum is 2 days

 Virtual machine and host alarms have more actions such as:

o Power on a VM

o Power off a VM

o Suspend a VM

o Reboot host

o Shut down host

 You can configure up to 4 receivers of SNMP traps.

o They must be configured in numerical order

o Each SNMP trap requires a corresponding host name, port and community



Data Protection

 After you configure, change the configuration, or upgrade an ESXi host, backup your

configuration

 The serial number is backed up and restored when you restore your configuration

 The serial number is not preserved when you run the recovery CD (ESXi Embedded) or

perform the repair operation (ESXi Installable)

 Use the vicfg-cfgbackup command to do the backup from the vCLI from Windows or

Linux

 Use the recovery CD or the repair option if the host does not boot up because the file

partitions or MBR on the installation disk might be corrupted



Use the following methods when backing up the Service Console:

 File backed backup

o Treat the service console as a physical machine with a deployed backup agent

 Image based backup

o Use third party software to create a backup imaged that you can restore quickly

Virtual Machine Backups

VMware Consolidated Backup (VCB)

 Used with supported third party software to do backups of virtual machine disks.

 Centralizes backup on the VCB proxy server

 VCB is the previous generation backup technology (vStorage APIs for Data Protection

and Data Recovery is most current)



vStorage APIs for Data Protection

 Allows backup and recovery of entire VM images across SAN storage or LANs

 Is an easy Smart Plug-in (SPI) that is directly integrated with backup tools from third

party vendors

 Enables you to remove load from the host and consolidates backup load onto a central

backup server

 Protects VMs that use any type of storage supported by ESX/ESXi (Fibre Channel, iSCSI,

NAS or local storage

 Part of a larger set of APIs know as vStorage APIs and consists of the following sets:

o Site Recovery Manager

o Array Integration

o Multipathing

o Data Protection



VMware Data Recovery (VDR)

 Agentless disk based backup and recovery appliance

 Based on the vStorage APIs for Data Protection

 VMware vCenter plugin

 Supports up to 10 appliances per vCenter Server instance

 Supports up to 100 VMs per appliance

 Intended for small to mediums sized environments

 Different backup appliances do not share information about backup jobs

 All backed up VMs are stored in a deduplicated store. The deduplicated store can be

stored in a VMFS, RDM, NFS, or Common Internet File System (CIFS) shares

 Requires an absolute minimum of 10GB of free space

 Need Essentials Plus, Advanced, Enterprise or Enterprise Plus licensing

 VDR components communicate with each other over TCP

o Connects to vCenter Server Web services on ports 80 and 443

o Client plugin and File Level Restore connect to the backup appliance over port

22024

o The backup appliance connects to an ESX/ESXi host over port 902

VDR Deduplication

 RDMs are recommended for deduplication stores

 To maximize deduplication rates, ensure that similar VMs are backed up to the same

destination

 The deduplication store completes the following processes:

o Integrity check

 Verifies and maintains data integrity

 VDR completes an incremental integrity check every 24 hours

 VDR performs an integrity check of all restore points once a week

o Recatalog

 Ensures that the catalog of restore points is synchronized with the

contents of the deduplication store

o Reclaim

 Reclaims space on the deduplication store

 Runs daily or when a backup job requires more space than is available on

the deduplication store

 Supports deduplication stores that are up to 1TB in size on VMDKs and RDMs and

500GB on CIFS shares

 Each backup appliance is limited to using 2 deduplication stores





VDR installation

 Install the client plugin

 Install the backup appliance

 Add a hard disk to the backup appliance



Configuration

 Power on the appliance and change the root password

 Configure network settings, and reboot if necessary

 Connect the appliance to the vCenter Server

 Configure the backup destination on the appliance

 Default username is root and password is vmw@re



VDR backup jobs

 A maximum of 8 jobs can run simultaneously

 Backup jobs can backup 100 VMs total

 By default, backup jobs run at night Monday through Friday and at any time Saturday

and Sunday



Rehearsal Restore

 Used to test how a VM would be restored through restore operations

 Does not replace the current VM

File Level Restore

 Can be installed in Windows or Linux VMs

 Requires administrator privileges

 Not supported on physical machines



vCenter Linked Mode

 Log in simultaneously to all vCenter Server systems

 View and search the inventories of all vCenter Server systems

 You cannot migrate hosts or VMs between vCenter Server systems in Linked Mode

 Can have up to 10 linked vCenter Server systems

 Can have up to 3,000 hosts across the linked vCenter Server systems

 Supports 30,000 powered on VMs and 50,000 registered VMs across linked vCenter

Server systems

 Uses Microsoft’s Active Directory Application Mode (ADAM) to store and synchronize

data across multiple vCenter Server instances

 Using peer to peer networking, the vCenter Server instances in Linked Mode replicated

shared global data to the LDAP directory

 The vSphere Client can connect to other vCenter Server instances by using the

connection information retrieved from ADAM.

 The Apace Tomcat Web service running on vCenter Server enables the search capability

across multiple vCenter Server instances

 For inventory searches, vCenter Linked Mode relies on a Java based Web application

called the query service, which runs in Tomcat Web services

 The search service queries Active Directory for information about user permissions. So

you must be logged in to a domain account to search all vCenter Server systems in

vCenter Linked Mode



When adding a vCenter Server instance to a Linked Mode group, the user running the installer

must be a local administrator on the machine where vCenter Server is being installed and on

the target machine of the Linked Mode group. Generally, the installer must be run by a domain

user who is an administrator of both systems





The following requirements apply to each vCenter Server system that is a member of a Linked

Mode group:

 DNS must be operational for Linked Mode replication to work

 The vCenter Server instances in a Linked Mode group can be in different domains if the

domains have a 2 way trust relationship. Each domain must trust the other domains on

which vCenter Server instances are installed

 All vCenter Server instances must have network time synchronization. The vCenter

Server installer validates that the machine clocks are no more than 5 minutes apart



Install the first vCenter Server instance as a standalone instance

The vCenter Server instances in a Linked Mode group do not need to have the same domain

user login



During vCenter Server installation, if you enter an IP address for the remote instance of vCenter

Server, the installer converts it into a fully qualified domain name



To join a vCenter Server system to a Linked Mode group click on

Start>Programs>VMware>vCenter Server Linked Mode Configuration



vCenter Server Status shows information such as:

 A list of all vCenter Server systems and their services

 A list of all vCenter Server plugins

 The status of all listed items

 The data and time of the last change in status

 Messages associated with the change in status



Roles are replicated when a vCenter Server system is joined to a Linked Mode group

 If role names differ on vCenter Server systems, they are combined into a single common

list and each server will have all the user roles

 If role names are identical, they are combined into a single role if they have the same

privileges

 If role names are identical, and the roles contain different privileges, these roles must be

reconciled



Use the vCenter Server Linked Mode Configuration wizard to isolate (remove) a vCenter Server

instance from a Linked Mode group

 Start>Programs>VMware>vCenter Server Linked Mode Configuration

 Click Modify linked mode configuration and click Next

 Click Isolate this vCenter Server instance from linked mode group





Host Profiles

 Basic workflow to implement host profiles:

o Setup and configure a host for a reference

o Use the Create Profile wizard to create a profile from the designated reference

host

o Attach the host or cluster to the profile

o Check the host’s compliance against a profile to ensure that the host continues

to be correctly configured

o Check new hosts for compliance against the host profile. You can easily apply the

host profile of the reference host to other hosts or clusters of hosts that are not

in compliance

You can also import and export a profile file to a host profile that is in the VMware profile

format (.vpf)



After the host profile is created and associated with a set of hosts or clusters, you can check the

compliance status from various places in the vSphere Client

 Host Profiles main view – Displays compliance status of hosts and clusters, listed by

profile

 Host Summary tab – Displays compliance status of the selected host

 Cluster Profile Compliance tab – Displays compliance status of the selected cluster and

all the hosts within the selected cluster



Whenever a new host is added into a cluster, it is checked for compliance against the host

profile that has been applied



You can also schedule tasks in vSphere to help automate compliancy checking



To apply a host profile:

 Go to Home>Management>Host Profiles

 Select the host profile in the inventory and click the Hosts and Clusters tab

 Right click the host and select Apply





vNetwork Distributed Switch

 vCenter Server owns the configuration of the distributed virtual switch. The

configuration will be consistent across all the hosts that use it

 A distributed virtual switch can support up to 350 hosts

 A distributed virtual switch can benefit from the performance of 10GbE physical NICs

 Provides support for private VLANs

 Distributed ports migrate with their clients



Private VLANs allow you to use VLAN IDs within a private network without having to worry

about duplicating VLAN IDs across a wider network

Some configuration is specific to the host. A host’s uplink ports are allocated to the distributed

virtual switch and are managed in the host’s network configuration. Similarly, the VMkernel and

service console ports are managed in the host’s network configuration as well.



You connect a virtual machine to a distributed virtual switch by connecting the VMs NIC to a

port group on the distributed virtual switch



A distributed virtual switch is a managed entity configured in vCenter Server

Each distributed virtual switch includes distributed ports. A distributed port represents a port to

which you can connect any networking entity, such as a VM, a VMkernel interface, or a service

console interface (ESX only)



Ports can exist without port groups



An uplink is an abstraction to associate the vmnics from multiple hosts to a single distributed

virtual switch



VMs on different hosts can communicate with each other only if both VMs have uplinks on the

same broadcast domain



The distributed virtual switch architecture consists of 2 planes: the control plane and the I/O

plane

 The control plane resides in vCenter Server and is responsible for configuring distributed

virtual switches, distributed port groups, distributed ports, uplinks, NIC teaming etc.

 The I/O plane is implemented as a hidden virtual switch in the VMkernel of each

ESX/ESXi host. The I/O plane manages the I/O hardware on the host and is responsible

for forwarding packets



Editing general switch properties

 The settings dialog box has 3 tabs: Properties, Network Adapters and Private VLAN

 The Network Adapters tab is a read only form that allows you to verify which physical

adapters are connected to the distributed virtual switch

 The Private VLAN tab allows you to setup private VLANs for the distributed virtual switch

 The Network Adapters and Private VLAN tabs are only available for distributed virtual

switches, not for distributed ports or distributed port groups

 Settings on the Properties tab are grouped into the categories General and Advanced.

General properties for the distributed virtual switch allow you to edit the information

specified when creating the distributed virtual switch

 Advanced properties on the distributed virtual switch allow you to define the maximum

transmission unit (MTU), the Cisco Discovery Protocol (CDP) status, and the

administrator contact details



MTU (Maximum Transmission Unit) determines the maximum size of frames in this distributed

virtual switch. The distributed virtual switch drops frames bigger than the specified size. If your

environment supports jumbo frames, use this option to enable or disable jumbo frames on the

distributed virtual switch. To enable jumbo frames on the distributed virtual switch, set the

Maximum MTU to 9000. To use jumbo frames, the network must support it end to end.

ESX/ESXi supports jumbo frames in the gest OS and on VMkernel ports.

CDP has 3 operation modes:

 Listen mode (default) – The ESX/ESXi host detects and displays information about the

associated Cisco switch port. But information about the virtual switch is not available to

the Cisco switch admin

 Advertise mode - The ESX/ESXi host makes information about the virtual switch

available to the Cisco switch admin

 Both mode – does both





Network resource pools determine the priority that different network traffic types are given on

a distributed virtual switch. By default, Network I/O Control is disabled. When Network I/O

Control is enabled, distributed virtual switch traffic is divided into the following network

resource pools:

 FT traffic

 iSCSI traffic

 vMotion traffic

 Management traffic

 NFS traffic

 VM traffic



Network shares and limits apply to a host’s outbound network I/O traffic only



To enable Network I/O Control

 Go to Home>Inventory>Networking

 Select the distributed virtual switch in the inventory and click the Resource Allocation

tab

 Click the Properties link and select Enable network I/O control on this vDS

 To modify the shares or limits of a particular network resource pool, right click the

resource pool and select Edit Settings

vMotion Migration

The state information includes the current memory content and all the information that defines

and identifies the virtual machine



vMotion Migration consists of the following steps:

1. The VMs memory state is copied over the vMotion network from the source host to the

target host

2. After most of the VMs memory is copied from the source host to the target host, the

VM is quiesced: no additional activity will occur on the VM

3. Immediately after the VM is quiesced on the source host, the VM is initialized and starts

running on the target host

4. Users access the VM on the target host instead of the source host

5. The VM is deleted from the source host



A Virtual Machine must meet the following requirements for vMotion:

 A VM must not have a connection to an internal vSwitch (vSwitch with zero uplink

adapters)

 A VM must not have a connection to a virtual device (Such as a CD-ROM or floppy) with

a local image mounted

 A VM must not have CPU affinity configured

 If the VM’s swap file is not accessible to the destination host, vMotion must be able to

create a swap file accessible to the destination host before migration can begin

 If a VM uses an RDM, the RDM must be accessible by the destination host

Host requirements for vMotion Migration

Source and destination hosts must have:

 Visibility to all storage (Fibre Channel, iSCSI, or NAS) used by the VM

o 128 concurrent vMotion migrations per vStorage VMFS datastore

 At least a Gigabit Ethernet network

o 4 concurrent vMotion migrations on a 1Gbps network

o 8 concurrent vMotion migrations on a 10Gbps network

 Access to the same physical networks

 Compatible CPUs:

o CPU feature sets of both the source and destination host must be compatible

o Some features can be hidden by using Enhanced vMotion Compatibility (EV) or

compatibility masks



If you are using standard virtual switches for networking, ensure that the network labels used

for VM port groups are consistent across hosts









AMD No eXecute (NX) and Intel eXecute Disable (XD) technologies serve the same security

purpose: to mark memory pages as data-only to prevent malicious software exploits and buffer

overflow attacks



If NX/XD technology is exposed on the source host, then it must be exposed on the destination

host. NX/XD technology is exposed by default for all guest operating systems that can use it

(trading off compatibility for security by default)

Hiding the NX/XD flag will increase vMotion compatibility between hosts, at the cost of

disabling certain CPU security features for some guest operating systems and applications



To hide the NX/XD flag from the guest OS:

 Right click the powered off VM and click Edit Settings

 Click the Options tab

 Select the CPUID Mask setting to hide or expose the flag



If the specifications of a server or its CPU features are unknown, you can use the VMware CPU

Identification Utility to boot a server and determine whether its CPUs contain features like

SSE3, SSSE3, and NX/XD



You can verify vMotion requirements by viewing the Maps tab of the VM being migrated



DRS Clusters

 A cluster is a collection of ESX/ESXi hosts and associated VMs with shared resources and

a shared management interface



The following cluster-level resource management capabilities are available:

 Initial placement – When you first power on a VM in the cluster, DRS either places the

VM on an appropriate host or makes a recommendation

 Load balancing – DRS continuously monitors the distribution and usage of CPU and

memory resources for all hosts and VMs in the cluster

 Power management – When VMware DPM is enabled, DRS compares cluster-level and

host-level capacity to the demands of the clusters VMs, including recent historical

demand. It places (or recommends placing) hosts in standby power mode if sufficient

excess capacity is found or powering on hosts if capacity is needed.



A system that is added to a DRS cluster must meet certain prerequisites to use cluster features:

 DRS works best if the VMs meet vMotion requirements

 To use DRS for load balancing, the hosts in your cluster must be part of a vMotion

network

 Configure all managed hosts to use shared storage (VMFS or NFS datastores)

 Place the disks of all VMs on shared storage that is accessible by source and destination

hosts

 Ensure that the shared storage is sufficiently large to store all virtual disks for you VMs

DRS automation levels

Manual – When you power on a VM, DRS displays a list of recommended hosts on which to

place the VM. When the cluster becomes unbalanced, DRS displays recommendations for VM

migration

Partially automated – When you power on a VM, DRS places it on the best suited host. When

the cluster becomes unbalanced, DRS displays recommendations for VM migration

Fully automated – When you power on a VM, DRS places it on the best suited host. When the

cluster becomes unbalanced, DRS migrates VMs from overutilized hosts to underutilized hosts

to ensure a balanced use of cluster resources.



The migration threshold determines how quickly DRS migrates VMs:

 Level 1 (Conservative) – Applies only priority 1 recommendations. vCenter Server will

apply only recommendations that must be taken to satisfy cluster constraints like

affinity rules and host maintenance

 Level 2 – Apply priority 1 and priority 2 recommendations. vCenter Server will apply

recommendations that promise a significant improvement to the cluster’s load balance

 Level 3 (default) - Apply priority 1 and priority 2 and priority 3 recommendations.

vCenter Server will apply recommendations that promise at least good improvement to

the cluster’s load balance

 Level 4 - Apply priority 1 and priority 2, priority 3 and priority 4 recommendations.

vCenter Server will apply recommendations that promise even a moderate

improvement to the cluster’s load balance

 Level 5 (Aggressive) – Apply all recommendations. vCenter Server will apply

recommendations that promise even a slight improvement to the cluster’s load balance



Enhanced vMotion Compatibility (EVC)

 Use EVC to help ensure vMotion compatibility for the hosts in a cluster

 EVC ensures that all hosts in a cluster present the same CPU feature set to VMs, even if

the actual CPUs on the hosts differ

 Hosts that cannot be configured to use the CPU baseline for an EVC cluster are not

permitted to join the cluster



EVC requirements for all hosts on the cluster

 Use CPUs from a single vendor (either Intel or AMD)

o Use Intel CPUs with Core 2 micro architecture or newer

o Use AMD first generation Opteron CPUs and newer

 Run ESX 3.5 Update 2 or later

 Be connected to vCenter Server

 Be enabled for hardware virtualization (AMD-V or Intel VT)

 Be enabled for execution-disable technology (AMD No eXecute (NX) or Intel eXecute

Disable (XD))

 Be configured for vMotion migration

 Applications in VMs must be well-behaved

VMware recommends creating an empty EVC cluster as the simplest way of creating an EVC

cluster with minimal disruption to your existing infrastructure



By default, swap files for VMs are on a VMFS datastore in the folder containing the other VM

files



If the swap file location specified on the destination host differs from the swap file location

specified on the source host, the swap file is copied to the new location. Copying the swap file

can result in slower migrations with vMotion



After a DRS cluster is created, you can edit its properties to create rules that specify affinity.

There are two types of rules:

 Affinity rules – DRS should try to keep certain VMs together on the same host

 Anti-affinity rules – DRS should try to make sure that certain VMs are not together



DRS Group

 A group of VMs

 A group of hosts

 A VM can belong to multiple VM DRS groups

 A host can belong to multiple DRS groups



A Virtual Machines to Host affinity rule specifies whether the members of a selected virtual

machine DRS group can run on the members of a specific host DRS group



A Virtual Machines to Host affinity rule includes 3 components

 One virtual machine DRS group

 One host DRS group

 A designation of whether the rule is a requirement (“must”) or a preference (“should”)

and whether it is affinity (“run on”) or anti-affinity (“not run on”)



The VMs and hosts that are included in a rule must all reside in the same cluster



A preferential rule is one that is softly enforced. Preferential rules can be violated to allow the

proper functioning of DRS, VMware High Availability, and VMware DPM



A Virtual Machines to Hosts affinity rule that is required, instead of preferential, can be used

when the software that you are running in your VMs has licensing restrictions

 The rule does not monitor the software running in the VMs nor does it know what non-

VMware licenses are in place on which ESX/ESXi hosts

You can customize the automation level for individual virtual machines in a DRS cluster to

override the automation level set on the entire cluster

 As a best practice, enable automation

 Partially Automated or Fully Automated

 Use Manual on VMs where you want more control





When adding a host with resource pools to a DRS cluster, you must decide on resource pool

placement.

 By default, the resource pool hierarchy is discarded and the host is added at the same

level as the VMs.

 You can choose to graft the host’s resource pools onto the cluster’s resource pool

hierarchy

 You can choose a name for the resource pool created to represent the host’s resources.

 By default, the resource pool created to represent the host’s resources is named

Grafted from



The VMware DRS pane in the cluster’s Summary tab appears only when DRS is enabled. This

section provides DRS information like:

 The automation levels selected

 The number of DRS recommendations and faults

 The migration threshold

 It also provides 2 standard deviation values

o Target host load standard deviation – A value derived from the migration

threshold setting that indicated the value under which load imbalance is to be

kept

o Current host load standard deviation – A value indicating the current load

imbalance in the cluster. This value should be less than the target host load

standard deviation, unless unapplied DRS recommendations or constraints

preclude attaining that level



Click the View Resource Distribution Chart link to open the Resource Distribution chart. This

chart provides CPU and memory use information, displayed per VM

 Green means that 100% of the VMs entitled resources has been delivered to it



There are 3 views from the DRS tab – Recommendations, Faults and History

 Recommendations

o Allows you to view and edit cluster properties

o Only manual recommendations awaiting user confirmation appear on this list

o To refresh the recommendations, click Run DRS

o To apply all recommendations, click Apply Recommendations

o To apply a subset of the recommendations, select the Override DRS

recommendations check box

Monitoring cluster status

An icon on the cluster object shows whether a cluster is valid, overcommitted (yellow triangle),

or invalid (red diamond)

 A cluster can become overcommitted if a host fails

 A cluster can become invalid if you use the vSphere Client to directly access the ESX/ESXi

host to power on or make changes to the VM

 A cluster can become invalid if the user reduces the reservation on a parent resource

pool while a VM is in the process of failing over



Maintenance Mode and Standby Mode

 You put a host in maintenance mode when you need to service it

 When a host is placed in standby mode, it is powered off

 Normally, hosts are placed in standby mode by VMware DPM to optimize power usage



Removing a host from the DRS Cluster

 To remove a host from a cluster, right click the host in the inventory and select Enter

Maintenance Mode. After the host is in maintenance mode, drag it to a different

inventory location

 When you remove a host from a cluster, the host retains only the root resource pool

 If you remove a host from a cluster, the resources available for the cluster decrease





VMware DPM

 DPM continuously monitors resource requirements and power consumption across a

DRS cluster.

 When the cluster needs fewer resources, it consolidates workloads and powers down

unused ESX/ESXi hosts to reduce power consumption

 DPM uses one of three power management protocols to bring a host out of standby

mode:

o Intelligent Platform Management Interface (IPMI)

o Hewlett Packard Integrated Lights Out (iLO)

o Wake on LAN (WOL)

o If a host does not support any of these protocols, it cannot be put into standby

mode by VMware DPM

 If a host supports multiple protocols, they are used in the following order:

o IPMI

o iLO

o WOL



Hosts powered off by DPM are marked by vCenter Server as being in standby mode

DPM operates by awakening ESX/ESXi hosts from a powered off state through WOL packets

 These packets are sent over the vMotion networking interface by another host in the

cluster, so DPM keeps at least one host powered on at all times.



DPM powers off the host when the cluster load is low

 DPM considers a 40 minute load history

 All VMs on the selected host are migrated to other hosts



DPM powers on a host when the cluster load is high

 It considers a 5 minute load history

 The WOL packet is sent to the selected host, which boots up

 DRS load balancing initiates, and some virtual machines are migrated to this host



When HA admission control is disabled, failover resource constraints are not passed on to DRS

and DPM

 The constraints are not enforced

 DRS does not evacuate VMs from hosts

 It places the hosts in maintenance or standby mode, regardless of the effect that this

might have on failover requirements

 DRS might undo (or recommend undoing) your change the next time that it runs

 To force a host to remain off, place it in maintenance mode and power it off



DPM is a cluster power management feature. Enhanced Intel SpeedStep and AM PowerNow!

are CPU power management technologies



Enabling DPM

 Configure the power management automation level, threshold, and host-level overrides

 The power management automation levels are different from the DRS automation levels

o Off - Disables the feature

o Manual – Sets DPM to make recommendations for host power operation and

related VM migration, but recommendations are not automatically executed

o Automatic – Sets DPM to execute host power operations if related VM

migrations can all be executed automatically



A priority 1 recommendation is mandatory, while a priority 5 recommendation brings only

slight improvement



When you disable DPM, hosts are taken out of standby mode



You can verify that DPM is functioning properly by viewing each host’s information in the Last

Time Exited Standby column on the Host Options page on the Hosts tab for each cluster

High Availability and Fault Tolerance

 HA refers to a system or component that is continuously operational for a desirably long

length of time

 FT describes a computer system or component that is designed so that, if a component

fails, a backup component or procedure can immediately take its place with no loss of

service

 HA and FT exist within a single physical datacenter

 HA and FT use shared storage for holding the data of the machines

 Storage path availability is accomplished by using the failover policies available with

multipathing

 Network availability is accomplished by using the failover feature in NIC teaming

 Site Recovery Manager (SRM) allows you to quickly restore your IT infrastructure

o SRM is disaster recovery workflow product that automates setup, failover, and

testing of disaster recovery plans

o SRM requires that vCenter Server be installed at the protected site and the

recovery site

o SRM requires preconfigured array based replication between the protected site

and the recovery site









Use HA to provide high availability to all the VMs in your cluster that require minimal downtime



Use FT or MSCS for applications that must be available at all times (zero downtime), especially

those that have long lasting client connections to maintain during hardware failure

vCenter Heartbeat provides deep and comprehensive levels of protection against unplanned

and, in some cases, planned vCenter Server downtime. HA is a good alternative for vCenter

Server running on a VM



VMware HA

 Provides automatic restart of VMs in case of physical host failures

 Is configured, managed, and monitored through vCenter Server

A cluster enabled for HA and DRS can have:

 Up to 32 hosts per cluster

 Up to 320 VMs per host (regardless of the number of hosts/clusters)

 Up to 3,000 VMs per cluster



HA is integrated with DRS



FT checks that individual VMs are functioning and responds to failures without interruption in

service

 FT creates a hidden duplicate copy of each running VM



Reasons why HA might not fail over VMs:

 HA admission control is disabled and DPM is enabled

 Required Virtual Machine to Hosts affinity rule prevents HA from failing over

 Sufficient aggregated resources exist, but they are fragmented across hosts



Detecting a Host Failure

 HA agent monitors the heartbeats between the primary and the secondary hosts to

detect host failure

 A heartbeat is sent every second (by default) over the heartbeat network

o On ESXi hosts, the management network is used

o On ESX hosts, the service console network is used



If a 15 second period elapses without the receipt of heartbeats from a host and the host cannot

be pinged, it is declared as failed



In a host failure, HA does not fail over VMs to a host that is in maintenance mode



If a host in the cluster loses connection to the heartbeat network but the host continues to run,

the host is isolated from the cluster

 HA waits 12 seconds before deciding that a host is isolated

When the isolated host’s network connection is not restored for 15 seconds or longer, the

other hosts in the cluster treat it as failed and try to fail over its VMs

 When an isolated host retains access to the shared storage, it also retains the disk lock

on virtual machine files

 vStorage VMFS disk locking prevents simultaneous write operations to the VM disk files

and tries to fail over the isolated host’s VMs



Architecture of a HA Cluster

 Each host in the cluster must have access to the same storage resources

 Distributed locking prevents simultaneous access to VMs, thus protecting data integrity

 The first 5 hosts added to the cluster are designated as primary hosts. Subsequent hosts

are designated as secondary hosts

 The primary hosts maintain and replicate all cluster state and are used to begin failover

actions

 A host that joins the cluster must communicate with a primary host to complete its

configuration (Except for the first host)

 At least one primary host must be functional for HA to operate correctly

 If all primary hosts are unavailable (not responding), no hosts can be successfully

configured for HA



HA provides the option to disable Host Monitoring to avoid affecting maintenance activities

 Host Monitoring is required for the best performance of FT



You can enable or disable Admission Control by selecting from the following options

 Enable: Do not power on VMs that violate availability constraints

 Disable: Power on VMs that violate availability constraints



Admission Control Policy Choices

 Host failures cluster tolerates

o HA reserves a certain amount of resources across a set of hosts

 Percentage of cluster resources reserved as failover spare capacity

o HA reserves a certain percentage of aggregate resources in the cluster to

accommodate failures

 Specify a failover host

o HA reserves a specific host to accommodate failures

Configuring Virtual Machine HA Options

The virtual machine restart policy determines the relative order in which virtual machines are

restarted after a host failure – Disabled, Low, Medium, High or Use cluster

 Disabled

o HA is disabled for VMs. VMs are not restarted on other ESX/ESXi hosts if a host

fails

o VM started on the same host

o Does not affect VM monitoring

 The host isolation response settings are Leave powered on, Power off, Shut down, and

Use cluster setting

 By default, VM monitoring is set to Disabled

o VM monitoring restarts individual VMs if their heartbeats are not received within

a set of time



HA Advanced Parameters

 Set specific attributes that affect how HA behaves

 das.vmMemoryMinMB defines the default memory resource value assigned to a VM if

its memory reservation is not specified or 0. If no value is specified, the default is 0MB

 das.vmCpuMinHz defines the default CPU resource value assigned to a VM if its CPU

reservation is not specified or 0. If no value is specified, the default is 256MHz

 das.slotMemInMB defines the maximum bound on the memory slot size

 das.slotCPUInMHZ defines the maximum bound on the CPU slot size



On ESXi hosts in the cluster, HA communications by default travel over VMkernel networks,

except those marked for use with vMotion, if necessary

 The VMkernel networks must be marked for management traffic



On ESX hosts in the cluster, HA communications travel over all networks that are designated as

service console traffic

 All its service console networks are used as heartbeat networks



One way to implement redundant heartbeat networks is to use NIC teaming

 To configure a NIC team, configure the virtual NICs in a virtual switch for active or

standby configuration and no failback



Another way to create redundancy for the heartbeat networks is to configure more

management ports on separate virtual switches



An isolation network is an IP address that is pinged to determine whether an ESX/ESXi host is

isolated from the network

 Hosts in the HA cluster test themselves for isolation by pinging the isolation address

o By default, ESXi hosts ping the VMkernel gateway IP address

o By default, ESX hosts ping the service console default gateway IP address

Advanced attributes to configure isolation addresses

 das.isolationaddressX, where X = 1 to 10

o The address to ping to determine whether a host is isolated from the network,

that is, when heartbeats are not received from any other host in the cluster

 das.failuredetectionstime

o The default detection time for Host Monitoring

o The default is 15,000 milliseconds (15 seconds)

 das.usedefaultisolationaddress (true or false)

o Specifies whether the default isolation address is used



Before changing the networking configuration on the ESX/ESXi hosts:

 Deselect Enable Host Monitoring

 Place the host in maintenance mode

 These steps prevent unwanted attempts to fail over virtual machines







Fault Tolerance

 You can use FT with DRS when Enhanced vMotion Compatibility (EVC) is enabled

 When a cluster has EVC enabled, DRS:

o Makes the initial placement recommendations for fault tolerant VMs

o Moves them during cluster load rebalancing

o Allows you to assign a DRS automation level to the primary VM



FT can be enabled on a VM in a cluster enabled for HA

 FT creates a duplicated, secondary copy of the VM on a different host

 The VMware record/replay technology is used to record all executions on the primary

VM and replay them on the secondary instance



VMware vLockstep technology ensures that the 2 copies stay synchronized and allows the

workload to run on 2 different ESX/ESXi hosts simultaneously

 The VMs have one IP address and one MAC address

 If either the primary or secondary VMs fails, FT creates a new copy of the VM on

another host in the cluster

 If the failed VM is the primary, the secondary takes over and a new secondary is

established

 If the secondary fails, another secondary takes over and a new secondary is established

FT Requirements

 Host certificate checking must be enabled for all hosts that will be used for FT. For

vSphere 4 ESX/ESXi installations, host certificate checking is enabled by default

 VM files must be stored on shared storage. Acceptable shared storages solutions include

Fibre Channel, iSCSI (hardware and software) and NFS/NAS

 VMs must be stored in virtual raw device mapping (RDM) or VMDK files that are thick

provisioned and enabled to support clustering features like FT

o If stored in a VMDK file that is thin provisioned you will get a message that it

must be converted

 Multiple gigabit NICs are required. The minimum is at least 2 VMkernel gigabit NICs with

1 dedicated to FT logging and the other dedicated to vMotion

 Uniprocessor VMs are supported on uniprocessor and symmetric multiprocessor

systems. SMP VMs are not supported

 VMs must be running a supported guest OS

 FT requires that Hardware Virtualization be turned on in the BIOS



How FT Works

 The primary and secondary VMs access the same virtual disks on a shared SAN

 The primary VM sends both reads and writes to the virtual disks

 The secondary VM sends only reads to the disks

 All writes by the secondary VM are marked as completed, but the writes are not issued

 To detect VMkernel and host failures, FT uses network heartbeats over the IP addresses

used for logging

 The primary and secondary VMs send ping packets to the logging IP address

 If one side does not receive pings within about one second, then that side initiates a

failover

 To detect VM failures, the VMkernel monitors the frequency of log updates from the

configuration file and virtual machine monitor



FT Guidelines

 Ensure enough ESX/ESXi hosts for fault tolerant virtual machines

o No more than 4 fault tolerant VMs (primaries or secondaries) on any single host

 Store ISOs on shared storage for continuous access

 Disable BIOS based power management

o Prevents the secondary VM from having insufficient CPU resources



To enable FT

 Create a network interface for FT logging

 Enable FT on the virtual machine

o Right click VM and choose Fault Tolerance>Turn on Fault Tolerance

You can view information about the primary and secondary VMs in the Fault Tolerance pane in

the VMs Summary tab

 Fault Tolerance Status

o Indicates whether FT is enabled or disabled on the VM

o Possible values are Protected and Not Protected (VM is not running)

 Secondary Location

o Displays the host on which the secondary VM is hosted

 Total Secondary CPU and Total Secondary Memory

o Indicate all secondary CPU and memory usage

 Secondary VM Lag Time

o Indicates the latency between the primary and the secondary virtual machines

 Log Bandwidth

o Indicates the amount of network being used for sending the FT log information

from the primary VMs host to the secondary VM’s host





Patch Management

 Can patch ESX/ESXi hosts, VMs and virtual appliances

 Automated patch downloading

 Contacts the following sources

o For ESX/ESXi patches: https://hostupdate.vmware.com

o For Windows and Linux VM and applications: https://shavlik.com

 Download information about a set of security patches and one or more of these patches

are aggregated to form a baseline

 A collection of VMs, virtual appliances, and ESX/ESXi hosts can be scanned for

compliance with a baseline or a baseline group and remediated

o This process can be started manually or through scheduled tasks



Major components of Update Manager

 Update Manager Server

o Can be installed directly on the vCenter Server or on a separate system

o Can be physical or virtual

 Patch database

o You can use the same database server that vCenter uses

o It will require a unique database with a DSN system ODBC connection already

configured

o Can also use SQL Server 2005 Express

 Update Manager plugin

o Runs on the same system on which the vSphere Client is installed

 Guest agents

o Installed into VMs from the Update Manager server and are used in the scanning

and remediation operations

 Optional Download server to download patches

Installing Update Manager

 Runs on Windows XP SP2, Windows Server 2003 and Windows Server 2008

o Must be 64 bit

 Must have 2 or more logical cores, each with a speed of 2GHz

 If Update Manager and vCenter are on different machines, 2GB of RAM is required

 If they are on the same machine, 4GB of RAM is required

 Before installing Update Manger, you must create a database instance and configure it

to ensure that all Update Manger tables are placed in it

 Supports the following databases

o SQL Server 2005

o SQL Server 2008

o Oracle 10g

o Oracle 11g

o SQL Server 2005 Express for small scale environments (up to 5 hosts and 50 VMs)

 Required information

o vCenter IP address or host name

o Port numbers ( 80 and 443 are defaults)

o Administrative credentials

o The system DNS name, plus the user name and password for the database that

Update Manager will work with



Configuring Update Manager Settings (Configuration tab)

 Network Connectivity

o Network settings such as IP address or host name for the patch store

 Patch Download Settings

 Patch Download Schedule

 Notification Check Schedule

 Virtual Machine Settings

o Whether to take a snapshot of the VMs before remediation to enable rollback

and how long to keep snapshots

 ESX Host/Cluster Settings

o How Update Manager responds to a failure that might occur when placing an

ESX/ESXi host in maintenance mode

 vApp Settings

o Enable or disable smart reboot of virtual appliances after remediation

Baseline and Baseline Groups

 Baselines contain a collection of one or more patches, extensions, service packs, bug

fixes, or upgrades

 Baselines can be classified as an upgrade, extension, or patch baselines

o An extension refers to additional software for ESX/ESXi hosts

 Baseline types

o Host patch – A set of patches to apply to a host or set of hosts, based on

applicability

o Host extension – A fixed set of extensions for your ESX/ESXi host

o Host upgrade – An upgrade release that allows you to upgrade host to a

particular release version

o Virtual machine patch – A set of patches that apply to one or more VMs, based

on applicability

o Virtual appliance upgrade – A set of patches to the operating system or

application in the virtual appliance

 Baseline groups are assembled from existing baselines



Creating a Baseline

 Click Create

 Specify name and description

 Choose a baseline type

 For a patch baseline, select a patch option: Fixed or Dynamic

 Select patches to add to the baseline

 A Fixed baseline remains the same even if new patches are added to the repository.

With a fixed patch baseline, the user manually specifies all updates included in the

baseline from all patches available in Update Manager

 A Dynamic baseline is updated when new patches meeting the specified criteria are

added to the repository



To view compliance information and remediate objects in the inventory against specific

baselines and baseline groups, attach existing baselines and baseline groups to these objects



Although you can attach baselines and baseline groups to individual objects, attaching them to

container objects, such as folders, hosts, clusters and datacenters is more efficient.



To attach baselines to virtual machines, templates and virtual appliances, go to the VMs and

Templates inventory view

Scanning for Updates

 Scanning is the process in which attributes of a set of hosts, virtual machines, virtual

appliances, are evaluated against patches, extensions, and upgrades in the attached

baselines and baseline groups

 If the object that you select is a container object, all child objects are also scanned

 To schedule the scan go to Home>Management>Scheduled Tasks



Viewing Compliancy

 To view compliancy, select the object in the appropriate inventory view and click the

Update Manger tab

o To view VM compliancy, you must use the VMs and Templates inventory view

 Staging allows you to download the patches and extensions from the Update Manger

server to the ESX/ESXi hosts, without applying the patches and extensions immediately.

o Staging patches and extensions speeds up the remediation process because the

patches and extensions are already available locally on the hosts.



Remediating Objects

 You can remediate VMs, virtual appliances, and hosts by user initiated remediation for

regularly scheduled remediation

o Right click the object and select Remediate

 For ESX/ESXi hosts in a cluster, the remediation process is sequential

 When you remediate a cluster of hosts and one of the hosts fails to enter maintenance

mode, Update Manager reports an error and the process fails

 For multiple clusters under a datacenter, the remediation processes run in parallel

 If the remediation process fails for one of the clusters within a datacenter, the

remaining clusters are still remediated

 To remediate virtual machines and virtual appliances together, they must be in one

container, such as a folder, a vApp, or a datacenter

 Templates are a type of VM so they can be remediated – take snapshots first, especially

if sealed

o A template that is sealed is stopped before the OS installation is completed

 VMs are rebooted at the end of the patch remediation process

 VMs must be powered on to be remediated

 For Linux and Windows operating systems, the guest agent is automatically installed the

first time a patch remediation is scheduled or when a patch scan is started on a

powered on machine

 Remediation of hosts in a cluster requires that you temporarily disable cluster features

like DPM and HA admission control

 You should also turn off FT if its enabled on any of the VMs on a host

 Disconnect the removable devices connected to the VM

 Before you start the remediation process, you can generate a report that shows which

cluster, host or VM is with enabled cluster features

Patch Recall Notification

 When patches with problems or potential problems are released, these patches are

recalled in the metadata, and Update Manager marks them as recalled

 Update Manager also detects all the recalled patches from the Update manager patch

repository



vCenter can migrate the VMs if the cluster is configured for vMotion and if DRS and EVC are

enabled





Installing ESX

 Hardware requirements:

o 64 bit processor (AMD Opteron, Intel Xeon, or Intel Nehalem)

o Minimum of 3GB RAM

o One or more Ethernet controllers

o A SCSI adapter, a Fibre Channel adapter, or an internal RAID controller

 ESX can be installed by either:

o The graphical user interface (default)

o Text mode (keyboard only)

o An installation script



During ESX installation, the following physical partitions are created:

 /boot

o Contains the ESX software and its support files

o The disk that you install /boot onto must be the disk the BIOS chooses to boot

from

 vmkcore

o Required to store core dumps for troubleshooting

o VMware does not support ESX host configurations without a vmkcore partition

 No scratch disks are created in ESX

You cannot use the graphical or text-based installation to define the size of the /boot,

vmkcore, and /vmfs/volumes partitions. But you can define their size by a scripted

installation

 The /vmfs/volumes partition also holds the ESX service console, located in a virtual disk

named esxconsole.vmdk. esxconsole.vmdk contains the Linux based service console

operating system. The files that make up the service console are found in a file system

called root (/)

 The service console has its own swap partition. Configure the swap partition size to be

at least twice the size of the service console’s RAM allocation

Installing ESXi

 Hardware requirements

o 64 bit processor (AMD Opteron, Intel Xeon, or Intel Nehalem)

o Up to 128 logical CPUs (cores or hyperthreads)

o Can support up to 512 virtual CPUs per host

o Minimum of 2GB of RAM with a maximum of 1TB

 The ESXi host must have

o One or more Ethernet controllers

o A basic SCSI controller

o An internal RAID controller

o A SCSI disk or a local RAID LUN

 ESXi Installable supports installing and booting from SATA, SCSI or SAS disk drives



ESXi Partitions

 The disk formatting software retains existing diagnostic partitions that are created by

the hardware vendor. In the remaining space, the software creates:

 One 4GB VFAT scratch partition for system swap

o Not required but used to store vm-support output which you need when you

create a support bundle

 One 110MB diagnostic partition for core dumps

 One vStorage VMFS data store on the remaining free space

 On other disks, the software creates one VMFS datastore per blank disk, using the whole

disk.

 The software formats blank disks only



You can take 2 approaches with ESXi scripted installation:

 Create multiple scripts, each containing unique network identification information

 Create one script (or use the default script) that uses DHCP to setup multiple ESXi hosts



To perform a scripted installation

 Create an installation script

o Custom script using script commands

o Default script: ks.cfg (kickstart)

o Make the installation media accessible to the host:

CD/DVD

o USB flash drive

o Media depot, accessible by HTTP/HTTPS, FTP, NFS

 Start the installer

o Boot from local CD/DVD ROM drive or PXE boot server

 Enter boot command to run the installations script

o Example: ks=cdrom: /ks.cfg



The default root password is mypassword

ks=cdrom: /ks.cfg – Calls the installation script located on the DVD-ROM drive attached to the

machine



ks=ftp: ///path/ks.cfg – Calls the installation script located at the given URL



Default installation script commands: ks.cfg

 vmaccepteula

o Accept the ESXi license agreement

 rootpw mypassword

o Set the root password to mypassword

 autopart –firstdisk –overwirtevmfs

o Choose the first discovered disk to install on to

 Install cdrom

o The installation media is the CD-ROM drive

 Network –bootproto=dhcp –device=vmnic0

o Set the network to DHCP on the first network adapter



If you are using a media depot, use install nfs or install url to point to the media depot







vSphere 4.1 Maximums



Virtual Machine Maximums

CPU

Virtual CPUs per virtual machine (Virtual SMP) – 8

Memory

RAM per virtual machine – 255GB

Virtual machine swap file size – 255GB

Storage Virtual Adapters and Devices

Virtual SCSI adapters per virtual machine - 4

Virtual SCSI targets per virtual SCSI adapter - 15

Virtual SCSI targets per virtual machine - 60

Disk size 2TB minus - 512 bytes

IDE controllers per virtual machine - 1

IDE devices per virtual machine - 4

Floppy controllers per virtual machine - 1

Floppy devices per virtual machine – 2

Networking Virtual Devices

Virtual NICs per virtual machine – 10

Virtual Peripheral Ports

USB controllers per virtual machine – 1

USB devices connected to a virtual machine 20

Parallel ports per virtual machine - 3

Serial ports per virtual machine - 4

Miscellaneous

Concurrent remote console connections to a virtual machine – 40



ESX Host Maximums

Compute Maximums

Host CPU maximums

Logical CPUs per host up to - 160

vSphere 4.1 supports up to 128, and vSphere 4.1 Update 1 supports up to 160

Virtual machine maximums

Virtual machines per host - 320

Virtual CPUs per host - 512

Virtual CPUs per core - 252

Fault Tolerance maximums

Virtual disks - 16

Virtual CPUs per virtual machine - 1

RAM per FT VM (GB) - 64

Virtual machines per host - 4



Memory Maximums

RAM per host - 1TB

Maximum RAM allocated to service console - 800MB

Minimum RAM allocated to service console - 272MB

Number of swap files - 1 per virtual machine

Swap file size - Same as maximum virtual machine RAM

Storage Maximums

iSCSI Physical

LUNs per server - 256

Qlogic 1Gb iSCSI HBA initiator ports per server - 4

Broadcom 1Gb iSCSI HBA initiator ports per server - 4

Broadcom 10Gb iSCSI HBA initiator ports per server - 4

NICs that can be associated or port bound with the software iSCSI stack per server- 8

Number of total paths on a server - 1024

Number of paths to a LUN (software iSCSI and hardware iSCSI) - 8

Qlogic iSCSI: dynamic targets per adapter port - 64

Qlogic iSCSI: static targets per adapter port - 62

Broadcom 1Gb iSCSI HBA targets - 64

Broadcom 10Gb iSCSI HBA targets - 64

Software iSCSI targets - 2561

NAS

NFS mounts per host - 64

Fibre Channel

LUNs per host - 256

LUN size - 2TB minus 512 bytes

LUN ID - 255

LUNs concurrently opened by all virtual machines - 256

Number of paths to a LUN - 32

Number of total paths on a server - 1024

Number of HBAs of any type - 8

HBA ports - 16

Targets per HBA – 256

VMFS

Raw device mapping (RDM) size - 2TB minus 512 bytes

Volume size - 64TB

Volumes per host - 256

Hosts per volume - 64

VMFS-3

Block size - 8MB

File size (1MB block size) - 256GB

File size (2MB block size) - 512GB

File size (4MB block size) - 1TB

File size (8MB block size) - 2TB minus 512 bytes

Files per volume - Approximately 30,720

Networking Maximums

Physical NICs

e1000 1GB Ethernet ports (Intel PCI‐x) - 32

e1000e 1GB Ethernet ports (Intel PCI‐e) - 24

igb 1GB Ethernet ports (Intel) - 16

tg3 1GB Ethernet ports (Broadcom) - 32

bnx2 1GB Ethernet ports (Broadcom) - 16

forcedeth 1GB Ethernet ports (NVIDIA) - 2

s2io 10GB Ethernet ports (Neterion) - 4

nx_nic 10GB Ethernet ports (NetXen) - 4

ixgbe Oplin 10GB Ethernet ports (Intel) - 4

bnx2x 10GB Ethernet ports (Broadcom) - 4

Infiniband ports (refer to VMware Community Support) - N/A

VMDirectPath limits

VMDirectPath PCI/PCIe devices per host - 8

VMDirectPath PCI/PCIe devices per virtual machine – 4

vNetwork Standard and Distributed Switch

Total virtual network switch ports per host (vDS and vSS ports) - 4096

Maximum ACTIVE ports per host (vDS and VSS) - 1016

Virtual network switch creation ports per standard switch - 4088

Port groups per standard switch - 512

Static or Dynamic Port groups per distributed switch - 5000

Ephemeral Port groups per distributed switch - 1016

Ports per distributed switch – 20,000

Distributed virtual network switch ports per vCenter – 20,000

Static or Dynamic Port groups per vCenter - 5,000

Ephemeral Port groups per vCenter - 1,016

Distributed switches per vCenter - 32

Distributed switches per Host - 16

Hosts per distributed switch – 350



ESX Resource Pool and Cluster Maximums

Cluster (all clusters including HA and DRS)

Hosts per cluster - 32

Virtual machines per cluster - 3,000

Virtual machines per host - 320

Maximum concurrent host HA failover - 4

Failover as percentage of cluster - 50%

Resource pools per cluster - 512

Resource Pool

Resource pool tree depth - 8

Resource pools per host - 4,096

Children per resource pool – 1,024



vCenter Server Maximums

vCenter Server Scalability

Hosts per vCenter Server - 1,000

Powered on virtual machines per vCenter Server - 10,000

Registered virtual machines per vCenter Server - 15,000

Linked vCenter Servers - 10

Hosts in linked vCenter Servers - 3,000

Powered on virtual machine in linked vCenter Servers - 30,000

Registered virtual machine in linked vCenter Servers - 50,000

Concurrent vSphere Clients - 100

Number of host per datacenter - 400

Concurrent Operations

Concurrent provisioning operations per host - 4

Concurrent provisioning operations per datastore - 4

Concurrent vMotion operations per host (1Gb/s network) - 4

Concurrent vMotion operations per host (10Gb/s network) - 8

Concurrent vMotion operations per VMFS3 datastore - 128

Concurrent Storage vMotion operations per host - 2

Concurrent Storage vMotion operations per datastore – 8



VMware vCenter Update Manager

vCenter Update Manager Scalability

Host scans in a single vCenter Server - 1,000

Virtual machine scans in a single vCenter Server - 10,000

Cisco VDS update and deployment - 70

Concurrent Operations

Virtual machine remediation per ESX host - 5

Powered‐on Windows virtual machine scans per ESX host - 5

Powered‐off Windows virtual machine scans per ESX host - 5

Powered‐on Linux virtual machine scans per ESX host - 2

VMware Tools scan per ESX host - 24

VMware Tools upgrade per ESX host - 24

Virtual machine hardware scan per host - 24

Virtual machine hardware upgrade per host - 24

Virtual machine remediation per VUM server - 48

Powered‐on Windows virtual machine scan per VUM server - 17

Powered‐off Windows virtual machine scan per VUM server - 10

Powered‐on Linux virtual machine scan VUM server - 8

VMware Tools scan per VUM server - 75

VMware Tools upgrade per VUM server - 75

Virtual machine hardware scan per VUM server - 75

Virtual machine hardware upgrade per VUM server - 75

ESX host scan per VUM server - 70

ESX host remediation per VUM server - 8

ESX host upgrade per VUM server - 44

ESX host upgrade per cluster – 1



VMware vCenter Orchestrator

Connected vCenter Server systems - 10

Connected ESX/ESXi instances - 100

Connected virtual machines - 15,000

Concurrent running workflows – 150



VMware vCenter Converter

Concurrent virtual machine to virtual machine import or export tasks 8

Concurrent physical machine to virtual machine import or export tasks 20



vSphere Storage Management Initiative - Specification (SMI-S)

Number of vCenter Server systems connected - 1

Number of ESX/ESXi hosts connected - 1,000

Number of ESX/ESXi hosts managed by vCenter Server - 320

Number of virtual machines registered in vCenter Server - 15,000