This topic describes creating a runtime environment for a deployment
manager.
Before you begin
Before
using the Profile Creation wizard,
install the core product files.
The Profile Creation wizard is
the graphical interface to the wasprofile command. See the description of the wasprofile command
for more information.
You must provide enough system temporary
space to create a profile. For information on the requirements, see the Profiles: file system requirements
topic.
About this task
After installing the core product files for the Network Deployment
product, you must create a profile. It can be a deployment manager profile,
an application server profile, or a custom profile. This procedure describes
creating a deployment manager profile using the graphical user interface that
is provided by the Profile Creation wizard.
The deployment manager provides a single administrative interface to a logical
group of application servers on one or more machines.
You
can use the Profile Creation wizard in
silent mode with a response file, without the graphical user interface. See responsefile.pct.NDdmgrProfile.txt
for
examples of using the Profile Creation wizard in silent mode.
You can also use the wasprofile command to create a deployment manager. See
the description of the wasprofile command for
more information.
- Start the Profile Creation wizard to
create a new runtime environment.
Several ways exist
to start the wizard:
This task selects the Profile Creation wizard from
the First steps console:
- Open a command window.
- Change directories to the firststeps directory
in the installation root directory:
- Issue the firststeps command to start the console:
./firststeps.sh
firststeps.bat
- Select the Profile Creation wizard option
on the console.
The Profile Creation wizard is
an InstallShield for Multiplatforms application. The wizard loads the Java
2 SDK and then displays its Welcome panel.
See the description of the firststeps command
for
more information.
- Click Next on the Welcome panel.
The wizard displays the Profile type selection panel.
- Select the
option for creating a deployment manager and click Next.
The Profile name
panel is displayed.
- Specify a name for the
profile, or accept the default: click Next.
Profile
naming guidelines: Double-byte characters are supported. The profile name
can be any unique name with the following restrictions. Do not use any of
the following characters when naming your profile:
- Spaces
- Illegal special characters that are not supported within the name of a
directory on your operating system, such as *&?
- Slashes (/) or (\)
The default profile
The first
profile that you create on a machine is the default profile. The default profile
is the default target for commands that are issued from the bin directory
in the product installation root. When only one profile exists on a machine,
every command works on the only server process in the configuration.
Addressing
a profile in a multiprofile environment
When two or
more profiles exist on a machine, certain commands require that you specify
the profile to which the command applies. These commands use the -profileName
parameter to identify which profile to address. You might find it easier to
use the commands that are in the bin directory of each profile.
The commands are found in the:
A command has two lines. The first line sets the WAS_USER_SCRIPT environment
variable for the command window. The variable sets up the command environment
to address the profile. The second line calls the actual command in the
The actual command queries the command shell to determine
the calling profile and to autonomically address the command to the calling
profile.
The
wizard then displays the Profile directory panel.
- Accept the default directory, specify a non-default
location, or click Browse to select a different location. Click Next.
If you click Back and change the name of the profile, you
must manually change the name on this panel when it displays again.
The
wizard displays the Node, host, and cell name panel.
- On the Node, host, and cell names
panel, specify a unique node name, the actual host name of the machine, and
a unique cell name. Click Next.
The deployment manager
node has the following characteristics.
Field name |
Default value |
Constraints |
Description |
Node name |
The name of your machine, or
a unique derivation of the machine name. |
Use a unique name for the deployment manager. |
The name is used for administration within the deployment
manager cell. |
Host name |
The DNS name of your
machine. |
The host name must be addressable through your network. See Host name considerations.
|
Use the actual DNS name or IP address of your machine
to enable communication with your machine. See additional information about
the host name that follows this table. |
Cell name |
The arbitrary name of the deployment
manager cell. The cell is a logical grouping of managed nodes, under the control
of the deployment manager. |
Use a unique name for the deployment manager cell.
If you plan to migrate a V5 deployment manager cell to this V6 deployment
manager, use the same cell name as the V5 deployment manager. A cell name must be unique in any circumstance in which the product is running on the same physical machine or cluster of machines, such as a sysplex. Additionally, a cell name must be unique in any circumstance in which network connectivity between entities is required either between the cells or from a client that must communicate with each of the cells. Cell names also must be unique if their name spaces are going to be federated. Otherwise, you might encounter symptoms such as a javax.naming.NameNotFoundException exception, in which case, you need to create uniquely named cells. |
All federated nodes become members of the deployment
manager cell, which you name in this panel. |
Reserved
names: Avoid using reserved folder names as field values. The use of reserved
folder names can cause unpredictable results. The following words are reserved:
- cells
- nodes
- servers
- clusters
- applications
- deployments
Directory
path considerations
The
number of characters in the profiles_directory_path\profile_name directory
must be less than or equal to 80 characters.
Host name considerations
The host name is the network name
for the physical machine on which the node is installed. The host name must
resolve to a physical network node on the server. When multiple network cards
exist in the server, the host name or IP address must resolve to one of the
network cards. Remote nodes use the host name to connect to and communicate
with this node. Selecting a host name that other machines can reach within
your network is extremely important. Do not use the generic identifier, localhost,
for this value. Also, do not attempt to install WebSphere Application Server
products on a machine with a host name that uses characters from the double-byte
character set (DBCS). DBCS characters are not supported when used in the host
name.
If you define coexisting nodes on the same computer
with unique IP addresses, define each IP address in a domain name server (DNS)
look-up table. Configuration files for stand-alone Application Servers do
not provide domain name resolution for multiple IP addresses on a machine
with a single network address.
The value that you specify
for the host name is used as the value of the hostName property in configuration
documents for the stand-alone Application Server. Specify the host name value
in one of the following formats:
- Fully qualified domain name server (DNS) host name string, such as xmachine.manhattan.ibm.com
- The default short DNS host name string, such as xmachine
- Numeric IP address, such as 127.1.255.3
The fully qualified DNS host name has the advantages
of being totally unambiguous and flexible. You have the flexibility of changing
the actual IP address for the host system without having to change the Application
Server configuration. This value for the host name is particularly useful
if you plan to change the IP address frequently when using Dynamic Host Configuration
Protocol (DHCP) to assign IP addresses. A format disadvantage is a dependency
on DNS. If DNS is not available, then connectivity is compromised.
The short host name is also dynamically resolvable. A short name
format has the added ability of being redefined in the local hosts file so
that the system can run the Application Server, even when disconnected from
the network. Define the short name to 127.0.0.1 (local loopback) in the hosts
file to run disconnected. A format disadvantage is a dependency on DNS for
remote access. If DNS is not available, then connectivity is compromised.
A numeric IP address has the advantage of not requiring name resolution
through DNS. A remote node can connect to the node that you name with a numeric
IP address without DNS being available. A format disadvantage is that the
numeric IP address is fixed. You must change the setting of the hostName property
in Express configuration documents whenever you change the machine IP address.
Therefore, do not use a numeric IP address if you use DHCP, or if you change
IP addresses regularly. Another format disadvantage is that you cannot use
the node if the host is disconnected from the network.
After
specifying deployment manager characteristics, the wizard displays the Port
value assignment panel.
- Verify that the ports specified
for the deployment manager are unique and click Next.
Ports are recognized as being in use if
- They are assigned to a profile created under an installation performed
by the current user.
Validation of ports occurs when you access the Port value assignment
panel. Conflicts can still occur between the Port value assignment panel and
the Profile Creation Complete panel because ports are not assigned until profile
creation completes.
- Choose whether to run
the dmgr process as a Windows service on a Windows platform and click Next.
WebSphere Application Server attempts to start Windows
services for dmgr processes that are started by a startManager command.
For example, if you configure a deployment manager as a Windows service and
issue the startManager command, the wasservice command attempts
to start the defined service.
If you chose to install a local system service, you do not have
to specify your user ID or password. If you create a specified user type of
service, you must specify the user ID and the password for the user who runs
the service. The user must have Log on as a service authority for the
service to run properly.
To
perform this installation task, the user ID must not have spaces in its name.
The ID must also belong to the administrator group and must have the advanced
user rights Act as part of the operating system and Log on as a
service. The Installation wizard grants the user ID the advanced user
rights if it does not already have them, if the user ID belongs to the administrator
group.
You can also create other Windows services after the
installation is complete, to start other server processes. See Automatically restarting server processes
for more information.
IPv6
considerations
Profiles created to run
as a Windows service fail to start when using IPv6 if the service is configured
to run as Local System. Create a user-specific environment variable
to enable IPv6. Since this environment variable is a user variable instead
of a Local System variable, only a Windows service that runs as that
specific user can access this environment variable. By default, when a new
profile is created and configured to run as a Windows service, the service
is set to run as Local System. When the Windows service for the dmgr
process tries to run, the service is unable to access the user environment
variable that specifies IPv6, and thus tries to start as IPv4. The server
does not start correctly in this case. To resolve the problem, when creating
the profile, specify that the Windows service for the dmgr process runs as
the same user ID under which the environment variable that specifies IPv6
is defined, instead of as Local System.
The
wizard displays the Profile Creation Summary panel.
- Click Next to create
the deployment manager or click Back to change the characteristics
of the deployment manager.
The wizard
displays a Status panel during the creation of the profile. When the installation
is complete, the wizard displays the Profile creation is complete panel.
- Click Finish to exit and then click Profile Creation wizard on
the First steps console to start the wizard again, if you intend to create
an application server profile.