Beruflich Dokumente
Kultur Dokumente
100-002685-E
COURSE DEVELOPERS
Bilge Gerrits
Steve Hoffer
Siobhan Seeger
Pete Toemmes
Graeme Gofton
Sean Nockles
Brad Willer
TECHNICAL
CONTRIBUTORS AND
REVIEWERS
Geoff Bergren
Kelli Cameron
Tomer Gurantz
Anthony Herr
James Kenney
Gene Henriksen
Bob Lucas
Paul Johnston
Rod Pixley
Clifford Barcliff
Danny Yonkers
Antonio Antonucci
Satoko Saito
Feng Liu
Table of Contents
Course Introduction
Lesson 1: Service Group Dependencies
Common application relationships ............................................................... 1-3
Service group dependencies........................................................................ 1-7
Service group dependency examples .......................................................... 1-9
Configuring service group dependencies ................................................... 1-12
Alternative methods of controlling interactions........................................... 1-16
Lesson 2: Reconfiguring Cluster Membership
Adding a system to a cluster ........................................................................ 2-3
Merging clusters ........................................................................................... 2-6
Additional reconfiguration tasks ................................................................. 2-10
Lesson 3: Startup and Failover Policies
Startup rules and policies ............................................................................. 3-3
Failover rules and policies.......................................................................... 3-10
Limits and Prerequisites ............................................................................. 3-18
Modeling startup and failover policies ........................................................ 3-22
Lesson 4: Alternate Network Configurations
Multiple service groups with NIC resources ................................................. 4-3
Multiple public interfaces .............................................................................. 4-8
Appendix A: Labs
Lab 1: Service group dependencies............................................................ A-3
Lab 2: Merging clusters .............................................................................. A-13
Lab 3: Failover policies............................................................................... A-25
Lab 4: Creating a parallel network service group ....................................... A-37
Appendix B: Lab Solutions
Lab 1: Service group dependencies............................................................ B-3
Lab 2: Merging clusters .............................................................................. B-37
Lab 3: Failover policies............................................................................... B-65
Lab 4: Creating a parallel network service group ....................................... B-89
Appendix C: Supplemental Content
Service group dependenciesFailover process......................................... C-2
Table of Contents
i
Copyright 2012 Symantec Corporation. All rights reserved.
ii
Course Introduction
The Veritas Cluster Server for UNIX curriculum is a series of courses that are
designed to provide a full range of expertise with Veritas Cluster Server (VCS)
high availability solutionsfrom design through implementation.
Veritas Cluster Server for UNIX: Install and Configure
This course covers installation and configuration of common VCS
environments, focusing on two-node clusters running application and database
services.
Veritas Cluster Server for UNIX: Manage and Administer
This course focuses on multinode VCS clusters and advanced topics related to
managing more complex cluster configurations.
The course includes two participant guides:
Veritas Cluster Server for UNIX: Example Application Configurations
This guide provides examples of common service group configurations.
The instructor may opt to present some or all of this material in class,
depending on time constraints and student interest.
Veritas Cluster Server for UNIX: Cluster Management
This guide provides detailed information about configuring and managing
more complex clusters.
Veritas Cluster Server eLearning Library
The eLearning Library is available with bundled training options and includes
content on advanced high availability and disaster recovery features.
12
Course Introduction
13
Copyright 2012 Symantec Corporation. All rights reserved.
Course overview
14
Lesson 1
10
12
11
The inherent relationships require that the database is started and running first,
then the application, and finally the Web server.
In a cluster environment, you must also consider which systems should run which
applications, as shown in the following examples.
13
Furthermore, the application must come online first, followed by the Web server. If
the application faults and fails over to another cluster node, the Web server must
then be brought online on that same failover target system.
12
14
In this scenario, the database must come online first, and then the application is
started only after the database is running. If the application faults and fails over to
another cluster node, the database can stay online on the original system, as long as
the application is not restarted on that same system.
13
15
The difference between this example and the previous example, is that in this case,
neither application requires the other to be online. The only requirement is that the
other application cannot be online on a system when one application is brought
online.
14
16
15
There are four basic criteria for defining how services interact when using service
group dependencies.
A service group can require another group to be online or offline in order to
start and run.
You can determine the startup order for service groups by designating one
group the child and another a parent. Parent groups depend on child groups. If
service group B requires service group A to be online in order to start, then B is
the parent and A is the child.
You can specify where the groups must be online or offline.
For all online dependencies, the child group must be online in order for the
parent to start. A location of local, global, or remote determines where the
parent can come online relative to where the child is online.
For offline local, there is no requirement for either service group to start
first. The child group must be offline for the parent to come online.
Failover behavior of linked service groups is specified by designating the
relationship soft, firm, or hard.
Detailed information about how these dependencies impact service group
operations is provided in Veritas Cluster Server Administrators Guide.
17
Failover types
The dependency type determines failover behavior for linked service groups.
Soft dependency
A soft dependency means that VCS does not immediately take the parent offline if
the child group faults.
When both groups are online, the child group can be taken offline while the parent
is online. Likewise, the parent group can be taken offline while the child is online.
The parent remains online if the child group faults and cannot fail over. The child
remains online if the parent group faults.
Firm dependency
Copyright 2012 Symantec Corporation. All rights reserved.
A firm dependency means that VCS imposes additional constraints on the parent
and child service groups. Specifically:
The parent group must be taken offline when the child group faults.
When the child is brought online on another system, the parent group is
brought online on a system determined by the location type (local, global,
remote).
If the parent group faults, the child continues to run.
16
Hard dependency
A hard dependency means that VCS takes the other service group offline when
either the child or parent group faults.
18
17
An online local firm is similar to soft, except the parent service group is taken
offline when the child faults. If the child fails over and is restarted on another
system, the parent is then also started on that system
Online local hard
In an online local hard dependency, the child service group is taken offline if the
parent service group faults. Online local is the only dependency supporting the
hard type.
19
An online remote firm is similar to soft, except the parent service group is taken
offline when the child faults.
18
110
In an offline local dependency, the parent service group can be started only if the
child service group is offline on the local system. Similarly, the child can only be
started if the parent is offline on the local system. This prevents conflicting
applications from running on the same system. This is commonly used when you
have a test version of an application running on the failover target system for the
production version of the application.
19
111
20
112
Linking constraints
21
To link a parent and child group with soft dependency, it is not required that the
child group be online if the parent is online. However, if the child group is also
online, the parent and child may not be linked in such a way that their online states
conflict with the type of link between the parent and child.
To link a parent and child group with firm dependency, the parent group must be
offline, or the parent and child group must be online in such a way that their online
states do not conflict with the type of link between the parent and child.
Removing service group dependencies
When removing a dependency, you do not need to specify the type of dependency,
because only one dependency is allowed between two service groups.
113
You can use the Simulator on Windows to model how different types of links
affect service group behavior. This enables you to fully understand the
implications and the effects of different dependency configurations before you
configure links in your production environment.
22
114
Note that propagation is supported only with local service group dependencies
type.
23
115
VCS provides several event triggers that can be used to enforce service group
relationships where dependencies cannot be configured, as described in the
previous example. These triggers are useful for managing relationships between
service groups:
VCS runs the preonline script before bringing a service group online.
The PreOnline trigger must be enabled for each applicable service group by
setting the PreOnline service group attribute. For example, to enable the
preonline trigger for groupa, type:
hagrp -modify GroupB PreOnline 1
The preonline script must start the service group with the -nopre option
to hagrp.
The postonline script is run after a service group is brought online.
The postoffline script is run after a service group is taken offline.
24
25
117
In the example shown in the slide, the database cluster is the lowest level tier and is
the most critical component in this service. Typically, you would find the database
to be shared across multiple services as well. If something happens in the database
tier, for example, an active server fails or the database runs into software issues on
one of the servers, VCS ensures the database is failed over to another node in the
cluster. Additionally, disaster recovery may be configured so that if the production
site goes down, VCS fails the database over to the DR site.
26
118
In other words, the parent service group with a restart dependency on a child
service group takes no action when a child service group faults until the child is
restarted.
27
This type of dependency is used when only start or stop ordering is required, and
no fault policy action is needed.
119
28
Labs and solutions for this lesson are located on the following pages.
Lab 1: Service group dependencies, page A-3.
Lab 1: Service group dependencies, page B-3.
120
Lesson 2
29
30
22
Assumptions
31
For illustration purposes, these assumptions are used in describing how to perform
this task:
The VCS cluster consists of two or more systems, all of which are up and
running.
There are multiple service groups configured in the cluster. All of the service
groups are online somewhere in the cluster.
The new system to be added to the cluster does not have any VCS software.
The new system has the same version of operating system and VERITAS
Storage Foundation as the systems in the cluster.
The new system may not have all the required application software.
The storage devices can be connected to all systems.
23
32
24
33
25
Merging clusters
Merging two running VCS clusters
The objective of this task is to merge two running VCS clusters with no or minimal
impact on application services. Also, ensure that the cluster configuration is
modified so that the application services can make use of the systems from both
clusters.
Assumptions
The following is a list of assumptions that you need to take into account while
planning a procedure for this task:
All the systems in both clusters are up and running.
There are multiple service groups configured in both clusters. All service
groups are online somewhere in the cluster.
All the systems have the same version of operating system and Veritas Storage
Foundation.
The clusters do not necessarily have the same application services software.
New application software can be installed on the systems to support
application services of the other cluster.
The storage devices can be connected to all systems.
The cluster interconnects of both clusters are isolated before the merger.
34
For this example, assume that two two-node clusters are merged into a single fournode cluster.
26
35
27
36
28
37
29
The following is a list of assumptions that you need to take into account while
planning a procedure for this task:
Application resources have been prepared
Service group definitions have been merged
Agents have been installed
38
210
39
211
40
212
41
Labs and solutions for this lesson are located on the following pages.
Lab 2: Merging clusters, page A-13.
Lab 2: Merging clusters, page B-37.
213
42
214
Lesson 3
43
44
32
45
33
46
34
47
where possible values for policy are Order, Priority, and Load. You can
also set this attribute using Veritas Operations Manager or the Cluster Manager
Java GUI.
Note: The configuration must be open to change service group attributes.
35
AutoStartPolicy=Order
When the AutoStartPolicy attribute of a service group is set to the default value of
Order, the first system available in AutoStartList is selected to bring the service
group online. The priority numbers in SystemList are ignored.
In the example shown on the slide, the A service group is brought online on s1,
although it is the system with the highest priority number in SystemList. Similarly,
the B service group is brought online on s2, and the C service group is brought
online on s3 because these are the first systems listed in the AutoStartList
attributes of the corresponding service groups.
Note: Because Order is the default value for the AutoStartPolicy attribute, it is
not listed in the service group definitions in the main.cf file.
48
36
AutoStartPolicy=Priority
When the AutoStartPolicy attribute of a service group is set to Priority, the system
with the lowest priority number in the SystemList that also appears in the
AutoStartList is selected as the target system during start-up. In this case, the order
of systems in the AutoStartList is ignored.
The same example service groups are now modified to use the Priority
AutoStartPolicy, as shown on the slide. In this example, the A service group is
brought online on s3, which has the lowest priority number in SystemList even
though it is listed as the last system in AutoStartList. Similarly, the B service group
is brought online on s1 (with priority number 0), and the C service group is
brought online on s1 (with priority number 1).
49
Notice how the startup systems have changed for the service groups by changing
the AutoStartPolicy attribute, although the SystemList and AutoStartList attributes
are the same for these two examples.
37
AutoStartPolicy=Load
When AutoStartPolicy is set to Load, VCS determines the target system based on
the existing workload of each system listed in the AutoStartList attribute and the
load that is added by the service group.
These attributes control load-based startup:
Capacity is a user-defined system attribute that contains a value representing
the total amount of load that the system can handle.
Load is a user-defined service group attribute that defines the amount of
capacity required to run the service group.
AvailableCapacity is a system attribute maintained by VCS that quantifies the
remaining available system load.
Copyright 2012 Symantec Corporation. All rights reserved.
In the example displayed on the slide, three servers have Capacity set to 300, 200,
and 100, respectively. The s2 system is selected as the target system for starting C
because it has the highest AvailableCapacity value of 200 after A and B are started
on s1.
50
38
When a service group is brought online, the value of its Load attribute is subtracted
from the system Capacity value, and AvailableCapacity is updated to reflect the
difference. Both the Capacity attribute of a system and the Load attribute of a
service group are static user-defined attributes based on your design criteria.
51
39
52
310
53
311
Failover policies
VCS supports a variety of policies that determine how a system is selected when
service groups must migrate due to faults. The policy is configured by setting the
FailOverPolicy attribute to one of these values:
Priority: The system with the lowest priority number is preferred for failover
(default).
RoundRobin: The system with the least number of active service groups is
selected for failover.
Load: The system with the highest value of the AvailableCapacity system
attribute is selected for failover.
54
312
FailOverPolicy=Priority
When FailOverPolicy is set to Priority, VCS selects the system with the lowest
assigned value from the SystemList attribute.
For example, the C service group has three systems configured in the SystemList
attribute and the same order for AutoStartList values:
SystemList = {s3=0, s1=1, s2=2}
AutoStartList = {s3, s1, s2}
55
Priority policy is the default behavior and is ideal for simple two-node clusters or
small clusters with few service groups.
313
FailOverPolicy=RoundRobin
The RoundRobin policy selects the system running the fewest service groups as
the failover target.
The RoundRobin policy is ideal for large clusters running many service groups
with essentially the same server load characteristics (for example, similar
databases or applications).
56
Ties are determined by the order of systems in the SystemList attribute. For
example, if two failover target systems have the same number of service groups
running, the system listed first in the SystemList attribute is selected for failover.
314
FailOverPolicy=Load
When FailOverPolicy is set to Load, VCS determines the target system based on
the existing workload of each system listed in the SystemList attribute and the load
that is added by the service group.
57
In the example displayed in the slide, the three servers that remain running after
the s3 system fails have Capacity set to 300, 200, and 100. Each service group has
a fixed load defined by the user, which is subtracted from the system capacity to
find the AvailableCapacity value of a system.
When failover occurs, VCS checks the value of AvailableCapacity on each
potential targeteach system in the SystemList attribute for the service group
and starts the service group on the system with the highest value.
Note: In the event that no system has a high enough AvailableCapacity value for a
service group load, the service group still fails over to the system with the highest
value for AvailableCapacity, even if the resulting AvailableCapacity value is zero
or a negative number.
315
To set Load from the CLI, use the hagrp -modify command as shown in this
example:
58
316
You can configure the loadwarning trigger to provide notification that a system
has sustained a predetermined load level for a specified period of time.
59
317
60
Limits
The Limits system attribute is used to define the resources and the corresponding
capacity of each system for that resource. You can use any keyword for a resource
as long as you use the same keyword on all systems and service groups.
318
CurrentLimits
CurrentLimits is an attribute maintained by VCS that contains the value of the
remaining available resources for a system. For example, if the limit for DBs is 2
and the A service group is online with a DBs prerequisite of 1, the CurrentLimits
setting for DBs is 1:
CurrentLimits = { DBs=1 }
61
Note: A value of 0 is assumed for systems that do not have some or all of the
resources defined in their Limits attribute. Similarly, a value of 0 is
assumed for service groups that do not have some or all of the resources
defined in their Prerequisites attribute.
319
Failover example
The configuration in the slide shows how a failover target system is selected for
the C service group when the s3 system faults. Because C has a prerequisite of 1
for DBs and there are no systems with a high enough CurrentLimits value for DBs
to support running the group, C stays offline.
System Limits are hard values, meaning that if a system does not meet the
requirements specified in the Prerequisites attribute for a service group, the service
group cannot be started on that system.
Contrast this with the Load and Capacity, which are soft limits. If the
AvailableCapacity value is not high enough to satisfy the Load requirement for a
service group, VCS still fails over the service group, even if the AvailableCapacity
becomes a negative value.
Copyright 2012 Symantec Corporation. All rights reserved.
Therefore, using Limits and Prerequisites may be a better method for controlling
service group startup and failover in cases where hard limits must be enforced. An
example case is an environment where licensing restrictions prevent two instances
of an application or database from running on a system.
62
320
To set Prerequisites from the CLI, use the hagrp -modify command as shown
in this example:
63
Notes:
To be able to set these attributes, open the VCS configuration to enable read/
write mode and ensure that the service groups that are already online on a
system do not violate the restrictions.
The order that the resources are defined within the Limits or Prerequisites
attributes is not important.
321
The VCS Simulator for Windows is a good tool for modeling the behavior that you
require before making changes to the running configuration. This enables you to
fully understand the implications and the effects of different workload
management configurations.
64
322
65
Labs and solutions for this lesson are located on the following pages.
Lab 3: Failover policies, page A-25.
Lab 3: Failover policies, page B-65.
Lesson 3 Startup and Failover Policies
Copyright 2012 Symantec Corporation. All rights reserved.
323
66
324
Lesson 4
67
68
42
In addition to the overhead of many monitor cycles for the same resource, a
disadvantage of this configuration is the effect of changes in NIC hardware. If you
must change the network interface (for example, in the event the interface fails),
you must change the Device attribute for each NIC resource monitoring that
interface.
69
43
70
44
A parallel service group is managed like any other service group in VCS, except
that it cannot be switched. The group is only started on a system listed in the
AutoStartList and the SystemList attributes. A parallel service group can also fail
over, in effect, if the service group faults on a system and there is an available
system (listed in the SystemList attribute) that is not already running the service
group. This is accomplished by VCS starting up the parallel group on the target
system.
71
To create a new parallel service group in a running cluster, set the Parallel attribute
to 1 (true) and then add resources.
You cannot change an existing failover service group that contains resources to a
parallel service group except by using the offline configuration procedure to edit
the main.cf file and add Parallel = 1 to the service group definition.
45
The example network service group also requires a Phantom resource in order to
ensure that the service group status is displayed properly. A service group shows
an online status only when all of its nonpersistent resources are online. Therefore,
if a service group has only persistent resources, VCS considers the group offline,
even if the persistent resources are running properly. When a Phantom resource is
added, the status of the service group is shown as online.
72
46
73
47
For example, on Linux, you can use the bond driver to treat multiple network
interfaces as one logical interface. In this case, you can use the VCS NIC agent to
monitor a bond-type interface and an IP resource to bring up a virtual IP address
on the virtual bond interface.
74
48
The following table shows an example configuration where eth0 and eth1 are
controlled by the bond0 interface.
ifcfg-eth0
ifcfg-eth1
ifcfg-bond0
DEVICE=eth0
ONBOOT=yes
BOOTPROTO=no
ne
USERCTL=no
PEERDNS=no
TYPE=Ethernet
MASTER=bond0
SLAVE=yes
DEVICE=eth1
ONBOOT=yes
BOOTPROTO=no
ne
USERCTL=no
PEERDNS=no
TYPE=Ethernet
MASTER=bond0
SLAVE=yes
DEVICE=bond0
ONBOOT=yes
USERCTL=no
TYPE=Ethernet
MTU=""
NETMASK=255.255.0.0
BOOTPROTO=none
IPADDR=10.10.2.20
BROADCAST=10.10.255.255
GATEWAY=10.10.2.1
NETWORK=10.10.2.0
NIC bond0nic (
75
Device = bond0
)
An advantage of using the Linux bond interface is that you can aggregate
interfaces to increase bandwidth.
49
Local interface failover can drastically reduce service interruptions to the clients.
Some applications have time-consuming shutdown and startup processes that
result in substantial downtime when the application fails over from one system to
another.
76
Failover between local interfaces can be completely transparent to users for some
applications.
Using multiple networks also makes it possible to eliminate any switch or hub
failures causing service group failover as long as the multiple interfaces on the
system are connected to separate hubs or switches.
410
The IPMultiNIC and MultiNICA resources provide essentially the same service as
the IP and NIC resources, but these resources monitor multiple interfaces instead
of a single interface. The dependency between these resources is the same as the
dependency between IP and NIC resources.
77
411
78
412
79
Labs and solutions for this lesson are located on the following pages.
Lab 4: Creating a parallel network service group, page A-37.
Lab 4: Creating a parallel network service group, page B-89.
413
80
414
Lesson 5
81
82
52
83
You can also use Veritas Operations Manager to visualize the state of multi-tier
applications and all subcomponents components and to start or stop the entire
logical application in an ordered fashion. Virtual Business Services provides this
capability of associating service groups across clusters and managing the
relationships in a high availability and disaster recovery environment.
53
Graphics and color coding are both used to draw attention to resources that require
attention because they are either faulted (not functioning) or are at risk (configured
to be fault-tolerant, but having one or more failed components that could result in
faulting if a further component failure occurs).
84
54
In the example shown in the slide, a new user account, OraOper, is given service
group operator privileges for all clusters containing Oracle service groups.
85
55
You can use VOM license deployment policies and reports to ensure all systems in
the data center meet licensing requirements.
86
56
VOM solutions
Veritas Operations Manager solutions are independent and optional feature packs
that you can download and use to enhance the functionality of Veritas Operations
Manager. These solutions are grouped into the following categories:
Add-on
Package
Patch
Hotfix
87
A core group of add-ons are bundled in VOM solutions repository. Other add-ons
must be downloaded from sort.symantec.com and then uploaded to your
VOM repository. Any solutions available in the repository can be installed on the
Management Server, managed hosts, or on both. To deploy solutions you must
have domain administrative privileges.
VOM Deployment Management enables you to install a solution on selected hosts,
or a selected platform (for example AIX). When you run the installation process, a
deployment request is sent. You can view that deployment request in the
Deployment Requests page.
After a solution is installed, you must enable the solution to use the management
tools within the add-on.
57
When you bring a fire drill service group online, the resources:
1 Create a snapshot of the replicated volume.
2 Mount the snapshot file system.
3 Start the application or database.
88
VCS logs any errors to enable you to identify configuration problems in your DR
site, as shown in the following examples. Starting with VCS 5.0, EMC SRDF and
Hitachi Truecopy agents are enhanced to take advantage of device tagging in
Volume Manager. This enables hardware snapshots to be imported and used for a
fire drill capability without any scripting or tasks required to be used outside of
VCS.
Note: The Percentage in slide based on Symantec Disaster Recovery Study
published 22 November 2010.
58
You can configure the VCS Management Console to run regularly scheduled fire
drills to ensure you are continuously validating your DR environment.
89
59
90
510
Other examples of configuration drift that cause problems for site migration
include:
Expired licenses
Operating system and application patch version mismatches
Configuration files not synchronized
91
511
Managing replication
You can use VOM to simplify configuration of storage and replication objects used
in replicated data clusters (RDCs) and global clusters with both 4.x- and 5.x-based
systems.
To generate reports about the details of data transfer between source and target
replication hosts, download the Add-on for Veritas Volume Replicator Bandwidth
Reporting from sort.symantec.com.
After you upload the add-on to the VOM repository, you can deploy the solution to
the management server and all hosts with replication solutions.
92
512
93
513
94
514
Virtualization support
95
Most customers are using more than one virtualization methodology within their
IT infrastructure. Each major UNIX operating system vendor has a virtualization
technology, each with unique terms such as control domain, host, guest, and virtual
I/O. The slide illustrates four virtualization approaches and examples of platforms
that utilize that type of virtualization.
Hardware partitioning subdivides a server, allocating resources to partitions,
each running an operating system. Sun Domains and HP nPartitions are
example solutions using this technology.
Bare metal hypervisor consists of a hypervisor process running directly on the
server hardware, which enables creating and maintaining virtual machines.
VMware ESX server is an example of this technology.
Hosted hypervisor has hosted operating system running on the server and a
hypervisor running within the OS. I/O flows from the virtual machines,
through the hypervisor, to the local operating system drivers, and then to the
server hardware. Linux KVM and Solaris LDOM are examples using this
technology.
Finally, operating system container technology allows for the partitioning of
the OS into containers to control access to the applications running within the
containers.
515
The slide shows how Storage Foundation and VCS works in each of the
virtualization environments.
Hardware partitioning
Storage Foundation and VCS run within partitions and generally partitions are
not migrated between physical machines.
Bare metal hypervisor
Storage Foundation and VCS run within the virtual machines. The VMs may
be migrated to different physical servers by virtualization HA technology.
Hosted hypervisor
Storage Foundation and VCS can run on both the host operating system and
virtual machines. VCS can be used to make VMs highly available, however
clusters cannot contain both VMs and the physical machine hosting VMs.
Operating system containers
Storage Foundation and VCS can run on the host operating system and can
control the startup of the security container as well as the applications within
them.
96
516
Veritas Operations Manager can be used to manage storage and clustering in both
physical and virtual environments.
97
517
98
518
Summary
This lesson summarized some of the features of Veritas Cluster Server, Storage
Foundation, and Veritas Operations Manager that specifically enhance high
availability and disaster recovery solutions in enterprise data centers.
99
519
100 520
Appendix A
Labs
101
102 A2
103
A3
Verify that the following virtual machines shown are powered on, that you are
logged using the indicated account and that the indicated terminal windows are
opened.
mgt
There is no need to log into this virtual machine at this time.
sym3
Log in with credentials:
Account: root
Password: veritas
Two terminal windows should be active on this server.
sym4
Log in with credentials:
Account: root
Password: veritas
Two terminal windows should be active on this server.
104 A4
If you have machines running that are not used in this lab, shut down the
operating system and power the machines off.
Note: If you are completing the lab exercises in order, this will require you to
shutdown cps, sym1 and sym2, and power on and log into sym3 and
sym4.
sym3
Note: The appsg and testappsg service groups were added in an early Install and
Configure lab using the Veritas Cluster Manager Java Console. If those
lab exercises were skipped, then consult with your instructor on how to
copy in a replacement main.cf file located at:
/student/labs/vcs/vcs60/maincf/IC/
main.cf.east.lab05
1
105
Display the state of all resources by listing the resources for each service group
and the dependencies between each resource in each service group.
Determine if all resources in all service groups other than the ClusterService
service group are set to critical. If any resources are set to non critical then
modify them. Open, save and close the cluster configuration as appropriate.
A5
sym3
Note: The service groups in this exercise will come online and go offline very
quickly. The hagrp -wait commands, while not necessary, have been
included to be consistent with other lab exercises.
1
Create and confirm dependencies between the appsg and dbsg and between
the websg and dbsg service groups as indicated below. Save but do not close
the VCS configuration.
106 A6
Use a single hagrp command with the propagate option to bring the dbsg,
appsg, and websg service groups online on sym3. Can all three service groups
be brought online using just one command? If not, then ensure that all three
service groups are brought online on sym3.
Use a single hagrp command to take the dbsg, appsg, and websg service
groups offline on sym3. Can all three service groups be brought offline using
just one command? If not, then ensure that all three service groups are brought
offline on sym3.
Using a single hagrp command with the propagate option, bring the dbsg
and websg service groups online on sym3. Attempt to bring the appsg service
group online on sym4. Were you successful? If not successful, then bring the
appsg service group online on sym3.
Clear the appip resource fault and bring the appsg service group back online
on sym3.
Fault the dbsg service group by removing the /var/tmp/dbip file followed
by a probe of the dbip resource on sym3. Do any service groups failover?
Clear the dbip resource fault and take the dbsg, websg, appsg, and testappsg
service groups offline on any system they may be online.
10 Link the testappsg service group as a parent of the appsg service group with
Copyright 2012 Symantec Corporation. All rights reserved.
an offline local dependency. Save, but do not close the VCS configuration.
107
Note: This will configure a preference for the appsg service group running on
a system over the testappsg service group and force the testappsg to
switch upon a successful failover of the appsg service group.
A7
11 Bring the dgsg, appsg, and websg service groups online on sym3. Attempt to
bring the testappsg service group online on sym3. Were you successful? If
not, attempt to bring the testappsg service group online on sym4.
12 Fault the dbsg service group by removing the /var/tmp/dbip file followed
13 Take the testappsg service group offline on sym3 and clear the dbip resource
fault. Then, use a single command to take the dbsg, websg, appsg, and
testappsg service groups offline on any system where they may be online.
14 Unlink the service group dependencies between the testappsg and appsg
service groups and between the dbsg, websg and appsg service groups. Save,
but do not close the VCS configuration.
108 A8
sym3
Note: The service groups in this exercise will come online and go offline very
quickly. The hagrp -wait commands, while not necessary, have been
included to be consistent with other lab exercises.
109
A9
110
A10
Using a single hagrp command with the propagate option, take the dbsg,
appsg, and websg service groups offline on sym3. Can all three service
groups be brought offline using just one command? If not, then ensure that all
three service groups are brought offline on sym3.
Using a single hagrp command with the propagate option, bring the dbsg,
appsg, and websg service groups online on sym3.
Clear the appip resource fault and use a single hagrp command with the
propagate option to bring the appsg service group back online on sym3.
Fault the dbsg service group by removing the /var/tmp/dbip file followed
by a probe of the dbip resource on sym3. Do any service groups failover?
Take the testappsg service group offline on sym3 and clear the dbip resource
fault. Then, use a single command to bring the dbsg, websg, appsg, and
testappsg service groups offline on any system they may be online.
Unlink the service group dependencies between the testappsg and appsg
service groups, between the dbsg and appsg service groups and between the
websg and appsg service groups. Save, but do not close the VCS
configuration.
sym3
Note: The service groups in this exercise will come online and go offline very
quickly. The hagrp -wait commands, while not necessary, have been
included to be consistent with other lab exercises.
111
Bring the testappsg service group online on sym4. Using a single hagrp
command with the propagate option, bring the dbsg, appsg, and websg
service groups online on sym3. Were you able to bring the websg, dbsg, and
appsg service groups online with the online propagate option?
Using a single hagrp command with the propagate option, take the websg,
dbsg, and appsg and service groups offline on sym3. Can all three service
groups be brought offline using just one command? If not, then ensure that all
three service groups are brought offline on sym3.
A11
Using a single hagrp command with the propagate option, bring the dbsg,
appsg, and websg service groups online on sym3.
Clear the webvip resource fault and bring the websg service group back online
on sym3.
Fault the dbsg service group by removing the /var/tmp/dbip file followed
by a probe of the dbip resource on sym3. Do any service groups failover?
Take the appsg service group offline on sym3. Then, bring the websg, appsg,
and dbsg service groups online on sym4. Account for the testappsg service
group.
10 Unlink the service group dependencies between the testappsg and appsg
service groups, between the dbsg and appsg service groups, and between the
websg and appsg service groups. Save and close the VCS configuration.
112
End of lab
A12
113
A13
Copyright 2012 Symantec Corporation. All rights reserved.
mgt
There is no need to log into this virtual machine at this time.
sym1
Log in with credentials:
Account: root
Password: veritas
Two terminal windows should be active on this server.
sym2
Log in with credentials:
Account: root
Password: veritas
Two terminal windows should be active on this server.
sym3
Log in with credentials:
Account: root
Password: veritas
Two terminal windows should be active on this server.
sym4
Log in with credentials:
Account: root
Password: veritas
Two terminal windows should be active on this server.
114
Verify that the following virtual machines shown are powered on, that you are
logged using the indicated account and that the indicated terminal windows are
opened.
A14
If you have machines running that are not used in this lab, shut down the
operating system and power the machines off.
Note: If you are completing the lab exercises in order, this will require you to
power on and log into sym1 and sym2.
115
A15
Copyright 2012 Symantec Corporation. All rights reserved.
sym1
Note: In this lab there are no iSCSI LUNs that are visible to all four systems, so
merging enabled I/O fencing using disks cannot be part of the cluster
merge lab. Therefore, for lab purposes, I/O fencing is re-configured to
disabled mode on both clusters.
sym3
116
A16
Note: In this lab, there is a conflicting duplicate named appsg service group in
both the west and east clusters. This would cause errors during a merge.
The appsg service group in the west cluster would have precedence so the
appsg service group in the east cluster must be renamed or, as performed
in this lab, deleted from the configuration.
5
Take the appsg service group offline and list its resources and dependencies.
Open the VCS configuration for update and delete each resource in the appsg
service group in resource dependency order. Then, delete the appsg service
group and save and close the VCS configuration.
Edit the eastmain.cmd file to remove all commands other than those
related to the websg, testappsg, dbsg service groups, their resources, and their
resource dependencies.
117
After you have made the edits the contents of the file should be:
hagrp -add dbsg
hagrp -modify dbsg SystemList sym3 0 sym4 1
hagrp -modify dbsg AutoStartList sym3 sym4
hagrp -modify dbsg SourceFile "./main.cf"
hares -add dbdg FileOnOff dbsg
hares -modify dbdg PathName "/var/tmp/dbdg"
hares -modify dbdg Enabled 1
hares -add dbip FileOnOff dbsg
hares -modify dbip PathName "/var/tmp/dbip"
hares -modify dbip Enabled 1
hares -add dblistener FileOnOff dbsg
hares -modify dblistener PathName "/var/tmp/
dblistener"
hares -modify dblistener Enabled 1
hares -add dbmnt FileOnOff dbsg
hares -modify dbmnt PathName "/var/tmp/dbmnt"
A17
Copyright 2012 Symantec Corporation. All rights reserved.
118
A18
119
A19
Copyright 2012 Symantec Corporation. All rights reserved.
Shutdown VCS, IO fencing, GAB, and LLT on both nodes of the east cluster.
Note: In order for the merge utility to work correctly, the main.cf files on
the east cluster nodes must not exist.
120 A20
sym1
1
Log file:
121
Perform a summary status of the cluster noting that none of the former east
cluster service groups have been added.
Use the lltstat command to check the LLT configured and active status.
Optionally, performed this step on all cluster nodes.
A21
Copyright 2012 Symantec Corporation. All rights reserved.
122 A22
sym3
1
123
A23
Copyright 2012 Symantec Corporation. All rights reserved.
sym3
1
Note: These service groups could be online on sym1 and sym2 and could
be modified to accommodate that.
End of lab
124 A24
125
A25
Copyright 2012 Symantec Corporation. All rights reserved.
Verify that the following virtual machines shown are powered on, that you are
logged using the indicated account and that the indicated terminal windows are
opened.
mgt
There is no need to log into this virtual machine at this time.
winclient
Log in with credentials:
Account: Administrator
Password: veritas
Note: The terminal windows are referred to as hostname:terminal#
throughout the labs. For example: sym1:terminal1, sym1:terminal2,
sym2:terminal1, and sym2:terminal2. It is not necessary to label the
terminal windows and you may decide which is terminal1 and which is
terminal2.
126 A26
If you have machines running that are not used in this lab, shut down the
operating system and power the machines off.
Note: If you are completing the lab exercises in order, this will require you to
shutdown sym1, sym2, sym3 and sym4, and power on and log into
winclient.
winclient
1
Without making any changes, review the contents of the main.cf file. Then,
close the Wordpad and Windows Explorer windows.
Note: The VCS Simulator may take a moment to display as it discovers any
running simulations even thought there will not be any.
127
Launch the VCS Java Console and log in using the following credentials.
User name: admin
Password: password
A27
Copyright 2012 Symantec Corporation. All rights reserved.
From the Java Console, verify that the Member Systems and their Priority
and their AutoStartList attribute are set to the values shown in the following
table.
Service Group
AutoStartList
A1
S1
A2
S1
B1
S2
B2
S2
C1
S3
C2
S3
D1
S4
D2
S4
128 A28
winclient
1
From the Java Console, verify that the failover policy of all service groups is
Priority.
Verify that all service groups are online as shown in the following table.
Service Group
System S1
A1
Online
A2
Online
System S2
B1
Online
B2
Online
System S3
C1
Online
C2
Online
System S4
D1
Online
D2
Online
If it faults, where will the A1 service group fail over to? Verify this failover by
faulting a critical resource in the A1 service group.
Service Group A1 will fail over to:
129
If the A1 service group faults again without clearing the previous fault, where
should it fail over? Verify the failover by faulting a critical resource in the A1
service group.
Service Group A1 will fail over to:
Clear the faults in the A1 service group. Where will the service group fail over
to now? Verify the failover by faulting a critical resource in the A1 service
group.
Service Group A1 will fail over to:
A29
Copyright 2012 Symantec Corporation. All rights reserved.
winclient
1
From the Java Console, open the cluster configuration for update.
Set the FailOverPolicy attribute to Load for the eight service groups.
Set the Load attribute for each service group as shown in the following table.
130 A30
Service Group
Load
A1
75
A2
75
B1
75
B2
75
C1
50
C2
50
D1
50
D2
50
AvailableCapacity
S1
50
S2
50
S3
S4
0
Veritas Cluster Server 5.1 for UNIX: Install and Configure
Verify that all service groups are online as shown in the following table.
Service Group
System S1
A1
Online
A2
Online
System S2
B1
Online
B2
Online
System S3
C1
Online
C2
Online
System S4
D1
Online
D2
Online
If it faults, where will the A1 service group fail over to? Verify this failover by
faulting a critical resource in the A1 service group.
Service Group A1 will fail over to:
AvailableCapacity
S1
125
S2
-25
S3
S4
10 If the S2 system fails, where will the A1, B1 and B2 service groups that are
131
Online on S2 fail over? Verify this failover by powering off the S2 system in
Cluster Manager.
The A1 service group will fail over to:
The B1 service group will fail over to:
The B2 service group will fail over to:
A31
Copyright 2012 Symantec Corporation. All rights reserved.
following table.
Service Group
AvailableCapacity
S1
-25
S2
200
S3
-75
S4
12 Power up the S2 system, clear all faults, and return the service groups to their
startup locations.
13 Verify that the systems AvailableCapacity attribute is as shown in the
following table.
Service Group
AvailableCapacity
S1
50
S2
50
S3
S4
132 A32
winclient
1
From the Java Console, open the cluster configuration for update.
Set the Limits attribute for each system to the key ABGroup with a value of 3.
Set the Prerequisites attribute for the A1, A2, B1 and B2 service groups to be
ABGroup with a value of 1.
CAUTION Do not modify the Prerequisites attribute for the C1, C2, D1,
and D2 service groups.
If the S1 system fails, where will the A1 and A2 service groups that are Online
on S1 fail over? Verify this failover by powering off the S1 system in Cluster
Manager. Explain the failover actions that occurred.
The A1 service group will fail over to:
133
A33
Copyright 2012 Symantec Corporation. All rights reserved.
If the S2 system fails, where will the A1, B1 and B2 service groups that are
Online on S2 fail over? Verify this failover by powering off the S2 system in
Cluster Manager. Explain the failover actions that occurred.
The A1 service group will fail over to:
The B1 service group will fail over to:
The B2 service group will fail over to:
If the S3 system fails, where will the A2, B1, C1 and C2 service groups that
are Online on S3 fail over? Verify this failover by powering off the S3 system
in Cluster Manager. Explain the failover actions that occurred.
The A2 service group will fail over to:
The B1 service group will fail over to:
The C1 service group will fail over to:
The C2 service group will fail over to:
134 A34
S4
winclient
1
End of lab
135
A35
Copyright 2012 Symantec Corporation. All rights reserved.
136 A36
137
A37
138 A38
Verify that the following virtual machines shown are powered on, that you are
logged using the indicated account and that the indicated terminal windows are
opened.
mgt
There is no need to log into this virtual machine at this time.
sym1
Log in with credentials:
Account: root
Password: veritas
Two terminal windows should be active on this server.
sym2
Log in with credentials:
Account: root
Password: veritas
Two terminal windows should be active on this server.
sym3
Log in with credentials:
Account: root
Password: veritas
Two terminal windows should be active on this server.
sym4
Log in with credentials:
Account: root
Password: veritas
Two terminal windows should be active on this server.
Note: The terminal windows are referred to as hostname:terminal#
throughout the labs. For example: sym1:terminal1, sym1:terminal2,
sym2:terminal1, and sym2:terminal2. It is not necessary to label the
terminal windows and you may decide which is terminal1 and which is
terminal2.
If you have machines running that are not used in this lab, shut down the
operating system and power the machines off.
Note: If you are completing the lab exercises in order, this will require you to
shutdown winclient, and power on and log into sym1, sym2, sym3
and sym4.
139
A39
sym1
140 A40
Modify the SystemList to allow the netsg service group to run on the sym1,
sym2, sym3, and sym4 in that order.
Modify the AutoStartList attribute to allow the service group to start on all
systems.
Modify the Parallel service group to configure the netsg service group as a
parallel service group.
Display the service group attributes for the netsg service group to confirm your
input and key default service group attribute values.
Add a resource of type NIC named netnic to the netsg service group.
Modify the Device and Critical resource attributes for the netnic resource
using the following information.
Device is eth0
Critical set to 0 (zero)
10 Display the resource attribute values for the netnic resource to confirm your
11 Enable the netnic resource and verify that it is enabled. Display the state of the
Note: The service group state is offline even though the NIC resource state is
online. This is because VCS does not consider persistent resources
when determining the state of a service group.
13 Save the VCS configuration, but do not close it.
14 Add a resource of type Phantom named netphantom to the netsg service
group.
15 Set the Critical resource attribute to 0 (zero) for the netphantom resource.
16 Display the resource attribute values for the netphantom resource to confirm
141
Note: The service group state is offline even though the NIC resource state is
online. This is because VCS does not consider persistent resources
when determining the state of a service group.
A41
sym1
1
From sym1:terminal1, identify the names of the resources of type NIC that are
configured.
Ignoring the resource of type NIC named netnic, list the resource
dependencies associated with the listed resources of type NIC.
Replace the resource of type NIC named appnic with a resource of type Proxy
named appproxy using the following information.
Replace the resource of type NIC named csgnic with a resource of type Proxy
named csgproxy using the following information.
142 A42
Replace the resource of type NIC named nfsnic with a resource of type Proxy
named nfsproxy using the following information.
Replace the resource of type NIC named oranic with a resource of type Proxy
named oraproxy using the following information.
End of lab
A43
144 A44
Appendix B
Lab Solutions
145
146 B2
B3
Verify that the following virtual machines shown are powered on, that you are
logged using the indicated account and that the indicated terminal windows are
opened.
mgt
There is no need to log into this virtual machine at this time.
sym3
Log in with credentials:
Account: root
Password: veritas
Two terminal windows should be active on this server.
sym4
Log in with credentials:
Account: root
Password: veritas
Two terminal windows should be active on this server.
148 B4
If you have machines running that are not used in this lab, shut down the
operating system and power the machines off.
Note: If you are completing the lab exercises in order, this will require you to
shutdown cps, sym1 and sym2, and power on and log into sym3 and
sym4.
sym3
Note: The appsg and testappsg service groups were added in an early Install and
Configure lab using the Veritas Cluster Manager Java Console. If those
lab exercises were skipped, then consult with your instructor on how to
copy in a replacement main.cf file located at:
/student/labs/vcs/vcs60/maincf/IC/
main.cf.east.lab05
1
hastatus -sum
-- SYSTEM STATE
-- System
State
Frozen
A
A
RUNNING
RUNNING
0
0
sym3
sym4
-- GROUP STATE
-- Group
System
Probed
AutoDisabled
State
B
B
B
B
B
B
B
B
B
B
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
N
N
N
N
N
N
N
N
N
N
ONLINE
OFFLINE
ONLINE
OFFLINE
ONLINE
OFFLINE
ONLINE
OFFLINE
ONLINE
OFFLINE
ClusterService
ClusterService
appsg
appsg
dbsg
dbsg
testappsg
testappsg
websg
websg
hagrp -dep
VCS WARNING V-16-1-50035 No Group dependencies are configured
End of Solution
B5
Display the state of all resources by listing the resources for each service group
and the dependencies between each resource in each service group.
Solution
150 B6
hares -state
#Resource
appdg
appdg
appmnt
appmnt
appnic
appnic
appproc
appproc
appip
appip
appvol
appvol
csgnic
csgnic
dbdg
dbdg
dbip
dbip
dblistener
dblistener
dbmnt
dbmnt
dbnic
dbnic
dboracle
dboracle
dbvol
dbvol
notifier
notifier
testappdg
testappdg
testappmnt
testappmnt
testappnic
testappnic
testappproc
testappproc
testappip
testappip
testappvol
testappvol
webapache
webapache
webdg
webdg
webip
webip
webmnt
webmnt
webnic
webnic
webvip
webvip
webvol
webvol
Attribute
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
State
System
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
Value
ONLINE
OFFLINE
ONLINE
OFFLINE
ONLINE
OFFLINE
ONLINE
OFFLINE
ONLINE
OFFLINE
ONLINE
OFFLINE
ONLINE
ONLINE
ONLINE
OFFLINE
ONLINE
OFFLINE
ONLINE
OFFLINE
ONLINE
OFFLINE
ONLINE
OFFLINE
ONLINE
OFFLINE
ONLINE
OFFLINE
ONLINE
OFFLINE
ONLINE
OFFLINE
ONLINE
OFFLINE
ONLINE
OFFLINE
ONLINE
OFFLINE
ONLINE
OFFLINE
ONLINE
OFFLINE
ONLINE
OFFLINE
ONLINE
OFFLINE
ONLINE
OFFLINE
ONLINE
OFFLINE
ONLINE
OFFLINE
ONLINE
OFFLINE
ONLINE
OFFLINE
csgnic
csgnic
appmnt
appproc
appproc
appip
appvol
appvol
appip
appmnt
appnic
appdg
151
dbip
dblistener
dblistener
dbmnt
dboracle
dbvol
dbnic
dboracle
dbip
dbvol
dbmnt
dbdg
B7
testappmnt
testappproc
testappproc
testappip
testappvol
testappvol
testappip
testappmnt
testappnic
testappdg
webapache
webapache
webmnt
webvip
webvol
webvip
webmnt
webvol
webnic
webdg
End of Solution
Determine if all resources in all service groups other than the ClusterService
service group are set to critical. If any resources are set to non critical then
modify them. Open, save and close the cluster configuration as appropriate.
Solution
152 B8
End of Solution
Solution
Attribute
FaultPropagation
ManageFaults
System
global
global
Value
1
ALL
FaultPropagation
ManageFaults
global
global
1
ALL
FaultPropagation
ManageFaults
global
global
1
ALL
FaultPropagation
ManageFaults
global
global
1
ALL
FaultPropagation
ManageFaults
global
global
1
ALL
End of Solution
haconf -makerw
End of Solution
Solution
153
hastatus
testappproc
sym4
OFFLINE
testappip
sym3
ONLINE
testappip
sym4
OFFLINE
------------------------------------------------------------------------testappvol
sym3
ONLINE
testappvol
sym4
OFFLINE
webapache
sym3
OFFLINE
webapache
sym4
ONLINE
webdg
sym3
OFFLINE
------------------------------------------------------------------------webdg
sym4
ONLINE
group
resource
system
message
--------------- -------------------- -------------------- -------------------
B9
webmnt
sym3
OFFLINE
webmnt
sym4
ONLINE
webnic
sym3
OFFLINE
webnic
sym4
ONLINE
------------------------------------------------------------------------webvip
sym3
OFFLINE
webvip
sym4
ONLINE
webvol
sym3
OFFLINE
webvol
sym4
ONLINE
End of Solution
154 B10
sym3
Note: The service groups in this exercise will come online and go offline very
quickly. The hagrp -wait commands, while not necessary, have been
included to be consistent with other lab exercises.
1
155
hagrp -state
#Group
ClusterService
ClusterService
appsg
appsg
dbsg
dbsg
testappsg
testappsg
websg
websg
Attribute
State
State
State
State
State
State
State
State
State
State
System
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
Value
|ONLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
End of Solution
B11
Create and confirm dependencies between the appsg and dbsg and between
the websg and dbsg service groups as indicated below. Save but do not close
the VCS configuration.
appsg depends on dbsg (online local firm)
websg depends on dbsg (online local firm)
Solution
hagrp -dep
#Parent
appsg
websg
Child
dbsg
dbsg
Relationship
online local firm
online local firm
haconf -dump
End of Solution
Use a single hagrp command with the propagate option to bring the dbsg,
appsg, and websg service groups online on sym3. Can all three service groups
be brought online using just one command? If not, then ensure that all three
service groups are brought online on sym3.
Both top level parent service groups are independent so only one of the parents and
the shared child dbsg service group can be brought online using the propagate option.
Solution
156 B12
hagrp -state
#Group
ClusterService
ClusterService
appsg
Attribute
State
State
State
System
sym3
sym4
sym3
Value
|ONLINE|
|OFFLINE|
|OFFLINE|
appsg
dbsg
dbsg
testappsg
testappsg
websg
websg
State
State
State
State
State
State
State
sym4
sym3
sym4
sym3
sym4
sym3
sym4
|OFFLINE|
|ONLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
hagrp -state
#Group
ClusterService
ClusterService
appsg
appsg
dbsg
dbsg
testappsg
testappsg
websg
websg
Attribute
State
State
State
State
State
State
State
State
State
State
System
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
Value
|ONLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
End of Solution
157
Use a single hagrp command to take the dbsg, appsg, and websg service
groups offline on sym3. Can all three service groups be brought offline using
just one command? If not, then ensure that all three service groups are brought
offline on sym3.
All three service groups can be taken offline with one command.
Solution
B13
hagrp -state
#Group
ClusterService
ClusterService
appsg
appsg
dbsg
dbsg
testappsg
testappsg
websg
websg
Attribute
State
State
State
State
State
State
State
State
State
State
System
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
Value
|ONLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
End of Solution
Using a single hagrp command with the propagate option, bring the dbsg
and websg service groups online on sym3. Attempt to bring the appsg service
group online on sym4. Were you successful? If not successful, then bring the
appsg service group online on sym3.
The appsg service group cannot be brought online on sym4 due to service group
dependencies.
Solution
158 B14
hagrp -state
#Group
ClusterService
ClusterService
appsg
appsg
dbsg
dbsg
Attribute
State
State
State
State
State
State
System
sym3
sym4
sym3
sym4
sym3
sym4
Value
|ONLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
testappsg
testappsg
websg
websg
State
State
State
State
sym3
sym4
sym3
sym4
|OFFLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
End of Solution
Only appsg faults. Neither the dbsg nor websg failover or change state.
Solution
rm /var/tmp/appip
hagrp -state
#Group
ClusterService
ClusterService
appsg
appsg
dbsg
dbsg
testappsg
testappsg
websg
websg
Attribute
State
State
State
State
State
State
State
State
State
State
System
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
Value
|ONLINE|
|OFFLINE|
|OFFLINE|FAULTED|
|OFFLINE|
|ONLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
End of Solution
159
Clear the appip resource fault and bring the appsg service group back online
on sym3.
Solution
B15
hagrp -state
#Group
ClusterService
ClusterService
appsg
appsg
dbsg
dbsg
testappsg
testappsg
websg
websg
Attribute
State
State
State
State
State
State
State
State
State
State
System
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
Value
|ONLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
End of Solution
Fault the dbsg service group by removing the /var/tmp/dbip file followed
by a probe of the dbip resource on sym3. Do any service groups failover?
The dbsg service group faults. The appsg and websg service groups are taken offline
on sym3. All three service groups failover to sym4.
Solution
160 B16
rm /var/tmp/dbip
hagrp -state
#Group
ClusterService
ClusterService
appsg
appsg
dbsg
dbsg
testappsg
testappsg
websg
websg
Attribute
State
State
State
State
State
State
State
State
State
State
System
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
Value
|ONLINE|
|OFFLINE|
|ONLINE|STOPPING|
|OFFLINE|
|OFFLINE|FAULTED|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|ONLINE|STOPPING|
|OFFLINE|
hagrp -state
#Group
ClusterService
ClusterService
appsg
Attribute
State
State
State
System
sym3
sym4
sym3
Value
|ONLINE|
|OFFLINE|
|OFFLINE|
appsg
dbsg
dbsg
testappsg
testappsg
websg
websg
State
State
State
State
State
State
State
sym4
sym3
sym4
sym3
sym4
sym3
sym4
|ONLINE|
|OFFLINE|FAULTED|
|ONLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|ONLINE|
End of Solution
Clear the dbip resource fault and take the dbsg, websg, appsg, and testappsg
service groups offline on any system they may be online.
Solution
hagrp -state
#Group
ClusterService
ClusterService
appsg
appsg
dbsg
dbsg
testappsg
testappsg
websg
websg
Attribute
State
State
State
State
State
State
State
State
State
State
System
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
Value
|ONLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
End of Solution
161
10 Link the testappsg service group as a parent of the appsg service group with
an offline local dependency. Save, but do not close the VCS configuration.
Note: This will configure a preference for the appsg service group running on
a system over the testappsg service group and force the testappsg to
switch upon a successful failover of the appsg service group.
Solution
B17
hagrp -dep
#Parent
appsg
testappsg
websg
haconf -dump
Child
dbsg
appsg
dbsg
Relationship
online local firm
offline local
online local firm
haconf -dump
End of Solution
11 Bring the dgsg, appsg, and websg service groups online on sym3. Attempt to
bring the testappsg service group online on sym3. Were you successful? If
not, attempt to bring the testappsg service group online on sym4.
The testappsg service group cannot be brought online on the same system as the
appsg service group.
Solution
162 B18
hagrp -state
#Group
ClusterService
ClusterService
appsg
appsg
dbsg
dbsg
testappsg
testappsg
websg
websg
Attribute
State
State
State
State
State
State
State
State
State
State
System
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
Value
|ONLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
hagrp -state
#Group
ClusterService
ClusterService
appsg
appsg
dbsg
dbsg
testappsg
testappsg
websg
websg
Attribute
State
State
State
State
State
State
State
State
State
State
System
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
Value
|ONLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
|OFFLINE|
|ONLINE|
|ONLINE|
|OFFLINE|
End of Solution
12 Fault the dbsg service group by removing the /var/tmp/dbip file followed
Solution
163
rm /var/tmp/dbip
hagrp -state
#Group
ClusterService
ClusterService
appsg
appsg
dbsg
dbsg
testappsg
testappsg
websg
websg
Attribute
State
State
State
State
State
State
State
State
State
State
System
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
Value
|ONLINE|
|OFFLINE|
|PARTIAL|STOPPING|
|OFFLINE|
|OFFLINE|FAULTED|
|OFFLINE|
|OFFLINE|
|ONLINE|
|PARTIAL|STOPPING|
|OFFLINE|
B19
hagrp -state
#Group
ClusterService
ClusterService
appsg
appsg
dbsg
dbsg
testappsg
testappsg
websg
websg
Attribute
State
State
State
State
State
State
State
State
State
State
System
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
Value
|ONLINE|
|OFFLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|FAULTED|
|ONLINE|
|ONLINE|
|OFFLINE|
|OFFLINE|
|ONLINE|
End of Solution
13 Take the testappsg service group offline on sym3 and clear the dbip resource
fault. Then, use a single command to take the dbsg, websg, appsg, and
testappsg service groups offline on any system where they may be online.
Solution
164 B20
hagrp -state
#Group
ClusterService
ClusterService
appsg
appsg
dbsg
dbsg
testappsg
testappsg
websg
websg
Attribute
State
State
State
State
State
State
State
State
State
State
System
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
Value
|ONLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
End of Solution
14 Unlink the service group dependencies between the testappsg and appsg
service groups and between the dbsg, websg and appsg service groups. Save,
but do not close the VCS configuration.
Solution
hagrp -dep
VCS WARNING V-16-1-50035 No Group dependencies are configured
haconf -dump
End of Solution
165
B21
sym3
Note: The service groups in this exercise will come online and go offline very
quickly. The hagrp -wait commands, while not necessary, have been
included to be consistent with other lab exercises.
Solution
166 B22
hagrp -dep
#Parent
appsg
testappsg
websg
Child
dbsg
appsg
appsg
Relationship
online local firm
offline local
online local firm
haconf -dump
End of Solution
Bring the testappsg service group online on sym3. Use a single hagrp
command with the propagate option to bring the dbsg, appsg, and websg
service groups online on sym3. Were you successful? If not, then switch the
testappsg service group to sym4 and retry. Can all three service groups be
brought online using just one command? If not, then ensure that all three
service groups are brought online on sym3.
The testappsg is already online on sym3 through normal mechanisms so the appsg
service group cannot come online there. It could, under the proper circumstances, fail
over to sym3 preferentially, but it cannot just come online there with the testappsg
already online.
All three service groups can be brought online using a single command. The dbsg
service group is brought online, followed by the appsg service group and finally the
websg service group.
Solution
167
hagrp -state
#Group
ClusterService
ClusterService
appsg
appsg
dbsg
Attribute
State
State
State
State
State
System
sym3
sym4
sym3
sym4
sym3
Value
|ONLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
|ONLINE|
B23
dbsg
testappsg
testappsg
websg
websg
State
State
State
State
State
sym4
sym3
sym4
sym3
sym4
|OFFLINE|
|OFFLINE|
|ONLINE|
|ONLINE|
|OFFLINE|
End of Solution
Using a single hagrp command with the propagate option, take the dbsg,
appsg, and websg service groups offline on sym3. Can all three service
groups be brought offline using just one command? If not, then ensure that all
three service groups are brought offline on sym3.
All three service groups can be taken offline with one command. The testappsg
service group remains online.
Solution
168 B24
hagrp -state
#Group
ClusterService
ClusterService
appsg
appsg
dbsg
dbsg
testappsg
testappsg
websg
websg
Attribute
State
State
State
State
State
State
State
State
State
State
System
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
Value
|ONLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
|OFFLINE|
End of Solution
Using a single hagrp command with the propagate option, bring the dbsg,
appsg, and websg service groups online on sym3.
Solution
hagrp -state
#Group
ClusterService
ClusterService
appsg
appsg
dbsg
dbsg
testappsg
testappsg
websg
websg
Attribute
State
State
State
State
State
State
State
State
State
State
System
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
Value
|ONLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
|OFFLINE|
|ONLINE|
|ONLINE|
|OFFLINE|
End of Solution
The appsg faults. The websg is taken offline. No service groups fail over.
Solution
Solution
169
rm /var/tmp/appip
hagrp -state
#Group
ClusterService
ClusterService
appsg
appsg
dbsg
dbsg
testappsg
testappsg
websg
websg
Attribute
State
State
State
State
State
State
State
State
State
State
System
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
Value
|ONLINE|
|OFFLINE|
|OFFLINE|FAULTED|
|OFFLINE|
|ONLINE|
|OFFLINE|
|OFFLINE|
|ONLINE|
|PARTIAL|STOPPING|
|OFFLINE|
End of Solution
B25
Clear the appip resource fault and use a single hagrp command with the
propagate option to bring the appsg service group back online on sym3.
Solution
hagrp -state
#Group
ClusterService
ClusterService
appsg
appsg
dbsg
dbsg
testappsg
testappsg
websg
websg
Attribute
State
State
State
State
State
State
State
State
State
State
System
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
Value
|ONLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
|OFFLINE|
|ONLINE|
|ONLINE|
|OFFLINE|
End of Solution
Fault the dbsg service group by removing the /var/tmp/dbip file followed
by a probe of the dbip resource on sym3. Do any service groups failover?
The dbsg service group faults. The appsg and websg service groups are brought
offline on sym3. The testappsg service group is switched from sym4 to sym3. The
170 B26
Solution
rm /var/tmp/dbip
hagrp -state
#Group
Attribute
ClusterService State
ClusterService State
System
sym3
sym4
Value
|ONLINE|
|OFFLINE|
appsg
appsg
dbsg
dbsg
testappsg
testappsg
websg
websg
State
State
State
State
State
State
State
State
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
|ONLINE|STOPPING|
|OFFLINE|
|OFFLINE|FAULTED|
|OFFLINE|
|OFFLINE|
|ONLINE|
|ONLINE|
|OFFLINE|
hagrp -state
#Group
ClusterService
ClusterService
appsg
appsg
dbsg
dbsg
testappsg
testappsg
websg
websg
Attribute
State
State
State
State
State
State
State
State
State
State
System
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
Value
|ONLINE|
|OFFLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|FAULTED|
|ONLINE|
|PARTIAL|STARTING|
|OFFLINE|
|OFFLINE|
|ONLINE|
End of Solution
Take the testappsg service group offline on sym3 and clear the dbip resource
fault. Then, use a single command to bring the dbsg, websg, appsg, and
testappsg service groups offline on any system they may be online.
Solution
171
hagrp -state
#Group
ClusterService
ClusterService
appsg
appsg
dbsg
dbsg
testappsg
testappsg
Attribute
State
State
State
State
State
State
State
State
System
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
Value
|ONLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
B27
websg
websg
State
State
sym3
sym4
|OFFLINE|
|OFFLINE|
End of Solution
Unlink the service group dependencies between the testappsg and appsg
service groups, between the dbsg and appsg service groups and between the
websg and appsg service groups. Save, but do not close the VCS
configuration.
Solution
hagrp -dep
VCS WARNING V-16-1-50035 No Group dependencies are configured
haconf -dump
End of Solution
172 B28
sym3
Note: The service groups in this exercise will come online and go offline very
quickly. The hagrp -wait commands, while not necessary, have been
included to be consistent with other lab exercises.
Solution
173
hagrp -dep
#Parent
testappsg
websg
websg
Child
appsg
appsg
dbsg
Relationship
offline local
online local firm
online local firm
haconf -dump
End of Solution
B29
Bring the testappsg service group online on sym4. Using a single hagrp
command with the propagate option, bring the dbsg, appsg, and websg
service groups online on sym3. Were you able to bring the websg, dbsg, and
appsg service groups online with the online propagate option?
Solution
hagrp -state
#Group
ClusterService
ClusterService
appsg
appsg
dbsg
dbsg
testappsg
testappsg
websg
websg
Attribute
State
State
State
State
State
State
State
State
State
State
System
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
Value
|ONLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
|OFFLINE|
|ONLINE|
|ONLINE|
|OFFLINE|
End of Solution
174 B30
Using a single hagrp command with the propagate option, take the websg,
dbsg, and appsg and service groups offline on sym3. Can all three service
groups be brought offline using just one command? If not, then ensure that all
three service groups are brought offline on sym3.
No, all three service groups cannot be taken offline with one command. The appsg
service group needs to be taken offline explicitly in this case.
Solution
hagrp -state
#Group
ClusterService
ClusterService
appsg
appsg
dbsg
dbsg
testappsg
testappsg
websg
websg
Attribute
State
State
State
State
State
State
State
State
State
State
System
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
Value
|ONLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
|OFFLINE|
hagrp -state
#Group
ClusterService
ClusterService
appsg
appsg
dbsg
dbsg
testappsg
testappsg
websg
websg
Attribute
State
State
State
State
State
State
State
State
State
State
System
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
Value
|ONLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
|OFFLINE|
End of Solution
175
Using a single hagrp command with the propagate option, bring the dbsg,
appsg, and websg service groups online on sym3.
Solution
B31
hagrp -state
#Group
ClusterService
ClusterService
appsg
appsg
dbsg
dbsg
testappsg
testappsg
websg
websg
Attribute
State
State
State
State
State
State
State
State
State
State
System
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
Value
|ONLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
|OFFLINE|
|ONLINE|
|ONLINE|
|OFFLINE|
End of Solution
Solution
176 B32
rm /var/tmp/webvip
hagrp -state
#Group
ClusterService
ClusterService
appsg
appsg
dbsg
dbsg
testappsg
testappsg
websg
websg
Attribute
State
State
State
State
State
State
State
State
State
State
System
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
Value
|ONLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|FAULTED|
|OFFLINE|
End of Solution
Clear the webvip resource fault and bring the websg service group back online
on sym3.
Solution
hagrp -state
#Group
ClusterService
ClusterService
appsg
appsg
dbsg
dbsg
testappsg
testappsg
websg
websg
Attribute
State
State
State
State
State
State
State
State
State
State
System
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
Value
|ONLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
|OFFLINE|
|ONLINE|
|ONLINE|
|OFFLINE|
End of Solution
Fault the dbsg service group by removing the /var/tmp/dbip file followed
by a probe of the dbip resource on sym3. Do any service groups failover?
The dbsg service group faults on sym3. The websg service group is taken offline on
sym3. The appsg service group is unaffected and no failover occurs. The testappsg
177
Solution
rm /var/tmp/dbip
hagrp -state
#Group
Attribute
ClusterService State
ClusterService State
System
sym3
sym4
Value
|ONLINE|
|OFFLINE|
B33
appsg
appsg
dbsg
dbsg
testappsg
testappsg
websg
websg
State
State
State
State
State
State
State
State
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
|ONLINE|
|OFFLINE|
|OFFLINE|FAULTED|
|ONLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
|OFFLINE|
End of Solution
Take the appsg service group offline on sym3. Then, bring the websg, appsg,
and dbsg service groups online on sym4. Account for the testappsg service
group.
Solution
178 B34
hagrp -state
#Group
ClusterService
ClusterService
appsg
appsg
dbsg
dbsg
testappsg
testappsg
websg
websg
Attribute
State
State
State
State
State
State
State
State
State
State
System
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
Value
|ONLINE|
|OFFLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|FAULTED|
|ONLINE|
|ONLINE|
|OFFLINE|
|OFFLINE|
|ONLINE|
End of Solution
hagrp -state
#Group
ClusterService
ClusterService
appsg
appsg
dbsg
dbsg
testappsg
testappsg
websg
websg
Attribute
State
State
State
State
State
State
State
State
State
State
System
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
Value
|ONLINE|
|OFFLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
|ONLINE|
|ONLINE|
|OFFLINE|
|OFFLINE|
|ONLINE|
End of Solution
10 Unlink the service group dependencies between the testappsg and appsg
service groups, between the dbsg and appsg service groups, and between the
websg and appsg service groups. Save and close the VCS configuration.
Solution
hagrp -dep
179
End of Solution
B35
End of lab
180 B36
181
B37
Copyright 2012 Symantec Corporation. All rights reserved.
182 B38
Verify that the following virtual machines shown are powered on, that you are
logged using the indicated account and that the indicated terminal windows are
opened.
mgt
There is no need to log into this virtual machine at this time.
sym1
Log in with credentials:
Account: root
Password: veritas
Two terminal windows should be active on this server.
sym2
Log in with credentials:
Account: root
Password: veritas
Two terminal windows should be active on this server.
sym3
Log in with credentials:
Account: root
Password: veritas
Two terminal windows should be active on this server.
sym4
Log in with credentials:
Account: root
Password: veritas
Two terminal windows should be active on this server.
Note: The terminal windows are referred to as hostname:terminal#
throughout the labs. For example: sym1:terminal1, sym1:terminal2,
sym2:terminal1, and sym2:terminal2. It is not necessary to label the
terminal windows and you may decide which is terminal1 and which is
terminal2.
If you have machines running that are not used in this lab, shut down the
operating system and power the machines off.
Note: If you are completing the lab exercises in order, this will require you to
power on and log into sym1 and sym2.
183
B39
Copyright 2012 Symantec Corporation. All rights reserved.
sym1
Note: In this lab there are no iSCSI LUNs that are visible to all four systems, so
merging enabled I/O fencing using disks cannot be part of the cluster
merge lab. Therefore, for lab purposes, I/O fencing is re-configured to
disabled mode on both clusters.
cd /opt/VRTS/install
./installsfha -fencing
Checking communication on sym1
Checking release compatibility on sym1
Checking VCS installation on sym1
184 B40
y
Checking
Done
Checking
Checking
Checking
Done
Checking release compatibility on sym2 ............................ Done
Checking VCS installation on sym2 .................. Version 6.0.000.000
Fencing is already started in enabled mode, do you want to reconfigure it?
[y,n,q] (y)
y
Fencing configuration
1) Configure Coordination Point client based fencing
2) Configure disk based fencing
3) Configure fencing in disabled mode
4) Online fencing migration
Select the fencing mechanism to be configured in this Application Cluster:
3
Installer will stop VCS before applying fencing configuration. To make sure
VCS shuts down successfully, unfreeze any frozen service group and unmount
the mounted file systems in the cluster.
Are you ready to stop VCS and apply fencing configuration on all nodes at
this time? [y,n,q,b] (y)
y
Deleting Coordination Point Agent Service Group westfensg ......... Done
Stopping VCS on sym1 ..............................................
Done
Stopping Fencing on sym1 ..........................................
Done
Stopping VCS on sym2 ..............................................
Done
Stopping Fencing on sym2 ..........................................
Done
Configuring fencing in disabled mode on sym1 ...................... Done
Configuring fencing in disabled mode on sym2 ...................... Done
Updating main.cf without fencing ..................................
Done
Starting VCS on sym1 ..............................................
Done
Starting VCS on sym2 ..............................................
Done
185
Done
I/O Fencing configuration completed successfully
...
installsfha log files, summary file, and response file are saved at:
/opt/VRTS/install/logs/installsfha-201112081226AJb
Would you like to view the summary file? [y,n,q] (n)
End of Solution
B41
Copyright 2012 Symantec Corporation. All rights reserved.
gabconfig -a
GAB Port Memberships
===============================================================
Port a gen
1b5701 membership 01
Port b gen
1b5707 membership 01
Port h gen
1b570a membership 01
vxfenadm -d
I/O Fencing Cluster Information:
================================
Fencing Protocol Version: 201
Fencing Mode: Disabled
Cluster Members:
* 0 (sym1)
1 (sym2)
RFSM State Information:
node
0 in state 8 (running)
node
1 in state 8 (running)
End of Solution
sym3
186 B42
cd /opt/VRTS/install
./installsfha -fencing
Checking communication on sym3
Checking release compatibility on sym3
Checking VCS installation on sym3
Cluster information verification:
Cluster Name: east
Cluster ID Number: 777
Systems: sym3 sym4
Would you like to configure I/O fencing on the cluster? [y,n,q]
y
Checking
Done
Checking
Checking
Checking
Done
Checking
Checking
y
Fencing configuration
1) Configure Coordination Point client based fencing
2) Configure disk based fencing
3) Configure fencing in disabled mode
4) Online fencing migration
Select the fencing mechanism to be configured in this Application Cluster:
187
3
Installer will stop VCS before applying fencing configuration. To make sure
VCS shuts down successfully, unfreeze any frozen service group and unmount
the mounted file systems in the cluster.
Are you ready to stop VCS and apply fencing configuration on all nodes at
this time? [y,n,q,b] (y)
y
Stopping VCS on sym3 .............................................
Stopping Fencing on sym3 .........................................
Stopping VCS on sym4 .............................................
Stopping Fencing on sym4 .........................................
Configuring fencing in disabled mode on sym3 ......................
Configuring fencing in disabled mode on sym4 ......................
Updating main.cf without fencing .................................
Starting VCS on sym3 .............................................
Starting VCS on sym4 .............................................
Done
Done
Done
Done
Done
Done
Done
Done
Done
B43
Copyright 2012 Symantec Corporation. All rights reserved.
End of Solution
gabconfig -a
GAB Port Memberships
===============================================================
Port a gen
1b5701 membership 01
Port b gen
1b5707 membership 01
Port h gen
1b570a membership 01
188 B44
vxfenadm -d
I/O Fencing Cluster Information:
================================
Fencing Protocol Version: 201
Fencing Mode: Disabled
Cluster Members:
* 0 (sym3)
1 (sym4)
RFSM State Information:
node
0 in state 8 (running)
node
1 in state 8 (running)
End of Solution
Note: In this lab, there is a conflicting duplicate named appsg service group in
both the west and east clusters. This would cause errors during a merge.
The appsg service group in the west cluster would have precedence so the
appsg service group in the east cluster must be renamed or, as performed
in this lab, deleted from the configuration.
Take the appsg service group offline and list its resources and dependencies.
Solution
hagrp -state
#Group
ClusterService
ClusterService
appsg
appsg
dbsg
dbsg
testappsg
testappsg
websg
websg
Attribute
State
State
State
State
State
State
State
State
State
State
System
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
Value
|ONLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
hagrp -dep
VCS WARNING V-16-1-50035 No Group dependencies are configured
189
Attribute
State
State
System
sym3
sym4
Value
|OFFLINE|
|OFFLINE|
appmnt
appproc
appproc
appip
appvol
appvol
appip
appmnt
appnic
appdg
End of Solution
B45
Copyright 2012 Symantec Corporation. All rights reserved.
Open the VCS configuration for update and delete each resource in the appsg
service group in resource dependency order. Then, delete the appsg service
group and save and close the VCS configuration.
Solution
haconf -makerw
hagrp -state
#Group
ClusterService
ClusterService
dbsg
dbsg
testappsg
testappsg
websg
websg
190 B46
Attribute
State
State
State
State
State
State
State
State
System
sym3
sym4
sym3
sym4
sym3
sym4
sym3
sym4
Value
|ONLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
|ONLINE|
|OFFLINE|
End of Solution
cd /etc/VRTSvcs/conf/config
hacf -cftocmd .
Veritas Cluster Server 5.1 for UNIX: Install and Configure
Copyright 2012 Symantec Corporation. All rights reserved.
cp -p main.cmd eastmain.cmd
ls *.cmd
eastmain.cmd
main.cmd
End of Solution
Edit the eastmain.cmd file to remove all commands other than those
related to the websg, testappsg, dbsg service groups, their resources, and their
resource dependencies.
191
After you have made the edits the contents of the file should be:
hagrp -add dbsg
hagrp -modify dbsg SystemList sym3 0 sym4 1
hagrp -modify dbsg AutoStartList sym3 sym4
hagrp -modify dbsg SourceFile "./main.cf"
hares -add dbdg FileOnOff dbsg
hares -modify dbdg PathName "/var/tmp/dbdg"
hares -modify dbdg Enabled 1
hares -add dbip FileOnOff dbsg
hares -modify dbip PathName "/var/tmp/dbip"
hares -modify dbip Enabled 1
hares -add dblistener FileOnOff dbsg
hares -modify dblistener PathName "/var/tmp/
dblistener"
hares -modify dblistener Enabled 1
hares -add dbmnt FileOnOff dbsg
hares -modify dbmnt PathName "/var/tmp/dbmnt"
hares -modify dbmnt Enabled 1
hares -add dbnic FileOnOff dbsg
hares -modify dbnic PathName "/var/tmp/dbnic"
hares -modify dbnic Enabled 1
hares -add dboracle FileOnOff dbsg
hares -modify dboracle PathName "/var/tmp/
dboracle"
hares -modify dboracle Enabled 1
hares -add dbvol FileOnOff dbsg
B47
Copyright 2012 Symantec Corporation. All rights reserved.
192 B48
193
hares
hares
hares
hares
hares
hares
hares
hares
hares
hares
hares
hares
hares
hares
hares
hares
hares
hares
hares
hares
hares
hares
hares
hares
hares
hares
hares
hares
hares
hares
hares
Solution
vi eastmain.cmd
B49
Copyright 2012 Symantec Corporation. All rights reserved.
End of Solution
Shutdown VCS, IO fencing, GAB, and LLT on both nodes of the east cluster.
Solution
hastop -all
gabconfig -a
GAB Port Memberships
===============================================================
Port a gen
a9e001 membership 01
Port b gen
a9e007 membership 01
194 B50
gabconfig -a
GAB gabconfig ERROR V-15-2-25022 unknown error
End of Solution
mv main.cf main.cf.east
ls -l main.cf
ls: main.cf No such file or directory
End of Solution
195
B51
Copyright 2012 Symantec Corporation. All rights reserved.
sym1
1
installsfha-yyyymmddhhmmxxx
Solution
cd /opt/VRTS/install
./installsfha -addnode
Logs are being written to /var/tmp/installsfha-201112081249IHO while
installsfha
is in progress.
196 B52
* The cluster to which you want to add the node must have all required
SFHA rpms installed
* New node must have all required SFHA rpms installed
* SFHA must be running on the cluster to which you want to add a node
* There should be no prior running SFHA processes on the new node
* New node must have the same SFHA version as that of the existing
cluster nodes
Press Enter.
Enter one node of the SFHA cluster to which you would like to add one or
more new nodes:
sym1
Checking communication on sym1 ....................................
Done
Checking release compatibility on sym1 ............................ Done
Following cluster information detected:
Cluster Name: west
Cluster ID: 5
Systems: sym1 sym2
Is this information correct? [y,n,q] (y)
y
Checking communication on sym2 ....................................
Done
Checking VCS running state on sym1 ................................ Done
Checking VCS running state on sym2 ................................ Done
Enter the system names separated by spaces to add to the cluster:
sym3 sym4
Checking
Done
Checking
Checking
Done
Checking
Do you want to add the system(s) sym3 sym4 to the cluster west? [y,n,q] (y)
197
y
Checking
Checking
Checking
Checking
Checking
installed
installed
installed
installed
installed
Each node to be added to the cluster should have the same LLT heartbeat
configuration as the existing node(s).
To connect to the existing node(s) properly, each new node is required to be
configured with 2 private and 1 low-priority heartbeat links.
Discovering NICs on sym3 ...... Discovered eth0 eth1 eth2 eth3 eth4 eth5
Enter the NIC for the first private heartbeat link on sym3: [q,?] (eth4)
eth4
Enter the NIC for the second private heartbeat link on sym3: [b,q,?] (eth5)
B53
Copyright 2012 Symantec Corporation. All rights reserved.
eth5
Enter the NIC for the low-priority heartbeat link on sym3: [b,q,?] (eth0)
eth0
Are you using the same NICs for private heartbeat links on all systems?
[y,n,q,b,?]
y
Checking
Checking
Checking
Checking
Checking
Checking
media
media
media
media
media
media
speed
speed
speed
speed
speed
speed
for
for
for
for
for
for
eth4
eth5
eth4
eth5
eth4
eth5
on
on
on
on
on
on
sym3
sym3
sym4
sym4
sym1
sym1
.........................
.........................
.........................
.........................
.........................
.........................
1000Mb/s
1000Mb/s
1000Mb/s
1000Mb/s
1000Mb/s
1000Mb/s
for sym3:
y
Cluster Virtual IP is configured on the cluster west
A public NIC device is required by following services on each of the newly
added nodes:
Cluster Virtual IP
Active NIC devices discovered on sym3: eth0 eth1 eth2 eth3 eth4 eth0
198 B54
m eth0
Is eth0 the public NIC used by other new nodes? [y,n,q] (y)
y
Starting llt
Starting
Starting
Starting
Done
Checking
Updating
Updating
Starting
Starting
seed
Done
Done
Done
Done
Starting
Starting
Starting
Starting
amf
had
amf
had
on
on
on
on
sym3
sym3
sym4
sym4
.............................................
.............................................
.............................................
.............................................
Done
Done
Done
Done
End of Solution
Perform a summary status of the cluster noting that none of the former east
cluster service groups have been added.
Solution
hastatus -sum
-- SYSTEM STATE
-- System
State
Frozen
A
A
A
A
RUNNING
RUNNING
RUNNING
RUNNING
0
0
0
0
sym1
sym2
sym3
sym4
-- GROUP STATE
-- Group
System
B
B
B
B
B
B
B
B
B
B
sym1
sym2
sym3
sym4
sym1
sym2
sym1
sym2
sym1
sym2
ClusterService
ClusterService
ClusterService
ClusterService
appsg
appsg
nfssg
nfssg
orasg
orasg
Probed
Y
Y
Y
Y
Y
Y
Y
Y
Y
AutoDisabled
N
N
N
N
N
N
N
N
N
State
ONLINE
OFFLINE
OFFLINE
OFFLINE
ONLINE
OFFLINE
ONLINE
OFFLINE
ONLINE
OFFLINE
End of Solution
199
gabconfig -a
GAB Port Memberships
===============================================================
Port a gen
1b5702 membership 0123
Port b gen
1b5709 membership 0123
Port h gen
1b570c membership 0123
End of Solution
B55
Copyright 2012 Symantec Corporation. All rights reserved.
vxfenadm -d
I/O Fencing Cluster Information:
================================
Fencing Protocol Version: 201
Fencing Mode: Disabled
Cluster Members:
* 0
1
2
3
(sym1)
(sym2)
(sym3)
(sym4)
End of Solution
Use the lltstat command to check the LLT configured and active status.
Optionally, performed this step on all cluster nodes.
Solution
1 sym2
200 B56
Link
Status
Address
eth4
eth5
eth0
UP
UP
UP
00:0C:29:5A:B1:16
00:0C:29:5A:B1:20
00:0C:29:5A:B1:EE
eth4
eth5
eth0
UP
UP
UP
00:0C:29:2D:77:9A
00:0C:29:2D:77:A4
00:0C:29:2D:77:72
eth4
eth5
eth0
UP
UP
UP
00:0C:29:97:FB:5D
00:0C:29:97:FB:67
00:0C:29:97:FB:35
eth4
eth5
eth0
UP
UP
UP
00:0C:29:C4:1D:0A
00:0C:29:C4:1D:14
00:0C:29:C4:1D:E2
OPEN
2 sym3
OPEN
3 sym4
State
OPEN
OPEN
State
OPEN
Link
eth4
eth5
Status
UP
UP
Address
00:0C:29:5A:B1:16
00:0C:29:5A:B1:20
1 sym2
eth0
UP
00:0C:29:5A:B1:EE
eth4
eth5
eth0
UP
UP
UP
00:0C:29:2D:77:9A
00:0C:29:2D:77:A4
00:0C:29:2D:77:72
eth4
eth5
eth0
UP
UP
UP
00:0C:29:97:FB:5D
00:0C:29:97:FB:67
00:0C:29:97:FB:35
eth4
eth5
eth0
UP
UP
UP
00:0C:29:C4:1D:0A
00:0C:29:C4:1D:14
00:0C:29:C4:1D:E2
OPEN
2 sym3
OPEN
3 sym4
OPEN
End of Solution
Solution
cat /etc/llthosts
0
1
2
3
sym1
sym2
sym3
sym4
cat /etc/llttab
set-node sym3
set-cluster 5
link eth4 eth-00:0c:29:97:fb:5d - ether - link eth5 eth-00:0c:29:97:fb:67 - ether - link-lowpri eth0 eth-00:0c:29:97:fb:35 - ether - -
cat /etc/gabtab
/sbin/gabconfig -c -n4
End of Solution
201
B57
Copyright 2012 Symantec Corporation. All rights reserved.
Solution
Note: Only the main.cf file though the ClusterService service group
definition is shown.
more /etc/VRTSvcs/conf/config/main.cf
...
include
include
include
include
include
"OracleASMTypes.cf"
"types.cf"
"Db2udbTypes.cf"
"OracleTypes.cf"
"SybaseTypes.cf"
cluster west (
UserNames = { admin = bQRjQLqNRmRRpZRlQO }
ClusterAddress = "10.10.2.51"
Administrators = { admin }
)
system sym1 (
)
system sym2 (
)
system sym3 (
)
system sym4 (
)
group ClusterService (
SystemList = { sym1 = 0, sym2 = 1, sym3 = 2, sym4 = 3 }
AutoStartList = { sym1, sym2, sym3, sym4 }
OnlineRetryLimit = 3
OnlineRetryInterval = 120
)
202 B58
IP webip (
Device = eth0
Address = "10.10.2.51"
NetMask = "255.255.255.0"
)
NIC csgnic (
Device = eth0
)
NotifierMngr notifier (
SmtpServer = "mgt.example.com"
SmtpServerVrfyOff = 1
SmtpRecipients = { student = Warning }
)
notifier requires csgnic
webip requires csgnic
...
End of Solution
sym3
1
cd /etc/VRTSvcs/conf/config
haconf -makerw
End of Solution
Solution
sh -x ./eastmain.cmd
+ hagrp -add dbsg
VCS NOTICE V-16-1-10136 Group added; populating SystemList and setting the
Parallel attribute recommended before adding resources
+ hagrp -modify dbsg SystemList sym3 0 sym4 1
+ hagrp -modify dbsg AutoStartList sym3 sym4
+ hagrp -modify dbsg SourceFile ./main.cf
+ hares -add dbdg FileOnOff dbsg
VCS NOTICE V-16-1-10242 Resource added. Enabled attribute must be set before
agent monitors
+ hares -modify dbdg PathName /var/tmp/dbdg
+ hares -modify dbdg Enabled 1
+ hares -add dbip FileOnOff dbsg
VCS NOTICE V-16-1-10242 Resource added. Enabled attribute must be set before
agent monitors
+ hares -modify dbip PathName /var/tmp/dbip
B59
Copyright 2012 Symantec Corporation. All rights reserved.
204 B60
set before
set before
set before
set before
set before
set before
End of Solution
B61
Copyright 2012 Symantec Corporation. All rights reserved.
sym3
1
hastatus -sum
206 B62
-- SYSTEM STATE
-- System
State
Frozen
A
A
A
A
RUNNING
RUNNING
RUNNING
RUNNING
0
0
0
0
sym1
sym2
sym3
sym4
-- GROUP STATE
-- Group
System
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B
B
sym1
sym2
sym3
sym4
sym1
sym2
sym3
sym4
sym1
sym2
sym1
sym2
sym3
sym4
sym3
sym4
ClusterService
ClusterService
ClusterService
ClusterService
appsg
appsg
dbsg
dbsg
nfssg
nfssg
orasg
orasg
testappsg
testappsg
websg
websg
Probed
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
AutoDisabled
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
State
ONLINE
OFFLINE
OFFLINE
OFFLINE
ONLINE
OFFLINE
OFFLINE
OFFLINE
ONLINE
OFFLINE
ONLINE
OFFLINE
OFFLINE
OFFLINE
OFFLINE
OFFLINE
End of Solution
more /etc/VRTSvcs/conf/config/main.cf
...
cluster west (
UserNames = { admin = bQRjQLqNRmRRpZRlQO }
ClusterAddress = "10.10.2.51"
Administrators = { admin }
)
...
group ClusterService (
SystemList = { sym1 = 0, sym2 = 1, sym3 = 2, sym4 = 3 }
AutoStartList = { sym1, sym2, sym3, sym4 }
OnlineRetryLimit = 3
OnlineRetryInterval = 120
)
...
group appsg (
SystemList = { sym1 = 0, sym2 = 1 }
AutoStartList = { sym1, sym2 }
TriggersEnabled @sym1 = { NOFAILOVER, RESFAULT }
TriggersEnabled @sym2 = { NOFAILOVER, RESFAULT }
)
...
group nfssg (
SystemList = { sym1 = 0, sym2 = 1 }
AutoStartList = { sym1, sym2 }
)
...
group orasg (
SystemList = { sym1 = 0, sym2 = 1 }
AutoStartList = { sym1, sym2 }
B63
Copyright 2012 Symantec Corporation. All rights reserved.
)
...
group websg (
SystemList = { sym3 = 0, sym4 = 1 }
AutoStartList = { sym3, sym4 }
)
...
group testappsg (
SystemList = { sym3 = 0, sym4 = 1 }
AutoStartList = { sym3, sym4 }
)
...
group dbsg (
SystemList = { sym3 = 0, sym4 = 1 }
AutoStartList = { sym3, sym4 }
...
End of Solution
End of lab
208 B64
B65
Copyright 2012 Symantec Corporation. All rights reserved.
Verify that the following virtual machines shown are powered on, that you are
logged using the indicated account and that the indicated terminal windows are
opened.
mgt
There is no need to log into this virtual machine at this time.
winclient
Log in with credentials:
Account: Administrator
Password: veritas
Note: The terminal windows are referred to as hostname:terminal#
throughout the labs. For example: sym1:terminal1, sym1:terminal2,
sym2:terminal1, and sym2:terminal2. It is not necessary to label the
terminal windows and you may decide which is terminal1 and which is
terminal2.
210 B66
If you have machines running that are not used in this lab, shut down the
operating system and power the machines off.
Note: If you are completing the lab exercises in order, this will require you to
shutdown sym1, sym2, sym3 and sym4, and power on and log into
winclient.
winclient
1
From the left pane, expand and select Computer > C: >
Program Files (x86) > VERITAS > VCS Simulator > wlm > conf >
config.
End of Solution
Solution
211
From the pop-up window, select Select a program from a list of installed
programs.
Click OK.
Clear Always use the selected program to open this kind of file.
Select WordPad.
Click OK.
End of Solution
B67
Copyright 2012 Symantec Corporation. All rights reserved.
Without making any changes, review the contents of the main.cf file. Then,
close the Wordpad and Windows Explorer windows.
Solution
From the Wordpad window, review the contents of the main.cf file.
Click the Close button that appears as an X in the upper right hand corner
of the window.
From the Windows Explorer window, click the Close button that appears
as an X in the upper right hand corner of the window.
End of Solution
End of Solution
212 B68
From the Symantec Veritas Cluster Server Simulation window, from the
left pane list, select wlm.
End of Solution
End of Solution
Launch the VCS Java Console and log in using the following credentials.
User name: admin
Password: password
Solution
Click OK.
End of Solution
213
B69
Copyright 2012 Symantec Corporation. All rights reserved.
From the Java Console, verify that the Member Systems and their Priority
and their AutoStartList attribute are set to the values shown in the following
table.
Service Group
AutoStartList
A1
S1
A2
S1
B1
S2
B2
S2
C1
S3
C2
S3
D1
S4
D2
S4
Solution
From the System Details section, verify Member Systems and their
Priority.
Repeat steps a through d for the A2, B1, B2, C1, C2, D1 and D2 service
groups.
End of Solution
214 B70
winclient
1
From the Java Console, verify that the failover policy of all service groups is
Priority.
Solution
From the right pane, verify that the FailOverPolicy attribute is set to
Priority.
Repeat steps a and b for the A2, B1, B2, C1, C2, D1 and D2 service
groups.
End of Solution
215
Verify that all service groups are online as shown in the following table.
Service Group
System S1
A1
Online
A2
Online
System S2
B1
Online
B2
Online
System S3
C1
Online
C2
Online
System S4
D1
Online
D2
Online
Solution
B71
Copyright 2012 Symantec Corporation. All rights reserved.
Verify the service groups are online as shown in the above table.
End of Solution
If it faults, where will the A1 service group fail over to? Verify this failover by
faulting a critical resource in the A1 service group.
Service Group A1 will fail over to:
S2
Solution
From left pane, expand wlm > A1 > FileOnOff > A1File.
From the right pane, notice that the A1 service group shows faulted on S1,
but Online on S2.
End of Solution
216 B72
If the A1 service group faults again without clearing the previous fault, where
should it fail over? Verify the failover by faulting a critical resource in the A1
service group.
Service Group A1 will fail over to:
S3
Solution
From the left pane, right-click wlm > A1> FileOnOff > A1File and select
Fault Resource > S2.
From the right pane, notice that the A1 service group shows faulted on S1,
and S2 but Online on S3.
End of Solution
Clear the faults in the A1 service group. Where will the service group fail over
to now? Verify the failover by faulting a critical resource in the A1 service
group.
Service Group A1 will fail over to:
S1
Solution
From the left pane, right-click wlm > A1 > FileOnOff > A1File and select
Clear Fault > Auto.
From the right pane, notice that the A1 service group shows faulted on S3,
but Online on S1.
End of Solution
From the left pane, right-click wlm > A1 > FileOnOff > A1File and select
Clear Fault > Auto.
217
End of Solution
B73
Copyright 2012 Symantec Corporation. All rights reserved.
winclient
1
From the Java Console, open the cluster configuration for update.
Solution
Set the FailOverPolicy attribute to Load for the eight service groups.
Solution
218 B74
From the Edit Attribute window, from the Scaler Value drop-down menu,
select Load.
Note: The drop-down menu will appear when you click in the Scaler
Value field.
Click OK.
Repeat steps a through e for the A2, B1, B2, C1, C2, D1 and D2 service
groups.
End of Solution
Set the Load attribute for each service group as shown in the following table.
Service Group
Load
A1
75
A2
75
B1
75
B2
75
C1
50
C2
50
D1
50
D2
50
Solution
219
From the Edit Attribute window, in the Scaler Value field, type: 75
Click OK.
Repeat steps a through g for the A2, B1, B2, C1, C2, D1 and D2 service
groups using the Load value shown in the table above.
End of Solution
B75
Copyright 2012 Symantec Corporation. All rights reserved.
From the pop-up window, in the Scaler Value field, type: 200
Click OK.
End of Solution
AvailableCapacity
S1
50
S2
50
S3
S4
Solution
220 B76
Verify that the AvaliableCapacity is set to the value shown in the table
above.
End of Solution
Verify that all service groups are online as shown in the following table.
Service Group
System S1
A1
Online
A2
Online
System S2
B1
Online
B2
Online
System S3
C1
Online
C2
Online
System S4
D1
Online
D2
Online
Solution
221
Select wlm.
Verify the service groups are online as shown in the above table.
End of Solution
If it faults, where will the A1 service group fail over to? Verify this failover by
faulting a critical resource in the A1 service group.
Service Group A1 will fail over to:
S2
Solution
From left pane, expand wlm > A1 > FileOnOff > A1File.
B77
Copyright 2012 Symantec Corporation. All rights reserved.
From the right pane, notice that the A1 service group shows faulted on S1,
but Online on S2.
End of Solution
AvailableCapacity
S1
125
S2
-25
S3
S4
Solution
222 B78
Verify that the AvaliableCapacity is set to the value shown in the table
above.
End of Solution
10 If the S2 system fails, where will the A1, B1 and B2 service groups that are
Online on S2 fail over? Verify this failover by powering off the S2 system in
Cluster Manager.
The A1 service group will fail over to:
S3
S1
S1
Solution
From the left pane, right-click wlm > S2 and select on Power Off.
Select wlm.
Notice that the A1 service group shows faulted on S1, but Online on S3.
End of Solution
following table.
Service Group
AvailableCapacity
S1
-25
S2
200
S3
-75
S4
Solution
Verify that the AvaliableCapacity is set to the value shown in the table
above.
B79
Copyright 2012 Symantec Corporation. All rights reserved.
End of Solution
12 Power up the S2 system, clear all faults, and return the service groups to their
startup locations.
Solution
From the left pane, right-click wlm > S2 and select Up > S2.
Solution
following table.
224 B80
Service Group
AvailableCapacity
S1
50
S2
50
S3
S4
Solution
Verify that the AvaliableCapacity is set to the value shown in the table
above.
End of Solution
End of Solution
B81
Copyright 2012 Symantec Corporation. All rights reserved.
winclient
1
From the Java Console, open the cluster configuration for update.
Solution
Set the Limits attribute for each system to the key ABGroup with a value of 3.
Solution
226 B82
From the Edit Attribute window, click the add an element button.
Click OK.
End of Solution
Set the Prerequisites attribute for the A1, A2, B1 and B2 service groups to be
ABGroup with a value of 1.
CAUTION Do not modify the Prerequisites attribute for the C1, C2, D1,
and D2 service groups.
Solution
From the Edit Attribute window, click the add an element button.
Click OK.
End of Solution
B83
Copyright 2012 Symantec Corporation. All rights reserved.
If the S1 system fails, where will the A1 and A2 service groups that are Online
on S1 fail over? Verify this failover by powering off the S1 system in Cluster
Manager. Explain the failover actions that occurred.
The A1 service group will fail over to:
S2
S3
A1 failed over to S2, but A2 failed over to S3 because the limit was reached on S2.
Solution
Select wlm.
From the left pane, right-click wlm > S1 and select on Power Off.
Select wlm.
End of Solution
228 B84
If the S2 system fails, where will the A1, B1 and B2 service groups that are
Online on S2 fail over? Verify this failover by powering off the S2 system in
Cluster Manager. Explain the failover actions that occurred.
The A1 service group will fail over to:
S4
S3
S4
A1 failed over to S4, B1 failed over to S3, and B2 failed over to S4 due to the values
of the Load attribute on each system.
Solution
From the left pane, right-click wlm > S2 and select on Power Off.
Select wlm.
End of Solution
If the S3 system fails, where will the A2, B1, C1 and C2 service groups that
are Online on S3 fail over? Verify this failover by powering off the S3 system
in Cluster Manager. Explain the failover actions that occurred.
The A2 service group will fail over to:
S4
S4
S4
All service groups fail over to S4 except B1. B1 is the last group to attempt to fail over
to S4. Since B1 has a prerequisite that cannot be met on S4, it cannot fail over. A1,
A2, B1, and B2 can run on the same system. B1 stays offline.
Solution
From the left pane, right-click wlm > S3 and select on Power Off.
Select wlm.
B85
Copyright 2012 Symantec Corporation. All rights reserved.
End of Solution
End of Solution
230 B86
winclient
1
End of Solution
231
End of Solution
B87
Copyright 2012 Symantec Corporation. All rights reserved.
Click the Close button that appears as an X in the upper right hand corner
of the window.
End of Solution
End of lab
232 B88
B89
234 B90
Verify that the following virtual machines shown are powered on, that you are
logged using the indicated account and that the indicated terminal windows are
opened.
mgt
There is no need to log into this virtual machine at this time.
sym1
Log in with credentials:
Account: root
Password: veritas
Two terminal windows should be active on this server.
sym2
Log in with credentials:
Account: root
Password: veritas
Two terminal windows should be active on this server.
sym3
Log in with credentials:
Account: root
Password: veritas
Two terminal windows should be active on this server.
sym4
Log in with credentials:
Account: root
Password: veritas
Two terminal windows should be active on this server.
Note: The terminal windows are referred to as hostname:terminal#
throughout the labs. For example: sym1:terminal1, sym1:terminal2,
sym2:terminal1, and sym2:terminal2. It is not necessary to label the
terminal windows and you may decide which is terminal1 and which is
terminal2.
If you have machines running that are not used in this lab, shut down the
operating system and power the machines off.
Note: If you are completing the lab exercises in order, this will require you to
shutdown winclient, and power on and log into sym1, sym2, sym3
and sym4.
B91
sym1
1
haconf -makerw
End of Solution
Solution
236 B92
Modify the SystemList to allow the netsg service group to run on the sym1,
sym2, sym3, and sym4 in that order.
Solution
Modify the AutoStartList attribute to allow the service group to start on all
systems.
Solution
Modify the Parallel service group to configure the netsg service group as a
parallel service group.
Solution
Display the service group attributes for the netsg service group to confirm your
input and key default service group attribute values.
Solution
Attribute
AdministratorGroups
Administrators
Authority
AutoFailOver
AutoRestart
AutoStart
AutoStartIfPartial
AutoStartList
AutoStartPolicy
ClusterFailOverPolicy
System
global
global
global
global
global
global
global
global
global
global
Value
Parallel
global
SystemList
sym4
global
0
1
1
1
1
sym1
Order
Manual
sym1
sym2
sym3
sym4
sym2
End of Solution
B93
haconf -dump
End of Solution
Add a resource of type NIC named netnic to the netsg service group.
Solution
End of Solution
Modify the Device and Critical resource attributes for the netnic resource
using the following information.
Device is eth0
Critical set to 0 (zero)
Solution
End of Solution
10 Display the resource attribute values for the netnic resource to confirm your
238 B94
Solution
Attribute
Group
Type
AutoStart
Critical
Enabled
System
global
global
global
global
global
Value
netsg
NIC
1
0
0
Device
global
eth0
End of Solution
11 Enable the netnic resource and verify that it is enabled. Display the state of the
Attribute
State
State
State
State
System
sym1
sym2
sym3
sym4
Value
ONLINE
ONLINE
ONLINE
ONLINE
End of Solution
Note: The service group state is offline even though the NIC resource state is
online. This is because VCS does not consider persistent resources
when determining the state of a service group.
Solution
Attribute
State
State
State
State
System
sym1
sym2
sym3
sym4
Value
|OFFLINE|
|OFFLINE|
|OFFLINE|
|OFFLINE|
End of Solution
haconf -dump
End of Solution
B95
group.
Solution
End of Solution
15 Set the Critical resource attribute to 0 (zero) for the netphantom resource.
Solution
16 Display the resource attribute values for the netphantom resource to confirm
Attribute
Group
Type
AutoStart
Critical
Enabled
System
global
global
global
global
global
Value
netsg
Phantom
1
0
0
End of Solution
240 B96
17 Enable the netphantom resource and verify that it is enabled. Display the state
Attribute
State
State
State
State
System
sym1
sym2
sym3
sym4
Value
ONLINE
ONLINE
ONLINE
ONLINE
End of Solution
Note: The service group state is offline even though the NIC resource state is
online. This is because VCS does not consider persistent resources
when determining the state of a service group.
Solution
Attribute
State
State
State
State
System
sym1
sym2
sym3
sym4
Value
|ONLINE|
|ONLINE|
|ONLINE|
|ONLINE|
End of Solution
haconf -dump
End of Solution
B97
sym1
1
From sym1:terminal1, identify the names of the resources of type NIC that are
configured.
Solution
sym1
sym2
sym1
sym2
sym3
sym4
sym1
sym2
sym3
sym4
sym1
sym2
sym1
sym2
End of Solution
242 B98
Ignoring the resource of type NIC named netnic, list the resource
dependencies associated with the listed resources of type NIC.
Solution
Parent
notifier
webip
appip
nfsip
oraip
Child
csgnic
csgnic
appnic
nfsnic
oranic
End of Solution
Replace the resource of type NIC named appnic with a resource of type Proxy
named appproxy using the following information.
Solution
Attribute
Group
Type
AutoStart
Critical
Enabled
System
global
global
global
global
global
Value
appsg
Proxy
1
0
0
TargetResName
global
netnic
Attribute
State
State
System
sym1
sym2
appip
Value
ONLINE
ONLINE
appproxy
End of Solution
B99
Replace the resource of type NIC named csgnic with a resource of type Proxy
named csgproxy using the following information.
Solution
Attribute
Group
Type
AutoStart
Critical
Enabled
System
global
global
global
global
global
Value
ClusterService
Proxy
1
0
0
TargetResName
global
netnic
244B100
Attribute
State
State
State
State
System
sym1
sym2
sym3
sym4
Value
ONLINE
ONLINE
ONLINE
ONLINE
csgproxy
csgproxy
End of Solution
Replace the resource of type NIC named nfsnic with a resource of type Proxy
named nfsproxy using the following information.
Solution
VCS NOTICE V-16-1-10242 Resource added. Enabled attribute must be set before
agent monitors
Attribute
Group
Type
AutoStart
Critical
Enabled
System
global
global
global
global
global
Value
nfssg
Proxy
1
0
0
TargetResName
global
netnic
Attribute
State
State
System
sym1
sym2
Value
ONLINE
ONLINE
B101
nfsip
nfsproxy
End of Solution
Replace the resource of type NIC named oranic with a resource of type Proxy
named oraproxy using the following information.
Solution
VCS NOTICE V-16-1-10242 Resource added. Enabled attribute must be set before
agent monitors
246 B102
Attribute
Group
Type
AutoStart
Critical
Enabled
System
global
global
global
global
global
Value
orasg
Proxy
1
0
0
TargetResName
global
netnic
Attribute
State
State
System
sym1
sym2
Value
ONLINE
ONLINE
oraip
oraproxy
End of Solution
End of lab
B103
248 B104
Appendix C
Supplemental Content
249
250 C2
251
The following steps describe what happens when a service group in a service
group dependency relationship is faulted due to a critical resource fault:
1 The entire service group is taken offline due to the critical resource fault
together with any of its parent service groups that have an online firm or hard
dependency (online local firm, online global firm, online remote firm, or
online local hard).
2 Then a failover target is chosen from the SystemList of the service group based
on the failover policy and the restrictions brought by the service group
dependencies. Note that if the faulted service group is also the parent service
group in a service group dependency relationship, the service group
dependency has an impact on the choice of a target system. For example, if the
faulted service group has an online local (firm or soft) dependency with a child
service group that is online only on that system, no failover targets are
available.
3 If there are no other systems to which the service group can fail over, both the
child service group and all of the parents that were already taken offline remain
offline.
4 If there is a failover target then VCS takes any child service group with an
online local hard dependency offline.
5 VCS then checks if there are any conflicting parent service groups that are
already online on the target system. These service groups can be parent service
groups that are linked with an offline local dependency or online remote soft
dependency. In either case, the parent service group is taken offline to enable
the child service group to start on that system.
6 If there is any child service group with an online local hard dependency, first
the child service group and then the service group that initiated the failover are
brought online.
7 After the service group is brought online successfully on the target system,
VCS takes any parent service groups offline that have an online local soft
dependency to the failed-over child.
8 Finally, VCS selects a failover target for any parent service groups that may
have been taken offline during steps 1, 5, or 7 and brings the parent service
group online on an available system.
9 If there are no target systems available to fail over the parent service group that
has been taken offline, the parent service group remains offline.
C3
252 C4