Vmware – vCenter Server Heartbeat – 6.5 – Administrator Guide

Category: Software and Application User Guides and Manuals

Download user manual for Vmware – vCenter Server Heartbeat – 6.5 – Administrator Guide 
Preview: Below is a preview of the manual as extracted from the PDF file


Administrator Guide
vCenter Server Heartbeat 6.5
This document supports the version of each product listed and
supports all subsequent versions until the document is replaced
by a new edition. To check for more recent editions of this
document, see http://www.vmware.com/support/pubs.
EN-000948-00You can find the most up-to-date technical documentation on the VMware Web site at:
http://www .vmw are.com/support/
The VMware Web site also provides the latest product updates.
If you have comments about this documentation, submit your feedback to:
docfeedback@vmw are.com
Copyright
©
2012 VMware, Inc. All rights reserved. This product is protected by U.S. and international copyright and intellectual
property laws. VMware products are covered by one or more patents listed at http://www .vmw are.com/go/patents.
VMware is a registered trademark or trademark of VMware, Inc. in the United States and/or other jurisdictions. All other marks
and names mentioned herein may be trademarks of their respective companies.
VMware, Inc.
3401 Hillview Ave.
Palo Alto, CA 94304
www.vmware.com
2 VMware, Inc.
Administrator GuideContents
About This Book 5
Introduction 9 1
vCenter Serv er Heartbeat Concepts 9
vCenter Serv er Heartbeat Protection Lev els 10
vCenter Serv er Heartbeat Communications 13
vCenter Serv er Heartbeat Switchov er and Failov er Processes 15
Configuring vCenter Serv er Heartbeat 19 2
Launching the Configure Serv er W izard 20
Configuring the Machine 21
Configuring the VMw are Channel 22
Configuring Public IP Addressing 24
Managing vCenter Serv er Heartbeat License Keys 26
Configuring the Logs 27
Serv er Protection 29 3
Checking the Serv er P air Status 29
Monitoring the Status of Serv ers 31
Configuring Heartbeat Settings 31
Netw ork Protection 41 4
Communications Status 41
Reviewing the VMw are Channel Status 41
Configuring Public Netw ork Connection Checks 42
Enabling Automatic Switchov er in a W AN 42
Setting Max Serv er Time Difference 43
Application Protection 45 5
Applications: Summary T ab 45
Applications: Services T ab 48
Applications: T asks T ab 51
Applications: Plug-ins T ab 53
VMware, Inc. 3Status and Control 55 6
About vCenter Serv er Heartbeat Console 55
W orking with Groups 57
W orking with P airs in vCenter Serv er Heartbeat Groups 57
Reviewing the Status of vCenter Serv er Heartbeat Groups and P airs 60
Exiting vCenter Serv er Heartbeat Console 60
Shutting Down W indows W ithout Stopping vCenter Serv er Heartbeat 60
Controlled Shutdown 61
vSphere W eb Client Plug-in 61
Uninstalling vCenter Serv er Heartbeat 63
P erformance Protection 65 7
Applications: Rules T ab 65
Data Protection 69 8
Replication 69
File Filters 74
Automatic Filter Discov ery 75
Alerts and Ev ents 77 9
Configuring Alerts 77
Configuring Ev ent Log Files 79
Reviewing Ev ent Logs 81
T roubleshooting 83 10
T w o Activ e or T w o P assiv e Serv ers 83
Synchronization Failures 86
Registry Status is Out-of-Sync 88
Channel Drops 89
Subnet or Routing Issues 93
MaxDiskUsage Errors 94
Application Slowdown 97
Glossary 101
4 VMware, Inc.
Administrator GuideAbout This Book
The Administrator Guide provides information about configuring VMw are vCenter Serv er Heartbeat serv er
netw ork protection, application protection, data protection, Split-brain A v oidance, and more. T o help y ou
protect y our VMw are vCenter Serv er , this book provides an ov erview of the protection offered by vCenter
Serv er Heartbeat and the actions that vCenter Serv er Heartbeat can take in the ev ent of a netw ork, hardw are,
or application failure.
Intended Audience
This guide assumes the reader has a w orking knowledge of netw orks including the configuration of TCP/IP
protocols and domain administration on the W indows™ 2003 and 2008 platforms, notably in Activ e Directory
and DNS.
VMware T echnical Publications Glossary
VMw are T echnical Publications provides a glossary of terms that might be unfamiliar to y ou. For definitions
of terms as they are used in VMw are technical documentation go to www .vmw are.com/support/pubs .
Overview of Content
This book is designed to provide guidance on the configuration and administration of vCenter Serv er Heartbeat,
and is organized into the following sections:
■ Preface — About This Book (this chapter) provides an ov erview of this guide and the conv entions used
throughout.
■ Chapter 1 — Intr oduction presents an ov erview of vCenter Serv er Heartbeat concepts including the
Switchov er and Failov er processes.
■ Chapter 2 — Configuring vCenter Server Heartbeat shows y ou how to use the Serv er Configuration wizard
to configure y our new installation of vCenter Serv er Heartbeat.
■ Chapter 3 — Server Pr otection presents an ov erview of how vCenter Serv er Heartbeat provides protection
against serv er system crash or serv er hardw are failure, shows y ou how to check the serv er pair status,
and explains how to configure settings, shutdown options, and Split-Brain A v oidance.
VMware, Inc. 5■ Chapter 4 — Network Pr otection describes how vCenter Serv er Heartbeat protects against netw ork failure
and provides a w ay to monitor communication status. It also explains how to configure public netw ork
connection checks and maximum serv er time difference.
■ Chapter 5 — Application Pr otection discusses how vCenter Serv er Heartbeat maintains the protected
application environment ensuring that applications and services stay aliv e on the netw ork.
■ Chapter 6 — Status and Contr ol introduces y ou to the vCenter Serv er Heartbeat Console and shows y ou
how to configure its look and feel.
■ Chapter 7 — Performance Pr otection describes how vCenter Serv er Heartbeat monitors system and application
attributes to prev ent an unexpected system or application failure.
■ Chapter 8 — Data Pr otection discusses how vCenter Serv er Heartbeat intercepts all data written by users
and protected applications and maintains a copy of this data for use in case of failure.
■ Chapter 9 — Alerts and Events discusses additional tasks for the administrator to configure system logging
and alerting functions.
■ Chapter 10 — T r oubleshooting provides techniques to troubleshoot common issues and unexpected behaviors.
Document Feedback
VMw are w elcomes y our suggestions for improving our documentation and invites y ou to send y our feedback
to docfeedback@vmw are.com.
Abbreviations Used in Figures
Description Abbreviation
VMw are Channel Channel
Netw ork Interface Card NIC
Physical to Physical P2P
Physical to V irtual P2V
V irtual to V irtual V2V
T echnical Support and Educational Resources
The following sections describe the technical support resources av ailable to y ou. T o access the current v ersion
of this book and other books, go to www .vmw are.com/support/pubs.
Online and T elephone Support
Go to www .vmw are.com/support to use online support to submit technical support requests, view y our
product and contract information, and register y our products.
Go to www .vmw are.com/support/phone_support.html to find out how to use telephone support for the fastest
response on priority 1 issues (applies to customers with appropriate support contracts).
6 VMware, Inc.
Administrator GuideSupport Offerings
Go to www .vmw are.com/support/services to find out how VMw are support offerings can help meet y our
business needs.
VMware Professional Services
Go to www .vmw are.com/services to access information about educational classes, certification programs, and
consulting services. VMw are Education Services courses offer extensiv e hands-on labs, case study examples,
and course materials designed for use as on-the-job reference tools. Courses are av ailable onsite, in the classroom,
and liv e online. For onsite pilot programs and implementation best practices, VMw are Consulting Services
provides offerings to help y ou assess, plan, build, and manage y our virtual environment.
VMware, Inc. 7
About This Book8 VMware, Inc.
Administrator Guide1 Introduction
This chapter includes the following topics:
■ “vCenter Serv er Heartbeat Concepts,” on page 9
■ “vCenter Serv er Heartbeat Protection Lev els,” on page 10
■ “vCenter Serv er Heartbeat Communications,” on page 13
■ “vCenter Serv er Heartbeat Switchov er and Failov er Processes,” on page 15
vCenter Server Heartbeat Concepts
vCenter Serv er Heartbeat is a W indows based service specifically designed to provide high av ailability protection
for vCenter Serv er configurations without requiring any specialized hardw are.
Architecture Overview
vCenter Serv er Heartbeat uses an “ Activ e–P assiv e” architecture which enables it to provide “High A v ailability
(HA)” in a Local Area Netw ork (LAN) or “Disaster Recov ery (DR)” in a W ide Area Netw ork (W AN) for
vCenter Serv er , V iew Composer and SQL Serv er . The vCenter Serv er Heartbeat softw are is installed on an
existing production serv er known as the “Primary” serv er running the protected applications (vCenter Serv er ,
V iew and SQL Serv er). An additional serv er , known as the “Secondary” serv er , operates as a ready standby
serv er to provide service in the ev ent of an application, system, or hardw are failure, or when the Administrator
needs to perform system maintenance. The terms Primary and Secondary refer to the “Identity” of each serv er
instance, and these identities do not change.
Active / Passive Roles
The applications protected by vCenter Serv er Heartbeat will run on the “ Activ e” serv er . Only one serv er can
be activ e at any one time and the activ e serv er will host the “Principal (Public) IP Address” which is used by
clients to access the application. The “P assiv e” serv er is only accessible on the netw ork via its assigned
“Management IP Address”. Activ e and passiv e refer to the “Role” that the serv er is performing. The role can
be changed by a “Failov er ” or when the administrator performs a “Switchov er ”. T o ensure the serv ers can
provide a seamless switchov er / failov er experience for clients, the serv ers need to be symmetrical. T o ensure
that the Secondary serv er has all the programs and components installed in the same location, the install process
includes a cloning procedure. Clients will continue to connect to vCenter Serv er or SQL Serv er using the original
and unique fully qualified domain name that w as used previously by clients. During installation, a service
name is configured in vCenter Serv er Heartbeat which will continue to resolv e in DNS to the Principal (Public)
IP address regardless of which serv er is hosting the Principal (Public) IP address.
VMware, Inc. 9Managing the Primary and Secondary Servers
T o allow management of vCenter Serv er Heartbeat serv er pairs using standard netw ork, domain policy , and
domain management procedures, vCenter Serv er Heartbeat is deploy ed so that Primary and Secondary serv ers
use unique domain names. Each domain name must differ from the fully qualified domain name used by the
original vCenter or SQL Serv ers. A Management IP address on each serv er ensures that the Administrator can
access the serv er ev en when it is passiv e. This allows monitoring with 3rd party monitoring tools and
maintenance operations such as updating anti-malw are definition files, operating system hot-fixes and updates.
Switchover/Failover in a LAN
When deploy ed in a LAN environment, the Principal (Public) IP address is mov ed betw een the Primary and
Secondary serv ers as the roles change from activ e to passiv e so that vCenter Serv er or SQL Serv er are av ailable
to clients only when the serv er assumes the activ e role. vCenter Serv er Heartbeat does not require updates to
DNS during the switchov er / failov er , how ev er the DNS serv er must be preconfigured with the Management
IP addresses. Adjacent IP addresses should be reserv ed and used for the Principal (Public) IP address and the
Management IP addresses for the Primary and Secondary Serv ers when installing vCenter Serv er Heartbeat
on serv ers running W indows Serv er 2008. When vCenter Serv er Heartbeat is started, the Principal (Public) IP
address is added to the activ e serv er . When a switchov er is requested the Principal (Public) IP address is
remov ed from the activ e serv er as it becomes passiv e and then added to the passiv e serv er which is being
made activ e.
Additionally , vCenter Serv er Heartbeat can be deploy ed in a stretched VLAN using the same subnet for the
production and the disaster recov ery site. Similar to a LAN installation, this configuration requires that both
the Primary and Secondary serv ers share the Principal (Public) IP address with the activ e serv er rev ealing the
Principal (Public) IP address while the passiv e serv er is hidden from the netw ork. This means that vCenter
Serv er Heartbeat can be deploy ed without any changes to DNS during switchov er or failov er operations.
Switchover/ Failover in a W AN
vCenter Serv er Heartbeat can be deploy ed in a W AN where each site uses different subnets. When deploy ed
in this manner , each site will hav e a different Principal (Public) IP address. When a switchov er or failov er
occurs, vCenter Serv er Heartbeat will update the DNS serv er with the Principal (Public) IP address of the new
site thereby allowing clients to connect to the new site.
vCenter Server Heartbeat Protection Levels
vCenter Serv er Heartbeat provides the following protection lev els:
■ Server Pr otection — vCenter Serv er Heartbeat provides continuous av ailability to end users through a
hardw are failure scenario or operating system crash. Additionally , vCenter Serv er Heartbeat ensures users
are provided with a replica serv er should the production serv er fail.
■ Network Pr otection — vCenter Serv er Heartbeat proactiv ely monitors the netw ork by polling up to three
predefined nodes to ensure that the activ e serv er is visible on the netw ork.
■ Application Pr otection — vCenter Serv er Heartbeat maintains the application environment ensuring that
applications and services stay aliv e and are av ailable on the netw ork.
■ Performance Pr otection — vCenter Serv er Heartbeat proactiv ely monitors system performance attributes
to ensure the system administrator is notified of problems. Additionally , it can be configured to take
pre-emptiv e action to prev ent an outage.
■ Data Pr otection — vCenter Serv er Heartbeat intercepts all data written by users and applications, and
maintains a copy of the data on the passiv e serv er that can be used in the ev ent of a failure.
10 VMware, Inc.
Administrator GuidevCenter Serv er Heartbeat provides all fiv e protection lev els continuously , ensuring all facets of the user
environment are maintained at all times, and that vCenter Serv er continues to operate through as many failure
scenarios as possible.
Server Protection
vCenter Serv er Heartbeat provides continuous av ailability to end users through a hardw are failure scenario
or operating system crash and ensures users are provided with a replica serv er and its IP address on the failure
of the production serv er .
T w o instances of vCenter Serv er Heartbeat regularly send “I’ m aliv e” messages and message acknowledgments
to one another ov er a dedicated netw ork connection referred to as the “VMw are Channel” to detect interruptions
in responsiv eness. If the passiv e serv er detects that this monitoring process (referred to as the “Heartbeat”)
has failed, it initiates a failov er as illustrated in Figure 1-1.
Figure 1- 1. Failover
A failov er occurs when the passiv e serv er detects that the activ e serv er is no longer responding. This can occur
when the activ e serv er hardw are fails, loses its netw ork connections, or otherwise becomes unav ailable. Rather
than the activ e serv er gracefully closing, the passiv e serv er determines that the activ e serv er has failed and
requires no further operations. In a failov er , the passiv e serv er immediately assumes the activ e serv er role.
The failov er process is discussed in detail later in this guide.
Network Protection
vCenter Serv er Heartbeat proactiv ely monitors the netw ork by polling up to three predefined nodes to ensure
that the activ e serv er is visible on the netw ork. vCenter Serv er Heartbeat polls by default the primary DNS
serv er , the default gatew ay , and the global catalog serv er at regular interv als. If all three nodes fail to respond,
for example in the case of a netw ork card or local switch failure, vCenter Serv er Heartbeat can initiate a
switchov er , allowing the Secondary serv er to assume the activ e role and service clients.
Application Protection
vCenter Serv er Heartbeat running on the activ e serv er locally monitors vCenter Serv er and its services (through
the use of plug-ins) to v erify that vCenter Serv er is operational and not in an unresponsiv e or stopped state.
This lev el of monitoring is fundamental in ensuring that vCenter Serv er remains av ailable to users.
VMware, Inc. 1 1
Chapter 1 IntroductionIf vCenter Serv er should fail, vCenter Serv er Heartbeat first tries to restart the application on the activ e serv er
(1) in Figure 1-2 .
If the application does not successfully restart, vCenter Serv er Heartbeat initiates a switchov er (2) in Figure
1-2. Refer to “vCenter Serv er Heartbeat Switchov er and Failov er Processes,” on page 15 for further information
about the switchov er process.
Figure 1- 2. Switchover
A switchov er gracefully closes vCenter Serv er running on the activ e serv er and restarts it on the passiv e serv er ,
including the component or service that caused the failure. For example, if the Primary serv er is activ e and
the Secondary serv er is passiv e, the Primary serv er is demoted to a passiv e role and is hidden from the netw ork
when the Secondary serv er is promoted to an activ e role and is made visible to the netw ork. The mechanics
of switchov ers are discussed in more detail later in this guide.
Performance Protection
Ensuring that vCenter Serv er is operational and providing service at an adequate lev el of performance to meet
user demands is important. The vCenter Serv er Heartbeat plug-in provides these monitoring and pre-emptiv e
repair capabilities.
vCenter Serv er Heartbeat proactiv ely monitors system performance attributes and can notify the system
administrator in the ev ent of a problem. Additionally , it can be configured to take pre-emptiv e action to prev ent
an outage.
In addition to monitoring vCenter Serv er services, vCenter Serv er Heartbeat can monitor specific attributes to
ensure that they remain within normal operating ranges. Similar to application monitoring, v arious rules can
be configured to trigger specific correctiv e actions whenev er these attributes fall outside of their respectiv e
ranges.
vCenter Serv er Heartbeat provides the same lev el of flexibility to define and perform multiple correctiv e actions
in the ev ent of problems on a service by service or ev en attribute by attribute basis.
Data Protection
Y ou can configure vCenter Serv er Heartbeat to protect the application environment. All data files that users
or the vCenter Serv er requires in the application environment are made av ailable should a failure occur . After
installation, vCenter Serv er Heartbeat configures itself to protect files, folders, and registry settings for vCenter
Serv er on the activ e serv er by mirroring them in real time to the passiv e serv er . If a failov er occurs, all files
protected on the failed serv er are av ailable to users after the failov er , hosted on the Secondary serv er .
12 VMware, Inc.
Administrator GuidevCenter Serv er Heartbeat intercepts all file system I/O operations on the activ e serv er . If the intercepted write
and update operations are within the protected set, these are placed in a queue on the activ e serv er referred
to as the activ e serv er ’ s “Send Queue”, pending transmission to the passiv e serv er . Each request is numbered
to maintain its order in the queue. W ith the request in the activ e serv er ’ s send queue, vCenter Serv er Heartbeat
allows the disk I/O to continue with the requested disk operation.
If the channel is connected, the activ e serv er ’ s send queue is transferred to the passiv e serv er , which places all
the requests in the passiv e serv er ’ s “Receiv e Queue” . The passiv e serv er confirms the changes w ere logged by
sending the activ e serv er an acknowledgment. The activ e serv er then clears the data from its queue.
Figure 1- 3. Apply Process
The apply process running on the passiv e serv er ’ s receiv e queue applies all updates in strict sequence,
duplicating an identical set of file operations on the passiv e serv er as illustrated in Figure 1-3.
vCenter Server Heartbeat Communications
The VMw are Channel is a crucial component of the setup and can be configured in a number of w ays.
vCenter Serv er Heartbeat supports installation using either multiple netw ork interface connections (NICs), (1
configured for the VMw are Channel) and (1 configured with the Principal(Public) and Management IP
addresses), or a single NIC configured with the VMw are Channel IP , Principal (Public) IP , and Management
IP . Both the Primary and Secondary serv ers must hav e the same number of NICs. The Principal (Public) IP
address provides client access and the Management IP address provides administrativ e access, while the
VMw are Channel provides for data transfer and control.
A second pair of NICs for the VMw are Channel can be used to provide a degree of redundancy . In this
configuration, the VMw are Channel has a dual channel if more than one dedicated NIC is provided for the
VMw are Channel on each serv er . T o provide added resilience, the communications for the second channel
should be completely independent from the first channel. They should not share any switches, virtual switches,
routers or the same W AN connection.
VMware, Inc. 13
Chapter 1 IntroductionFigure 1- 4. Single NIC Configuration
Figure 1- 5. Multi-NIC Configuration
The IP address a client uses to connect to the activ e serv er (the Principal (Public) IP address) must be configured
as a static IP address and not DHCP (Dynamic Host Configuration Protocol) enabled. In Figure 1-4 and Figure
1-5, the Principal (Public) IP address is configured as 192.168.1.127.
LAN Deployment
When deploy ed in a LAN, the Principal (Public) NIC on the activ e serv er is configured with both a unique
permanently assigned Management IP address for administrativ e access and the Principal (Public) IP address
which allows traffic from clients. The Principal (Public) NIC on the passiv e serv er is configured to use its
unique permanently assigned Management IP address. When a switchov er or failov er occurs, the Principal
(Public) IP address assigned to the currently activ e serv er is remov ed and reassigned to the new activ e serv er .
The new passiv e serv er remains accessible to administrators via the Management IP address but is not visible
to clients. The newly activ e serv er then starts accepting traffic from clients.
The NICs on the activ e and passiv e serv ers used for the VMw are Channel are configured so that their IP
addresses are outside of the subnet range of the Principal (Public) netw ork. These addresses are referred to as
VMw are Channel addresses.
During installation, setup will switch off NetBIOS for the VMw are Channel(s) on the activ e and passiv e serv ers
as this connection remains liv e and both the passiv e and activ e machines hav e the same NetBIOS name.
Following vCenter Serv er Heartbeat installation (runtime), NetBIOS is disabled across the channel(s).
14 VMware, Inc.
Administrator GuideThe NICs that support connectivity across the VMw are Channel can be standard 100BaseT Ethernet cards
providing a throughput of 100 Mbits per second across standard Cat-5 cabling. In its most basic form, a dedicated
channel requires no hubs or routers, but the direct connection requires crossov er cabling.
W AN Deployment
When configured for a W AN deployment, configure the VMw are Channel to use static routes ov er switches
and routers to maintain continuous communications independent from corporate or public traffic.
vCenter Server Heartbeat Switchover and Failover Processes
vCenter Serv er Heartbeat uses four different procedures — managed switchov er , automatic switchov er ,
automatic failov er , and managed failov er — to change the role of the activ e and passiv e serv ers depending on
the status of the activ e serv er .
Managed Switchover
Y ou can click Make Active on the vCenter Serv er Heartbeat Console Server: Summary page to manually initiate
a managed switchov er . When a managed switchov er is triggered, the running of protected applications is
transferred from the activ e machine to the passiv e machine in the serv er pair . The serv er roles are rev ersed.
Figure 1- 6. Switchover
A managed switchov er performs the following steps:
1 Stop the protected applications on the activ e serv er . After the protected applications stop, no more disk
updates are generated.
2 Send all updates that are still queued on the activ e serv er to the passiv e serv er . After this step, all updates
are av ailable on the passiv e serv er .
3 Re-designate the Secondary serv er as the new activ e serv er . After this step, vCenter Serv er Heartbeat:
■ Reassigns the Principal (Public) IP address to the Secondary serv er .
■ Makes the newly activ e serv er visible on the netw ork. The newly activ e serv er begins to intercept
and queue disk I/O operations for the newly passiv e serv er .
VMware, Inc. 15
Chapter 1 Introduction4 vCenter Serv er Heartbeat causes the newly passiv e serv er to begin accepting updates from the activ e
serv er .
5 vCenter Serv er Heartbeat starts the same protected applications on the new activ e serv er . The protected
applications become accessible to users. The managed switchov er is complete
Automatic Switchover
Automatic switchov er (auto-switchov er) is similar to failov er (discussed in the next section) but is triggered
automatically when system monitoring detects failure of a protected application.
Like managed switchov er , auto-switchov er changes the serv er roles but then stops vCenter Serv er Heartbeat
on the previously activ e serv er to allow the administrator to inv estigate the cause of the auto-switchov er and
v erify the integrity of the data.
After the cause for the auto-switchov er is determined and problems are corrected, the administrator can use
vCenter Serv er Heartbeat Console to return the serv er roles to their original state.
Automatic Failover
Automatic failov er is similar to automatic switchov er (discussed abov e) but is triggered when the passiv e
serv er detects that the activ e serv er is no longer running properly and assumes the role of the activ e serv er .
Figure 1- 7. Automatic Failover
16 VMware, Inc.
Administrator GuideDuring the automatic failov er , the passiv e serv er performs the following steps:
1 Apply any intercepted updates currently in the passiv e serv er ’ s receiv e queue as identified by the log of
update records that are sav ed on the passiv e serv er but not y et applied to the replicated files.
The amount of data in the passiv e serv er ’ s receiv e queue affects the time required to complete the failov er
process. If the passiv e serv er ’ s receiv e queue is long, the system must w ait for all updates to the passiv e
serv er to complete before the rest of the process can take place. An update record can be applied only if
all earlier update records are applied, and the completion status for the update is in the passiv e serv er ’ s
receiv e queue. When no more update records can be applied, any update records that cannot be applied
are discarded.
2 Switch mode of operation from passiv e to activ e.
This enables the public identity of the serv er . The unique Management IP address is remov ed from the
passiv e serv er and the shared Principal (Public) IP address is assigned. The passiv e serv er becomes activ e
and av ailable to clients that w ere connected to the previously activ e serv er before the automatic failov er
and clients are able to reconnect.
3 Start intercepting updates to protected data. Any updates to the protected data are sav ed in the send queue
on the local serv er .
4 Start all protected applications. The applications use the replicated application data to recov er , and then
accept re-connections from any clients. Any updates that the applications make to the protected data are
intercepted and logged.
At this point, the originally activ e serv er is offline and the originally passiv e serv er is filling the activ e role
and running the protected applications. Any updates that completed before the failov er are retained.
Application clients can reconnect to the application and continue running as before.
Managed Failover
Managed failov er is similar to automatic failov er in that the passiv e serv er automatically determines that the
activ e serv er has failed and can w arn the system administrator about the failure; but no failov er actually occurs
until the system administrator manually triggers this operation.
Automatic Switchover and Failover in a W AN Environment
Automatic switchov er and failov er in a W AN environment differ from an automatic switchov er and failov er
in a LAN environment due to the nature of the W AN connection. In a W AN environment, automatic switchov er
and failov er are disabled by default in the ev ent that the W AN connection is lost.
Should a condition arise that w ould normally trigger an automatic switchov er or failov er , the administrator
will receiv e vCenter Serv er Heartbeat alerts. The administrator must manually click the Make Active button
on the Server: Summary page of the vCenter Serv er Heartbeat Console or vSphere Client to allow the roles of
the serv ers to switch ov er the W AN.
VMware, Inc. 17
Chapter 1 Introduction18 VMware, Inc.
Administrator Guide2 Configuring vCenter Server Heartbeat
After installation of vCenter Serv er Heartbeat, initial operational configuration is performed using the vCenter
Serv er Heartbeat Console. The VMw are vCenter Serv er Heartbeat – Server Configuration wizard ( Configur e
Server wizard) sets up and maintains communications betw een the vCenter Serv er Heartbeat serv ers. After
the system is set up and is functioning correctly , y ou do not normally need to reconfigure the system. The
Configure Serv er wizard becomes redundant during daily operations of the softw are.
This chapter includes the following topics:
■ “Launching the Configure Serv er W izard,” on page 20
■ “Configuring the Machine,” on page 21
■ “Configuring the VMw are Channel,” on page 22
■ “Configuring Public IP Addressing,” on page 24
■ “Managing vCenter Serv er Heartbeat License Keys,” on page 26
■ “Configuring the Logs,” on page 27
VMware, Inc. 19Launching the Configure Server W izard
The VMw are vCenter Serv er Heartbeat – Serv er Configuration wizard (Configur e Server wizard) is used to
modify the communications betw een the Primary and Secondary serv ers and reconfigure other components
of vCenter Serv er Heartbeat. When using the Configur e Server wizard, vCenter Serv er Heartbeat must not be
running.
Procedure
1 Stop vCenter Serv er Heartbeat.
2 Click the Configure Server icon on the desktop or navigate to Start > All Programs > VMware > VMware
vCenter Server Heartbeat > Configure Server to launch the Configur e Server wizard.
The Configur e Server wizard is display ed. If vCenter Serv er Heartbeat w as not stopped before launching
the Configur e Server wizard, a message stating that changes made in the Configur e Server wizard will not
be sav ed is display ed.
Figure 2- 1. Configure Server Wizard Introduction
20 VMware, Inc.
Administrator GuideConfiguring the Machine
The Machine tab is used to set the Server Identity, Active Server, and Client Connection Port .
Configuring the Machine Identity
The machine identity is either Primary or Secondary and once assigned does not change during the life of the
serv er .
Caution The machine Identity should only be changed when directed to do so by VMw are Support or when
instructed to do so by a knowledge base article.
Procedure
1 Click the Machine tab and select a Physical Hardwar e Identity for the local machine.
2 Click either Next or Finish.
Figure 2- 2. Configure Server wizard Machine T ab
Configuring the Server Role
Note Before changing the role of the local serv er , v erify that the other (remote) serv er in the pair is not already
performing the same role. vCenter Serv er Heartbeat is designed not to allow tw o passiv e or tw o activ e serv ers
to connect.
Procedure
◆ T o change the serv er role, click the Machine tab, select the serv er that is currently activ e (Primary or
Secondary), and click Next or Finish.
VMware, Inc. 21
Chapter 2 Configuring vCenter Server HeartbeatConfiguring the Client Connection Port
Clients such as the vCenter Serv er Heartbeat Console use the Client Connection Port to connect to vCenter Serv er
Heartbeat. Do not change this port unless another application is using it.
Procedure
◆ T o change the Client Connection Port, click the Machine tab, edit the default entry (52267) and click Next
or Finish .
Configuring the VMware Channel
The Configur e Server wizard Channel tab is used to configure the Channel Routing, Default Channel Port , and
Low Bandwidth Optimization .
Configuring Channel Routing
Channel IP routing defines the IP addresses used to communicate betw een the Primary and Secondary serv ers.
Each link has a pair of addresses, one for the Primary , and one for the Secondary .
T o add an additional VMw are Channel after installing the NICs and configuring them:
Procedure
1 Click the Channel tab. Click Add Row to add the new IP addresses for both the Primary and Secondary
serv er to the VMwar e Channel IP Routing table.
2 Use the drop-down menu to view a list of av ailable IP addresses on the local serv er .
3 T ype the remote serv er IP address.
Figure 2- 3. Configure Server wizard — Channel T ab
22 VMware, Inc.
Administrator Guide4 T o change the VMw are Channel IP addresses, select and edit the entry in the table.
5 Click Next or Finish .
Configuring the Default Channel Port
VMw are Channel uses the Default Channel Port to communicate betw een the Primary and Secondary serv er .
Do not change this port unless another application is using it.
Procedure
◆ T o change the Default Channel Port, click the Channel tab, edit the default entry (57348), and click Next or
Finish.
Configure Low Bandwidth Optimization
Low Bandwidth Optimization is configured automatically during installation based upon the configuration
options selected during Setup. Low Bandwidth Optimization can be configured for: High A v ailability (HA)
when deploy ed as a pair in a LAN or Disaster Recov ery (DR) when deploy ed ov er a W AN.
In a High A v ailability (HA) serv er pair , the queues and buffers are optimized for a high-speed local area
netw ork (LAN) connection, compression is disabled, and automatic failov er betw een serv ers is enabled.
In a Disaster Recov ery (DR) pair , the queues and buffers are optimized for a low-bandwidth wide area netw ork
(W AN) connection, compression may be used, and automatic failov er betw een serv ers is disabled.
In a vCenter Serv er Heartbeat serv er pair y ou can choose the HA or DR topology . How ev er , if y ou hav e
manually configured a non-standard topology , for example, by changing the Auto-Failov er setting, then
Non-Standard will appear in the menu and y ou can choose to leav e the non-standard topology option as it is,
or reset it to one of the standard topologies.
Note The same HA/DR configuration must be set on both serv ers in the pair .
VMware, Inc. 23
Chapter 2 Configuring vCenter Server HeartbeatProcedure
◆ T o change Low Bandwidth Optimization after installation, select the Channel tab of the Configur e Server
wizard and use the HA/DR T opology drop-down to select the appropriate topology . Click Next or Finish
to accept changes.
Figure 2- 4. Configure Server wizard – Channel tab
Configuring Public IP Addressing
vCenter Serv er Heartbeat serv ers are typically configured with one shared Principal (Public) IP address in a
LAN or tw o Principal (Public) IP address in a W AN. These are the addresses used by clients to connect to the
protected application.
During installation, vCenter Serv er Heartbeat configures the Principal (Public) IP address(s) and further
configuration is not normally necessary .
In the ev ent that y ou must add a Management IP address or Principal (Public) IP address for y our environment,
vCenter Serv er Heartbeat allows y ou to add these addresses using the Configure Serv er wizard.
Configuring Principal (Public) IP Addressing
vCenter Serv er Heartbeat`s Configure Serv er wizard is used to add or modify Management IP and Principal
(Public) IP addresses when required by y our environment such as during a migration from a LAN deployment
to W AN deployment.
Procedure
1 Select the Public tab of the Configur e Server wizard and v erify that Non-Identical is selected in the Public
Identity Mode pane.
2 V erify the vCenter Serv er or SQL Serv er Public name in the Name used to connect to vCenter or SQL Server
field.
24 VMware, Inc.
Administrator Guide3 In the NIC field, select the Principal (Public) netw ork connection in the drop-down.
Note Adjacent IP addresses should be reserv ed and used for the Principal (Public) IP address and the
management IP addresses for the Primary and Secondary serv ers when installing vCenter Serv er Heartbeat
on serv ers running W indows Serv er 2008.
4 Enter the Principal (Public) IP address in the Public IP field.
5 Enter the Principal (Public) IP address Subnet Mask in the first Mask field.
6 Enter the reserv ed Management IP address in the Mgmt IP field.
7 Enter the reserv ed Management IP address Subnet Mask in the second Mask field.
Note The Public IP Addresses table allows multiple entries. The table also allows y ou to configure Public
(Public IP)and Management (Mgmt IP) IPs for the same netw ork adapter on separate lines to accommodate
multiple Public IPs in the same or different subnets and multiple Mgmt IPs in the same or different subnets.
The following rules apply to the table:
■ Each row must identify a netw ork adapter
■ Each row must identify either a Public IP/subnet mask or Mgmt IP/subnet mask
■ For each netw ork adapter listed, y ou must hav e at least 1 Public IP/subnet and 1 Mgmt IP/subnet,
how ev er these may be configured on separate rows
8 Click Next or Finish .
Figure 2- 5. Configure Server wizard — Public tab
VMware, Inc. 25
Chapter 2 Configuring vCenter Server HeartbeatManaging vCenter Server Heartbeat License Keys
T o manage vCenter Serv er Heartbeat license keys, select the License tab of the Configur e Server wizard.
Adding an Entry to the License Keys T able
Procedure
1 Click the Add Row icon and enter y our VMw are vCenter Serv er Heartbeat serial number .
2 Manually type or paste (using Ctrl -V) y our license key into the table.
3 Click Next or Finish .
Figure 2- 6. Configure Server wizard — License tab
26 VMware, Inc.
Administrator GuideConfiguring the Logs
vCenter Serv er Heartbeat allows y ou to change the default location for the logs used for storing data in the
queue.
Configuring Message Queue Logs
The serv er temporarily stores replication data receiv ed in the passiv e serv er ’ s receiv e queue and the replication
data w aiting in the activ e serv er ’ s send queue in message queue logs. This configuration only affects the local
serv er . Logs can be in different locations on the Primary and Secondary serv ers.
Procedure
1 Click the Logs tab.
2 Click Browse to navigate to the folder to store the message queue logs.
3 Select the folder and click Next or Finish.
Figure 2- 7. Configure Server Wizard — Logs
Configuring the Maximum Disk Usage
The Configure Serv er wizard allows y ou to configure the maximum disk space allocated for logging.
Log files can increase in size on the activ e serv er under the following conditions:
■ If the activ e serv er cannot communicate with the passiv e serv er
■ Certain operations on the passiv e serv er
■ If the serv er is under heavy load
When the disk reaches quota, replication stops and the system is no longer protected.
VMware, Inc. 27
Chapter 2 Configuring vCenter Server HeartbeatIf using a dedicated disk for log files, consider disabling the quota. T o do this, set the quota to zero. If vCenter
Serv er Heartbeat runs out of disk space, it must be shut down before it can resume replication. Set the quota
with sufficient ov erflow space so vCenter Serv er Heartbeat can stop replicating gracefully .
Procedure
◆ T o configure Maximum Disk Usage , click the Logs tab, type the maximum dedicated disk space allocated
for message queue log files, and click Finish.
28 VMware, Inc.
Administrator Guide3 Server Protection
Protection against operating system or hardw are failure affecting the activ e serv er is facilitated by tw o instances
of the vCenter Serv er Heartbeat that monitor one another by sending “I’ m aliv e” messages and reciprocating
with acknowledgments ov er the VMw are Channel. If the passiv e serv er detects that this process (the heartbeat)
has failed, an automatic switchov er is initiated.
Additionally , vCenter Serv er Heartbeat proactiv ely monitors the capability of the activ e and passiv e serv ers
to communicate with the rest of the netw ork by polling defined nodes around the netw ork, including by
default, the primary DNS serv er , default gatew ay , and the Global Catalog serv er at regular interv als. If all
three nodes fail to respond, for example, due to a netw ork card or local switch failure, vCenter Serv er Heartbeat
can initiate an automatic switchov er , allowing the passiv e serv er to assume the role of the activ e serv er .
This chapter includes the following topics:
■ “Checking the Serv er P air Status,” on page 29
■ “Monitoring the Status of Serv ers,” on page 31
■ “Configuring Heartbeat Settings,” on page 31
Checking the Server Pair Status
The Server: Summary page is the default page that opens when administering a pair of serv ers. The Server:
Summary page allows y ou to view the roles that the serv ers are performing (activ e or passiv e), the actions that
the serv ers are currently performing, and summary information on the status of communications and data
replication betw een serv ers. The low er pane displays status information for each serv er in the pair .
Note T o change the currently display ed serv er , click the serv er graphical representation in the upper pane,
or select the serv er Identity tab (Primary or Secondary Serv er) in the bottom pane.
The following table lists the possible system statuses and their meanings.
T able 3- 1. System Status
Description Icon Status
The Heartbeat service is shut down. Heartbeat service
shutdown
The Heartbeat service is starting up. Initializing
(Normal status) File and registry changes on the activ e serv er are intercepted and
replicated to the passiv e serv er . The protected applications are monitored.
Replicating
VMware, Inc. 29T able 3- 1. System Status
Description Icon Status
File and registry replication is in the process of stopping and all protected
applications may be closing down.
Not replicating
The system is in the process of performing a switchov er . Switching activ e serv er
VMw are Channel connections hav e been established betw een the tw o serv ers. Connecting to peer
serv er
VMw are Channel connections hav e been lost betw een the tw o serv ers. Disconnecting from peer
serv er
File replication is in the process of being stopped and, optionally , all protected
applications may be closing down.
Stopping replication
The replication process is starting and protected applications are optionally starting. Starting replication
The Heartbeat service is initializing on the activ e serv er and starting protected
applications.
Starting as activ e serv er
The Heartbeat service is stopping. The Heartbeat service is shutting down, and will
no longer participate in replication. Optionally , protected applications may be
stopped.
Heartbeat service
shutting down
The passiv e serv er has lost connection to the activ e serv er . If this condition persists
for the failov er timeout, and failov er is permitted betw een the pair of serv ers, then
a failov er will occur .
Lost activ e serv er
A failov er has occurred, Activ e following failov er
The Heartbeat service cannot be contacted on the serv er . Serv er not responding
When viewing the status of the passiv e serv er , the status of the file system and registry are display ed graphically .
The following tables list possible synchronization statuses and their meanings.
T able 3- 2. File System and Registry Status
Description Icon Status
Fully synchronized. Synchronized
There are files that are currently unchecked. A full system check did not complete. Unchecked
Not synchronized. Out-of-Sync
Display ed when the Heartbeat service is starting up or shutting down. Uninitialized
The registry is currently in the process of synchronization. Checking
Not synchronized. Error
When the vCenter Serv er Heartbeat pair establishes a connection, it triggers a file synchronization and
v erification process to ensure all protected files on both serv ers are identical. The process checks each 64K
block of each protected file and performs a checksum to determine whether the blocks differ . If the blocks are
the same, the block is marked as synchronized. If the blocks differ , then the block is replicated to the passiv e
serv er and then marked as synchronized. The file v erification and synchronization process is finished after all
blocks of all stipulated files are marked as synchronized.
30 VMware, Inc.
Administrator GuideMonitoring the Status of Servers
The Server: Monitoring page provides additional information about the status of communications betw een the
serv ers within the pair . The graphical representation provides an ov erview of the status of communications
betw een the serv ers. A green channel icon indicates that the channel is connected and healthy while a y ellow
dashed channel icon indicates that communications are not operational betw een the indicated serv ers. In
addition to the heartbeat sent betw een the serv ers, vCenter Serv er Heartbeat also sends a ping to ensure that
the serv ers remain visible to one another .
Configuring Heartbeat Settings
The Server: Monitoring page provides three configuration features: Configur e Pings, Configur e Failover, and
Configur e Response T imes .
Configuring Pings
IP addresses of all NICs used for the VMw are Channel and configured Management IP addresses are, by
default, automatically added during startup of vCenter Serv er Heartbeat. The Server: Monitoring Ping
Configuration dialog allows y ou to add additional ping targets. This configuration allows vCenter Serv er
Heartbeat to send pings across the VMw are Channel in addition to the heartbeat (“I’ m aliv e” messages) and
to the other serv er in the pair ov er the Public netw ork via the Management IP address to confirm that the
serv er is still operational and providing service.
T o add additional ping targets:
Procedure
1 Click Configure Pings to open the Server Monitoring: Ping Configuration dialog.
2 Select the Ping Settings tab to configure the Ping Interval .
3 Select the Ping Routing tab to add additional ping targets.
Configuring Failover and Active Server Isolation
vCenter Serv er Heartbeat continuously monitors the serv ers in the pair and the netw ork to ensure av ailability
and uses nativ e logic and a combination of elapsed time, administrator configured rules, current serv er netw ork
status, and configured ping routing to determine if failov er or isolation of the activ e serv er is w arranted should
the serv ers experience missed heartbeats.
T o configure failov er:
Note For information on configuring ping routing, see “Configuring Pings,” on page 31 and Configuring
Public Netw ork Connection Checks .
VMware, Inc. 31
Chapter 3 Server ProtectionProcedure
1 Navigate to Server: Monitoring > Configure Failover to open the Server Monitoring: Failover
Configuration dialog.
2 The Failover timeout can be customized by changing the default v alue (60 seconds) to a custom v alue. T ype
a new numeric v alue (seconds) in the Failover timeout text box or use the arrow buttons to configure how
long vCenter Serv er Heartbeat w aits for a missed heartbeat before it takes a pre-configured action to
failov er or isolate the activ e serv er from the netw ork.
3 Select or clear check boxes for the items listed below to select the actions to take if the specified Failover
timeout is exceeded.
When the configured Failover timeout v alue has elapsed, vCenter Serv er Heartbeat will ev aluate, in order ,
the following pre-configured rules before taking action:
■ Failov er from Primary serv er to Secondary serv er if channel heartbeat is lost for failov er timeout
■ Failov er from Secondary serv er to Primary serv er if channel heartbeat is lost for failov er timeout
■ Prev ent failov er or auto switchov er while not synchronized
■ Prev ent Failov er if channel heartbeat is lost but Activ e serv er is still visible to other serv ers
■ Make the serv er passiv e if the Channel and Public netw orks are lost for the configured failov er timeout
Note If a rule is not selected, vCenter Serv er Heartbeat will skip the rule and mov e to the next rule in
the list. After all selected rules hav e been ev aluated vCenter Serv er Heartbeat will take action.
4 Click OK.
Important If either Server: Monitoring Ping Routing or Network Monitoring Ping Routing is misconfigured,
unpredictable behavior can occur .
T ypical Failover and Active Server Isolation Scenarios
The following scenarios assume that vCenter Serv er Heartbeat is deploy ed in a LAN with all rules selected in
the Server: Monitoring > Configure Failover > Failover Configuration dialog.
Failover
The following scenario assumes the activ e serv er has failed and is no longer av ailable.
Upon detection of missed heartbeats, vCenter Serv er Heartbeat on the passiv e serv er performs the following
steps:
1 As soon as the passiv e serv er detects that the VMw are Channel is experiencing missed heartbeats. It will
attempt to ping the activ e serv er`s Management IP address via the Public netw ork using the passiv e serv er`s
NIC configured with the Management IP address. If the ping is successful, the passiv e serv er will v eto
the failov er . If the ping is unsuccessful, it will continue to the next step.
Note Since the passiv e serv er assumes that activ e serv er has failed, the passiv e serv er will not attempt
to v erify synchronization with the activ e serv er .
2 At this point, the passiv e serv er checks the configured v alue of the Failover timeout and starts a “Heartbeat
lost” countdown The passiv e serv er continues with the next step.
32 VMware, Inc.
Administrator Guide3 The passiv e serv er attempts to ping the configured w aypoints on the Public netw ork to determine if the
it is still visible on the Public netw ork. If pings to the w aypoints on the Public netw ork are unsuccessful,
the passiv e serv er will remain passiv e and take no action. If the pings to the w aypoints on the Public
netw ork are successful, the passiv e serv er will continue to the next step.
4 At this point, failov er to the passiv e serv er is postponed until the v alue of the Failover timeout has elapsed.
5 The passiv e serv er changes its role to activ e, remov es the packet filter , and starts all services.
6 As the new activ e serv er , it will begin accepting traffic from clients.
Active Server Isolation
The figure below illustrates a scenario where the activ e serv er has lost connection with the passiv e serv er via
the VMw are Channel.
Figure 3- 1. Network Isolation W orkflow Diagram
VMware, Inc. 33
Chapter 3 Server ProtectionUpon detection of missed heartbeats vCenter Serv er Heartbeat performs the following steps:
1 As soon as the activ e serv er detects that the VMw are Channel is experiencing missed heartbeats, it will
determine if a valid failover tar get (the passive server) is pr esent.
Simultaneously , once the passiv e serv er detects missed heartbeats, it will determine if it is a valid failover
tar get.
2 Next, the activ e serv er will determine if it is synchronized with the failov er target (the passiv e serv er). If
synchronized, it will continue to the next step. If it is not synchronized, it will v eto a failov er .
Simultaneously , the passiv e serv er checks to see if it is synchronized with the activ e serv er . If synchronized,
it will continue to the next step. If it is not synchronized, it will v eto a failov er .
3 At this point, both the activ e and passiv e serv ers check the configured v alue of the Failover timeout and
start a “Heartbeat lost” countdown. Both serv ers should start the countdown at approximately the same
time.
4 Failov er or isolation of the activ e serv er is postponed until the configured Failover timeout v alue (in seconds)
has elapsed and it is during this period that both serv ers accomplish steps 1 & 2.
5 Once the configured Failover timeout period has elapsed, the activ e serv er assumes the VMw are Channel
is lost and will attempt to ping the failov er target (passiv e serv er) via the Public netw ork. If the ping is
successful, activ e serv er isolation is v etoed. If the attempt to ping the failov er target is unsuccessful, the
activ e serv er will proceed to the next step.
Simultaneously , the passiv e serv er assumes the VMw are Channel is lost and attempts to ping the activ e
serv er via the Public netw ork. If the ping is successful, failov er is v etoed. If the ping attempt is unsuccessful,
the passiv e serv er proceeds to the next step.
Note If the serv ers hav e reached this point, then neither serv er can see the other serv er .
6 Both the activ e and passiv e serv ers will check their connectivity to the Public netw ork. If the activ e serv er
has lost connectivity to the Public netw ork, it will isolate itself by making itself passiv e (potential activ e).
Should the activ e serv er reconnect with the passiv e, it will become activ e again. Otherwise, it will remain
passiv e. If the passiv e serv er has lost connectivity to the Public netw ork, it will v eto a failov er .
Configuring Response T imes
vCenter Serv er Heartbeat also allows y ou to configure the time to w ait following a channel connection before
starting replication and the time to w ait following channel disconnection before stopping replication.
Procedure
1 Click Configure Response T imes to open the Server Monitoring: Response T imes dialog.
2 T ype new numeric v alues (seconds) into the text boxes or use the arrow buttons to select new v alues.
3 Click OK.
34 VMware, Inc.
Administrator GuideConfiguring Split-brain A voidance
Split-brain A v oidance ensures that only one serv er becomes activ e if the VMw are Channel connection is lost,
but both serv ers remain connected to the Principal (Public) netw ork. Split-brain A v oidance w orks by pinging
from the passiv e serv er to the activ e serv er across the Principal (Public) netw ork. If the activ e serv er responds,
the passiv e serv er does not failov er , ev en if the VMw are Channel connection is lost. W AN installations require
different IP addresses on the Principal (Public) netw ork for the local and remote serv ers.
Procedure
1 T o enable Split-brain A v oidance, open the Server: Monitoring page in the vCenter Serv er Heartbeat Console.
2 Click Configure Failover.
3 Select Pr event failover if channel heartbeat is lost but Active server is still visible to other servers (r ecommended).
Common Administrative T asks in vCenter Server Heartbeat
The Server: Summary page provides the following buttons that allow y ou to quickly perform common
administrativ e tasks:
■ Make Active — Prompts to v erify that y ou w ant to make the passiv e serv er in the pair activ e. Click Y es.
■ Shutdown — Prompts y ou to select the serv er(s) to shut down. If y ou select the activ e serv er , additional
options to stop or not stop protected applications appear in the dialog. Click OK.
■ Start Replication — Opens the Start Replication Options dialog. Select to start or not start the protected
applications and click OK. By default, all protection modes start when vCenter Serv er Heartbeat starts
and a manual start is not required unless the system stopped in response to an automated stop.
■ Stop Replication — Opens the Stop Replication Options dialog. Use this method to stop replication, such
as to contain a malw are infection or to upgrade a protected application. Select whether to stop or not stop
protected applications and click OK. Replication of data files stops and, if selected, protected applications
also stop.
Note The vCenter Serv er Heartbeat service continues to run on the serv ers, providing heartbeats and protecting
the system and netw ork facets of the activ e serv er .
■ Start Applications — Click to start the protected applications on the activ e serv er .
■ Stop Applications — Click to stop the protected applications on the activ e serv er .
■ Configure — Click to open the Configur e dialog. Select the radio button corresponding to whether y ou
w ant to stop or leav e the protected applications running when vCenter Serv er Heartbeat is shut down.
Y ou can select whether to leav e protected applications running upon shutdown when anet stop command
is issued, and to start protected applications upon startup when a net start command is issued. T ype a
number (seconds) or use the arrow buttons to select an alert threshold v alue for time difference betw een
serv ers, which is checked at handshake following startup. Click OK.
Forcing a Switchover
After configuring vCenter Serv er Heartbeat to protect all the required applications and data, the Secondary
serv er can take ov er from the Primary serv er in a managed and seamless manner called a managed switchov er .
This is particularly useful when maintenance w ork performed on the Primary serv er requires rebooting the
serv er .
VMware, Inc. 35
Chapter 3 Server ProtectionSince a managed switchov er cannot be performed during synchronization, it is important to review the queue
information prior to attempting a managed switchov er . If the queues are large, file operations on the activ e
serv er are high and for this reason it may be prudent to delay a managed switchov er due to the length of time
required to completely clear the queue. Queue lengths can be view ed in the Data: T raffic/Queues page of the
vCenter Serv er Heartbeat Console.
Prior to performing w ork on the Primary serv er , a managed switchov er can be triggered by selecting the
Secondary serv er and clicking Make Active on the Server: Summary page. This changes the serv er roles such
that the activ e serv er becomes passiv e and the passiv e serv er becomes activ e. This means users are able to
w ork continuously while the Primary serv er is off line.
When the Primary serv er is back up and running, the managed switchov er can be triggered again so that the
passiv e serv er becomes activ e and the previously activ e serv er becomes passiv e.
Note The managed switchov er process may be performed at any time as long as the systems are fully
synchronized with respect to data files and registry replication. Switchovers cannot be performed if either server
is in an unsynchr onized or unknown state.
Recovering From a Failover
A failov er differs from a switchov er . A switchov er is a controlled switch (initiated manually from the vCenter
Serv er Heartbeat Console, or initiated by vCenter Serv er Heartbeat when preconfigured) betw een the activ e
and passiv e serv ers. A failov er happens when any of the

Leave a Comment