Consider common installation scenarios for Network Deployment to
determine how to install your application serving environment. This topic
links to procedures for installing each scenario.
Before you begin
IBM WebSphere Application Server Network Deployment is an integrated
platform that contains an application server, Web development tools, a Web
server, and additional supporting software and documentation.
In
Version 6.0, installing WebSphere Application Server Network Deployment is
a two-step process. The first step is using the installation wizard to install
a shared set of core product files. The second step is using the Profile Creation wizard to
create a deployment manager profile, a stand-alone
application server profile, or a custom profile.
A profile is a separate data partition
that includes the files that define a runtime environment for an application
server process, such as a deployment manager or an application server.
A necessary part of installation and configuration
is the creation of profiles, which is shown as step 2 in the following diagram.
You must create a deployment manager profile, for example, to have a functional
cell.
A
running application server process can create, read, update, or delete the
configuration files, data files, and log files in its profile. The application
server process can access the system files, which include command files and
other shared product binary files. However, most system files are updated
only by installing refresh packs, fix packs, interim fixes, or products that
extend Network Deployment.
Avoid trouble: It
is suggested that you configure WebSphere Application Server Network Deployment
with a single subnet for network traffic. You can use one Network interface
card (NIC) on a physical machine or logical partition (LPAR). You can also
reference a single Domain name system (DNS) server in the network configuration
for the physical machine or LPAR.
gotcha
Scenarios
for installation
The following information describes scenarios for
installing the product in various topologies on one or more machines. Two
types of WebSphere Application Server topologies are possible using the Network
Deployment product:
- Topologies for a stand-alone application server
- Each stand-alone application server has its own administrative console
and runs independently of other application servers.
The following topologies
are described in this topic.
- Topologies for a managed group of application servers in a cell
- A cell consists of one deployment manager and one or
more federated application servers that are managed nodes.
A
node can become a managed node in several ways:
- By federating the node within an application server profile into the cell
- By federating the node within a custom profile into the cell
The deployment manager is the single point of administration for
all of the managed nodes in the cell. The deployment manager maintains the
configuration files for nodes that it manages and deploys applications to
those managed nodes.
The following topologies for a cell are described
in this topic.
Scenarios 5, 6, 7, and 8 assume that all nodes in a cell reside
on a particular machine and operating system. However, this precise node assignment
does not need to apply. The deployment manager node can exist on Machine A,
other managed nodes (that have been federated into the deployment manager)
can exist on differing machines and operating systems. Such a configured cell
of differing machines or operating systems is called a heterogeneous cell and
expands the possible topologies you can consider for your network deployment.
About this task
Each of the following scenarios includes a diagram and a list
of detailed installation steps.
Some scenarios are more typical in
production environments. For example, Scenario 1 supports a lighter workload
than Scenario 3 or Scenario 4. However, Scenario 1 is a fully functional environment.
Scenarios 3 and 4 are typical production environments for a stand-alone application
server. Scenario 8 is a typical production scenario for a cell environment.
- Scenario 1: Install a stand-alone application server
on a single machine.
Installing WebSphere Application Server
Network Deployment by itself on a single machine lets you create a stand-alone
Application Server profile. Each stand-alone application server profile includes
a server1 application server process. Installing Network Deployment creates
the set of system files.
The Profile creation wizard creates the profile for
the application server. The profile is a separate data partition with files
that define the stand-alone application server environment.
In this
scenario, the application server uses its internal HTTP transport chain for
communication instead of a using a separate web server (on a separate machine)
to possibly offload some processing.
- Install WebSphere Application Server Network Deployment.
- Create an Application Server profile
using the Profile Creation wizard.
- Scenario 2: Install a stand-alone application server
and a Web server on a single machine.
Installing a Web server,
such as IBM HTTP Server, on the same machine as the application server provides
more configuration options. Installing a Web server plug-in enables the Web
server to communicate with the application server. This installation scenario
supports rigorous testing environments or production environments that do
not require a firewall. However, this scenario is not a typical production
environment. When everything is on one machine, neither the web server or
the application server will run as fast as if they were on separate machines
because they are both competing for the same CPU resources.

- Install WebSphere Application Server Network Deployment.
- Create a stand-alone application server
profile using the Profile Creation wizard.
- Install IBM HTTP Server or another supported
Web server.
- Install the Web server plug-ins and configure
the Web server using the Plug-ins installation wizard.
- Scenario 3: Install a stand-alone application server
and a Web server on separate machines.
In the typical production
environment, the application server on one machine communicates with a Web
server on a separate (remote) machine through the Web server plug-in. Optional
firewalls can provide additional security for the application server machine.
- Install WebSphere Application Server Network Deployment on Machine
A.
- Create a stand-alone application server
profile on Machine A using the Profile creation wizard.
- Install IBM HTTP Server or another supported
Web server on Machine B.
- On Machine B, install the Web server
plug-ins and configure the Web server using the Plug-ins installation wizard.
- The Plug-ins installation wizard creates
a script named configureWeb_server_name in
the plugins_root/bin directory
on Machine B. Copy the script to the app_server_root/bin directory
on Machine A.
- Run the configureWeb_server_name script
to create a Web server definition in the administrative console. You can then
use the administrative console to manage the Web server.
- Propagate the plugin-cfg.xml file
from the application server to the Web server using the administrative console.
Click Servers > Web server > Propagate Plug-in.
(Web servers other than IBM HTTP Server require manual propagation.)
- Scenario 4: Install multiple stand-alone application
servers on one machine and a Web server on a separate machine.
The Profile Creation wizard can create a deployment
manager profile, an application server profile, or a custom profile. Each
profile is a separate data partition containing the files that define the
run-time environment. After creating a profile and installing a dedicated
Web server, use the Plug-ins installation wizard to install a plug-in and
to update the Web server configuration file. The Web server can then communicate
with the application server.
This topology lets each profile have unique
applications, configuration settings, data, and log files, while sharing the
same set of system files. Creating multiple profiles creates multiple application
server environments that you can dedicate to different purposes.
For
example, each application server on a Web site can serve a different application.
In another example, each application server can be a separate test environment
that you assign to a programmer or a development team.
Updating
the core product files
Another feature of having
multiple profiles is enhanced serviceability. When a refresh pack or a fix
pack updates the core product files on a machine, all of the application server
profiles that were created from the core product files begin using the updated
files. However, the profile itself is not updated unless the maintenance package
contains required maintenance for existing profiles. The Update Installer
issues a message for a required update to an existing profile as you install
the maintenance package.
In some situations, you
might prefer to not update all of the application servers on a machine. In
such situations, simply install the product a second time and install the
new maintenance. Create application server profiles from both installations
to manage the product updates incrementally.

- Install WebSphere Application Server Network Deployment on Machine
A.
- Create the first stand-alone application
server profile on Machine A using the Profile creation wizard.
- Install IBM HTTP Server or another supported
Web server on Machine B.
- On Machine B, install the Web server
plug-ins and configure the first Web server using the Plug-ins installation
wizard.
- The Plug-ins installation wizard creates
a script named configureWeb_server_name in
the plugins_root/bin directory
on Machine B. Copy the script to the app_server_root/bin directory
on Machine A.
- Run the configureWeb_server_name script
to create a Web server definition in the administrative console. You can then
use the administrative console to manage the Web server.
- Propagate the plugin-cfg.xml file
from the application server to the Web server using the administrative console.
Click Servers > Web server > Propagate Plug-in.
(Web servers other than IBM HTTP Server require manual propagation.)
- Create the second Application Server
profile using the Profile Creation wizard on
Machine A. Make the profile the default profile during the profile creation
by selecting the check box on the appropriate panel.
The script
that the Plug-ins installation wizard creates works on the default profile
only. So, this script can only create a Web server definition on the profile
that is the default profile at the time that the script runs.
- Install a second IBM HTTP Server or another
supported Web server on Machine B.
- On Machine B, install the Web server
plug-ins to configure the second Web server using the Plug-ins installation
wizard. Both Web servers share a single installation of the plug-in binaries
but must be configured individually.
- The Plug-ins installation wizard creates
a script named configureWeb_server_name for
the second Web server. The script is in the plugins_root/bin directory
on Machine B. Copy the script to the app_server_root/bin directory
on Machine A.
- Run the configureWeb_server_name script
to create a Web server definition in the administrative console. You can then
use the administrative console to manage the Web server.
- Propagate the plugin-cfg.xml file
from the second application server to the Web server using the administrative
console. Click Servers > Web server > Propagate Plug-in.
(Web servers other than IBM HTTP Server require manual propagation.)
- Scenario 5: Install a cell
of managed application server nodes on one machine.
WebSphere
Application Server Network Deployment can create a cell of managed application
servers on a single machine from one installation of the core product files.
The
Profile creation wizard creates the deployment manager. After starting the
deployment manager, use the Profile Creation wizard to
create one or more application servers for the cell. Application server profiles
have a default application server, called server1, and default applications.
An Application Server node becomes a managed node after federating the node
into the deployment manager cell.
The deployment manager provides the
administration for all managed nodes that are in its cell. Periodically the
configuration and application files on a managed node refresh from the master
copy of the files hosted on the deployment manager during synchronization.
In
certain secure environments, the Profile Creation wizard cannot
federate a custom profile into a cell. Such cases require you to use the addNode command
instead. If you have configured the deployment manager to use a JMX connector
type other than the default SOAP connector, use the addNode command
to add the node to the cell.
The deployment manager provides the administration
for all managed nodes that are in its cell. Periodically the deployment manager
refreshes the configuration files and application files on the managed node.
Copying the master version of the files hosted on the deployment manager to
the managed nodes is a process called synchronization.
In
a cell environment, only the managed nodes serve applications, not the deployment
manager. The managed node in this scenario uses its internal HTTP transport
chain for communication, instead of a using a separate web server (on a separate
machine) to possibly offload some processing.

- Install WebSphere Application Server Network Deployment.
- Create a deployment manager profile using the Profile Creation wizard.
- Start the deployment manager using the First steps console or
the startManager command.
- reate a stand-alone application server profile using the Profile Creation wizard.
- Start the application server using the First steps console or
the startServer server1 command.
- Add the application server node to the cell using the administrative
console of the deployment manager. Click System Administration >
Nodes to add the node.
- Scenario 6: Install a cell
of managed application server nodes and a Web server on one machine.
Installing a Web server, such as IBM HTTP Server, on the same machine
as the application server provides a richer set of configuration options.
Installing a Web server plug-in is required for the Web server to communicate
with the server in the managed node. This type of installation can support
testing in a cell environment all on one machine (when a firewall separating
the web server and application server is not required) but higher throughput
will usually be obtained when the web server is on a separate machine.

- Install WebSphere Application Server Network Deployment.
- Create a deployment manager profile using the Profile Creation wizard.
- Start the deployment manager using the First steps console or
the startManager command.
- Create a stand-alone application server profile using the Profile Creation wizard.
- Start the application server using the First steps console or
the startServer server1 command.
- Add the application server node to the cell using the administrative
console of the deployment manager. Click System Administration >
Nodes to add the node.
- Install IBM HTTP Server or another supported Web server.
- Install the Web server plug-ins and configure the Web server
using the Plug-ins installation wizard.
- The Plug-ins installation wizard creates a script named configureWeb_server_name in
the plugins_root/bin directory.
Run the configureWeb_server_name script
to create a Web server definition in the administrative console. You can then
use the administrative console to manage the Web server.
- Scenario 7: Install a cell
of managed application server nodes on one machine and a Web server on a separate
machine.
In a typical production environment, a managed node
in a cell communicates with a Web server on a separate (remote) machine through
the Web server plug-in. An optional firewall can provide additional security
for the application server machine.

- Install WebSphere Application Server Network Deployment on Machine
A.
- Create a deployment manager profile using the Profile Creation wizard on
Machine A.
- Start the deployment manager using the First steps console or
the startManager command on Machine A.
- Create a stand-alone application server profile on Machine A
using the Profile creation wizard.
- Start the application server using the First steps console or
the startServer server1 command.
- Add the application server node to the cell using the administrative
console of the deployment manager. Click System Administration >
Nodes to add the node.
- Install IBM HTTP Server or another supported Web server on Machine
B.
- On Machine B, install the Web server plug-ins and configure
the Web server using the Plug-ins installation wizard.
- The Plug-ins installation wizard creates a script named configureWeb_server_name in
the plugins_root/bin directory
on Machine B. Copy the script from Machine B to the app_server_root/bin directory on Machine A.
You have the option of using the
script to create the Web server definition in the configuration of the deployment
manager or using the administrative console of the deployment manager to create
the Web server definition.
- Run the configureWeb_server_name script
on Machine A to create a Web server definition or use the administrative console
of the deployment manager to create the definition. You can then use the administrative
console to manage the Web server.
- Propagate the plugin-cfg.xml file from
the deployment manager on Machine A to the Web server on Machine B using
the administrative console. Click Servers > Web server >
Propagate Plug-in. (Web servers other than IBM HTTP Server require
manual propagation.)
- Scenario 8: Install a deployment manager on one
machine, multiple managed application server nodes on a second machine, and
a Web server on a third machine.
The primary advantage of a
cell over a stand-alone application server is its scalability. Managing a
cell to keep it in proportion with workload levels is possible. In this scenario,
managed nodes exist on Machine C. All of the managed nodes are federated into
the same deployment manager. Depending on your needs, an application server
in each managed node could serve the same or different applications.
Machine
A and Machine C represent both types of scaling, vertical and horizontal scaling:
- Vertical scaling creates multiple managed nodes on
the same physical machine.
- Horizontal scaling creates cell members on multiple
physical machines.
The managed nodes in this scenario communicate with the same Web
server. However, an alternative strategy could have a dedicated Web server
for each managed node.

- Install WebSphere Application Server
Network Deployment on Machine A.
- Create a deployment manager profile using
the Profile Creation wizard on Machine A.
- Start the deployment manager using the
First steps console or the startManager command on Machine
A.
- Install WebSphere Application Server
Network Deployment on Machine C.
- Create the first Application Server profile
using the Profile Creation wizard on Machine
C.
- Start the first application server using
the First steps console or the startServer server1 command
on Machine C.
- Create the second Application Server
profile and make this profile the default profile on Machine C.
Assign
unique ports. Write down the SOAP port for later use when federating this
node from the administrative console of the deployment manager.
- Start the second Application Server using
the First steps console or the startServer server1 command
on Machine C.
- Add both application server nodes to
the cell using the administrative console of the deployment manager on Machine
A. Click System Administration > Nodes to add the
nodes.
- Install IBM HTTP Server or another supported
Web server on Machine B.
- Install the Web server plug-ins and configure
the Web server using the Plug-ins installation wizard on Machine B.
- The Plug-ins installation wizard creates
a script named configureWeb_server_name in
the plugins_root/bin directory
on Machine B. Copy the script from Machine B to the app_server_root/bin directory on Machine A.
You have the option of using the
script to create the Web server definition in the configuration of the deployment
manager or using the administrative console of the deployment manager to create
the Web server definition.
- Run the configureWeb_server_name script
on Machine A to create a Web server definition or use the administrative console
of the deployment manager to create the definition. You can then use the administrative
console to manage the Web server.
- Propagate the plugin-cfg.xml file
from the deployment manager on Machine A to the Web server on Machine B using
the administrative console. Click Servers > Web server >
Propagate Plug-in. (Web servers other than IBM HTTP Server require
manual propagation.)
What to do next
After determining a possible topology, follow the steps in the
overall procedure.
One step not mentioned in any of these scenarios
is to prepare the operating system for installation. Useful links to the installation
procedures for each installable component are in the list of related topics.