This product is protected by United States and international copyright laws. The product’s underlying
technology, patents, and trademarks are listed at http://www.parallels.com/trademarks.
Microsoft, Windows, Windows Server, Windows NT, Windows Vista, and MS-DOS are registered trademarks of
Microsoft Corporation.
Linux is a registered trademark of Linus Torvalds.
Mac is a registered trademark of Apple, Inc.
All other marks and names mentioned herein may be trademarks of their respective owners.
Setting Up a Domain.................................................................................................................................. 14
Creating a Domain.......................................................................................................................... 15
Adding Nodes to the Domain..........................................................................................................22
Connecting Servers to a Shared Storage.................................................................................................... 24
Setting Up a Cluster................................................................................................................................... 25
Creating a Cluster Service Account................................................................................................ 26
Deploying a Failover Cluster in Windows Server 2008 and Windows Server 2008 R2................. 27
Deploying a Failover Cluster in Windows Server 2003 ................................................................. 42
Installing Parallels Containers on Cluster Nodes....................................................................................... 65
Setting Up a SAN-Based Cluster............................................................................................................... 77
Setting Up a 'Loopback File'-Based Cluster............................................................................................... 79
Setting Up an NLB Cluster 83
Index 87
4
Microsoft Cluster Types
The Windows Server operating system comes with two main built-in clustering technologies:
Failover cluster
Microsoft Network Load Balancing (NLB)
Failover Clusters
Failover clusters (known as server clusters in Windows Sever 2003 and earlier Windows Server
versions) allow you to provide high availability for your mission-critical applications (including
the Parallels Containers software itself) through the failover of resources. In Parallels
Containers, you can deploy failover clusters in one of the following ways:
Create a failover cluster of two or more Hardware Nodes that will host the Parallels
Containers software including all its mission-critical services and Containers. If one
Hardware Node fails, another Node will take its responsibilities.
Make Containers participate as full members in the cluster, like any other standalone server
running the Windows Server operating system, and increase the availability of your missioncritical applications.
The process of deploying failover clusters in Parallels Containers-based systems is described in
Deploying Parallels Containers Failover Clusters (p. 5) and Creating Clusters From Containers (p.
76).
Note: Throughout this guide, we use the term failover cluster to refer to the clustering
technology in Windows Server 2008 R2, Windows Server 2008, and Windows Server 2003
though in Windows Server 2003 this technology is called server cluster. When there are
differences between these two clustering models, we describe them in separate sections.
NLB Clusters
The NLB technology allows you to provide scalability and high availability for TCP- and UDPbased services and applications (e.g., for Web servers and Terminal Services) by load-balancing
incoming IP traffic across several servers in the cluster. For detailed information on deploying
NLB clusters, see Setting Up an NLB Cluster (p. 83).
5
C HAPTER 1
Deploying Parallels Containers Failover
Clusters
In Parallels Containers 4.6, you can use the Windows failover cluster feature to create failover
clusters consisting of two or more Hardware Nodes. Including Hardware Nodes in a failover
cluster increases the availability of the Parallels Containers software and Containers. If one
Hardware Node in the cluster fails, another Node will take its responsibilities.
The process of deploying a Parallels Containers failover cluster includes the following steps:
1 Choosing the cluster model that best suits your needs.
2 Setting up hardware for the cluster.
3 Configuring all physical servers to participate in the cluster (hereinafter, we will also refer to
them as cluster nodes or nodes).
4 Setting up an Active Directory domain and adding the cluster nodes to it.
Note: You can also add the cluster nodes to an existing Active Directory domain, if you
have one.
5 Setting up a shared storage and connecting it to the cluster nodes.
6 Creating a failover cluster and adding all your nodes to the cluster.
7 Setting up the failover cluster.
All these steps are described in the following subsections in detail.
In This Chapter
Designing a Cluster Configuration........................................................................................ 6
Setting Up Hardware for the Cluster.....................................................................................12
Before you start deploying a Parallels Containers failover cluster, you should decide on the
cluster configuration that will meet your demands best of all. You can create Parallels
Containers failover clusters of one of the following types:
Active/passive clusters. An active/passive cluster includes both active and standby nodes. In
this type of clusters, standby nodes are only used if one of the active nodes fails.
Active/active clusters. An active/active cluster consists of active nodes only, each running
the Parallels Containers software and hosting a number of Containers. In the event of a
failover, all Containers running on the problem node are failed over to one of the healthy
active nodes.
The following subsections discuss both types of failover clusters in detail.
Active/passive clusters include both active and standby nodes. An example of an active/passive
cluster is a four-node cluster where three nodes are active, running the Parallels Containers
software and hosting a number of Containers, and one node is in the standby state, waiting for
an active node failure and ready to take the responsibility of its resources. An active/passive
cluster can also consist of five active cluster nodes and two standby nodes where each of the
active nodes is able to fail over to any of the two standby nodes.
The following picture shows an example configuration of an active/passive failover cluster:
Note: We will use the given configuration throughout this chapter to demonstrate the Parallels
Containers failover cluster creation.
In this clustering scenario:
The cluster comprises two active nodes (Active Node #1 and Active Node #2) hosting a
number of Containers and one passive node (Standby Node). All nodes in the cluster have
one and the same version of the Windows Server operating system and the Parallels
Containers software installed. If one of the active nodes in the cluster fails or is taken offline
as part of planned maintenance, the services and Containers from this node are failed over to
the standby node.
All the nodes in the cluster are connected via Fibre Channel optical cables to a reliable and
fault tolerant shared SCSI storage device. It is used to store all the cluster configuration data
and Parallels Containers resources and allows active nodes in the cluster to access these
resources.
All the cluster nodes are located in the same Active Directory (AD) domain. A dedicated
server is set up in this domain to act as the domain controller (shown as Domain Controller
in the picture).
Note: In our example, we will create an Active Directory domain from scratch. However,
you can make use of an existing AD domain and include all your cluster nodes in this
domain.
Parallels Containers allows you to create active/active failover clusters where all Hardware
Nodes are acting as active nodes. The concept of active/active clusters is based on the ability to
simultaneously run more than one group with Parallels Containers resources (i.e. with Parallels
Containers-related services and Containers) on one and the same Hardware Node. In such
clusters, each node is capable of taking on the additional workload, the failed over services and
Containers, from a problem node along with running their own version of Parallels Containers
and hosting their own Containers.
The following picture shows you an example of an active/active cluster configuration:
This clustering configuration is similar to that described in the previous section for an
active/passive cluster. The cluster comprises three nodes: Active Node #1, Active Node #2,
Active Node #3. All nodes are running one and the same version of the Windows Server
operating system and the Parallels Containers software, located in the same AD domain, and
connected via Fibre Channel optical cables to a shared SCSI storage device.
However, unlike the active/passive configuration, all nodes in this cluster are active and host a
number of running Containers. If any of the nodes in the cluster fails or is taken offline as part
of planned maintenance, the Parallels Containers-related services and Containers from this node
are automatically failed over to one of the two remaining active nodes. If, say, Active Node #1
hosting Container 101, 102, and 103 fails, all these Containers will be moved to and started on
either Active Node #2 or Active Node #3. The decision on where to move the resources from the
problem node is made by the clustering software.
Setting Active/Active Failover Cluster to Work
The process of deploying an active/active Parallels Containers failover cluster slightly differs
depending on whether you are creating your failover cluster from scratch or transforming any of
your existing active/passive clusters:
If you are creating a new Parallels Containers failover cluster, follow the instructions given
in the following subsections to set up the cluster.
If you already have a Parallels Containers active/passive cluster and want to transform it
into an active/active one, do the following:
1. Reinstall Parallels Containers on each passive node in the cluster by deleting the current
software version and installing it again using the instructions in the Installing Parallels Containers on Cluster Nodes section (p. 65).
Note: When reinstalling the Parallels Containers software, you must select the Active node
radio button for each node that will participate in the active/active cluster.
2. Install all available Parallels Containers updates. Refer to the Parallels Containers 4.6
Important! We highly recommend that you install the same set of Parallels Containers and
Windows Server updates on all nodes in the Parallels Containers failover cluster.
Active/Active Failover Cluster Restrictions
Before starting to deploy an active/active Parallels Containers failover cluster, you should have
a clear understanding of all limitations imposed on your virtual infrastructure by this kind of
cluster:
Parallels Containers failover clusters do not currently have a mechanism to correctly process
Containers with identical IDs. So, you should always see to it that all Containers residing on
all nodes in the cluster have different IDs.
Containers failed over from a problem node to a healthy one make use of the OS and
application templates installed on the healthy node. So, to provide an error-free operation of
your Containers, make sure that all nodes in the cluster have the same set of OS and
application templates installed.
Deploying an active/active cluster requires the proper hardware configuration of all cluster
nodes. As the resources from a problem node may be failed over to any of the remaining
active nodes, all nodes in the active/active cluster must have enough resources (CPU, main
memory, etc.) to take on the additional workload from a failed node and ensure adequate
performance in the case of a failover.
If any of the nodes does not have sufficient resources to support this additional workload
(e.g. the node's own services and Containers consume the majority of system resources), it is
recommended to deploy an active/passive failover cluster to avoid degradation in your
overall cluster performance.
Once you know which cluster model to deploy, you can start setting up hardware for your
cluster. This procedure includes the following basic operations:
1 Deciding on the hardware to be included in the cluster. The amount and type of hardware
may vary depending on the purpose and availability requirements of your cluster. However,
the following types of hardware are always present in any cluster:
Cluster nodes represented by two or more physical servers capable of running the
Windows Server operating system.
Shared storage device usually represented by a common disk array on a SAN and used
to store all Parallels Containers-related data (Container data, templates, backups, and so
on).
Switches (Fibre Channel or SCSI) providing client access to the cluster and enabling the
communication between each cluster node and the shared cluster storage.
2 Uniting all hardware components into one subnet and ensuring that each hardware
component can access all the other components in the subnet.
For more information about installing and configuring cluster hardware, follow these links:
for Windows Server 2003:
To deploy our example cluster configuration that will include two active nodes and one passive
node, we use the following hardware:
A shared SCSI storage device.
A switch.
Fibre optic cables.
Four physical servers capable of running the Windows Server operating system. Three
servers will act as cluster nodes and one server will be used as a domain controller.
All the physical servers in our example meet the following hardware requirements:
Pentium IV processor
2048 MB of memory
80 GB of disk space
two network interface cards (NICs) for each cluster server and one network card for the
After you have prepared the hardware for your failover cluster, you need to configure three
physical servers for participating in the cluster and one physical server for converting into a
domain controller. To do this, perform the following operations:
1 Install the Windows Server operating system on each server (mandatory). The process of
installing Windows Server is described in detail in the documentation shipped with this
operating system.
Note: All cluster nodes must run the same version of the operating system, for example, the
64-bit version of Windows Server 2008 (Enterprise Edition).
2 Install the failover clustering feature on the physical server (mandatory). This operation
must be performed only on physical servers running Windows Server 2008 and Windows
Server 2008 R2. For more information, see http://technet.microsoft.com/enus/library/cc772178.aspx.
3 Configure the network adapters on all physical servers (mandatory):
For each cluster node:
Assign a static IP address to the first network adapter. This adapter will be used to
connect the server to the private network providing communication among all nodes in
the cluster.
Assign a static IP address to the second network adapter. This adapter will be used to
connect the server to the public network (usually, your local area network). In
comparison to the IP address assigned to the first network adapter, this IP address must
belong to a different subnet.
For the domain controller:
Assign a static IP address to the network adapter. This adapter will be used to connect
the domain controller to other servers in the Active Directory domain.
4 Validate your physical server configuration (optional though highly recommended). This
operation can be performed only on physical servers running Windows Server 2008 or
Windows Server 2008 R2. For more information, see http://technet.microsoft.com/enus/library/cc732035(WS.10).aspx.
In the next step, you need to set up an Active Directory domain and make your servers members
of this domain. This procedure includes the following steps:
1 Deploying an Active Directory domain.
2 Creating a domain administrator account.
3 Adding the cluster nodes to the domain.
All these steps are described in the following subsections in detail.
Notes:
1. If you already have an existing Active Directory domain, you can add your cluster nodes to
this domain. In this case you can skip Steps 1 and 2 and proceed directly to Step 3.
2. Throughout this section, we assume that you are setting up a domain on a server running
Windows Server 2008 R2. The steps you need to perform to set up a domain on a server running
Windows Server 2003 or Windows Server 2008 may slightly differ. For information on creating
and managing domains under these operating systems, refer to their documentation.
First, you have to set up an Active Directory domain. To do this, you need a dedicated physical
server capable of running the Windows Server operating system. This server will act as a
domain controller and have the Active Directory service installed. The process of setting up an
Active directory domain includes two steps:
1 Installing the Active Directory Domain Controller role.
2 Creating the Active Directory domain.
Installing the Active Directory Controller role
In the first step, you need to install the Active Directory Domain Controller role on the physical
server that is to act as the domain controller. By default, this role is not installed on servers
during the Windows Server 2008 R2 installation. To install the role:
1 Open the Server Manager console.
2 Click the Roles item in the left pane of the console.
3 In the Roles pane, click Add Roles to launch the Add Roles wizard.
9 In the Directory Services Restore Mode Administrator Password window, specify an arbitrary
password for the Directory Services Restore Mode administrator. You will need this
password to boot the domain controller in Directory Services Restore Mode (e.g., if the
Active Directory service has failed or needs to be restored). Make sure the password meets
the Windows length and complexity requirements. Click Next.
10 The last window allows you to review the parameters provided by you in the previous steps
of the wizard. To modify any parameters, click Back; otherwise, click Next.
11 A window will open showing you the operation progress. Select the Reboot on completion
check box to automatically restart the server once the Active Directory domain is created.
For more information on creating and configuring Active Directory domains, refer to
http://support.microsoft.com/kb/324753/en-us.
Now that you have created the Active Directory domain and the domain administrator account,
you can add your nodes to the domain. To do this:
1 Log in to the first node you want to add to the domain, and click Start > Control Panel >
System and Security > System.
2 In the Computer name, domain, and workgroup settings section, click Change settings.
3 On the Computer Name tab of the displayed window, click Change.
4 In the Computer Name Changes window, do the following:
In the Computer name field, specify a server hostname. This name will be used to
uniquely identify the given node among other nodes in the cluster. By default, you are
offered to use the hostname assigned to the node during the Windows Server installation.
However, we recommend that you change this hostname to something more descriptive
(e.g., CLUSTERNODE1).
Select the Domain radio button and type the domain DNS name (you specified this name
during the Active Directory domain). In our example, the domain DNS name must be set
to mycompany.local.
After providing the necessary information, your window should look like the following:
Next, you should configure the shared storage and connect it to all your cluster nodes. The way
of configuring storage devices and connecting them to cluster nodes varies from vendor to
vendor. So we recommend that you refer to the vendor’s guidelines and recommendations to
properly set up your storage.
For your reference, we list below the general steps that you usually have to perform to configure
a shared SCSI device:
1 Power off all your cluster nodes.
2 Prepare the necessary controllers for the shared storage bus.
3 Connect the SCSI device to the shared bus, and then join the bus to your cluster nodes.
4 Power on the first cluster node.
5 Format all the disks on the cluster storage device as NTFS, and configure them as basic
disks.
6 Partition the disks on the cluster storage device.
Pay special attention to the disk to be allocated to the cluster quorum resource. The quorum
resource maintains the cluster configuration data; so, we highly recommend that you use a
separate disk (ca. 500 MB in size) for this resource.
7 Assign drive letters to the disks on the cluster storage device.
Note: All the disks on the cluster storage device must be assigned drive letters. If a disk does
not have a letter, it cannot be added to a clustered service (known as a resource group in
Windows Server 2003) intended for hosting Parallels Containers resources. For information
on clustered services, see Configuring the Cluster (p. 35, p. 51).
For more information about configuring shared storage devices in cluster environments, refer to
http://technet2.microsoft.com/windowsserver/en/library/1bce08cc-2caf-417c-b53d035f57c4e78b1033.mspx?mfr=true.
Now you are ready to set up a failover cluster that will ensure the high availability of your
Parallels Containers installations. The process of setting up this kind of cluster includes the
following steps:
1 Creating a Cluster Service user account on the domain controller.
Note: Creating a Cluster service account is not necessary if your failover cluster will include
only nodes running Windows Server 2008 and Windows Server 2008 R2.
2 Creating the failover cluster and configuring the first cluster node.
3 Configuring all the remaining nodes in the cluster.
These steps are described in the following subsections in detail.
Before creating a failover cluster, you need to create a domain user account under which the
Cluster service (this service is responsible for managing all cluster-related activities) will run.
You will need this account later on when making the failover cluster and joining your nodes to
it.
Creating a Cluster service account is not necessary if your failover cluster will include only
nodes running Windows Server 2008 and Windows Server 2008 R2. So, the example below
demonstrates how to create a Cluster service account in Windows Server 2003:
1 Log in to the domain controller.
2 Click Start, point to Administrative Tools, and select Active Directory Users and Computers.
3 In the left pane of the Active Directory Users and Computers window, expand the contents of
the domain.
4 Right-click the Users folder, point to New, and select User.
5 In the displayed window, do the following:
Type arbitrary names in the First name and Last name fields (the choice of these names
is not important for the account being created).
In the User logon name field, type a descriptive name that will be used to log in to the
6 Specify a password for the domain account, and click Next.
7 The last screen allows you to review the parameters provided in the previous steps. To
modify any settings, click Back; otherwise, click Finish to create the Cluster service account.
Deploying a Failover Cluster in Windows Server 2008 and Windows
Server 2008 R2
This section describes the process of setting up a failover cluster on servers running Windows
Server 2008 and Windows Server 2008 R2. For information on deploying a failover cluster on
servers running Windows Server 2003, see Deploying a Failover Cluster in Windows Server 2003
(p. 42).
Loading...
+ 60 hidden pages
You need points to download manuals.
1 point = 1 manual.
You can buy points or you can get point for every manual you upload.