Before you begin
The uninstall command calls the uninstaller program
that is created during installation. The uninstaller program is customized
for each product installation, with specific disk locations and routines for
removing installed features.
The uninstaller program removes registry entries, uninstalls
the product, and removes all related features. The uninstaller program
does not remove log files in the installation root directory.
About this task
This procedure uninstalls the WebSphere Application
Server Network Deployment product.
See the
following topics to uninstall other components on the product disc:
The
time required to uninstall is dependent on the processing speed of your machine.
As a rough guideline, uninstalling the core product files and one application
server profile takes approximately 10 minutes when using the uninstall command.
Log on using the same user ID
that was used when the product was installed. Issue the ls
-al command at the root of the WebSphere Application Server installation
to find the user ID that was used to install the product.
Log on as a user who belongs to the Administrators
group or as the user who installed the product.
- Run the uninstaller program for the Web server
plug-ins for WebSphere Application Server.
If a Web
server is configured to run with the Application Server, uninstall
the plug-ins to remove the configuration from the Web server. See Uninstalling the Web server plug-ins for WebSphere Application Server.
- Stop the node agent process with the stopNode command.
Stop the node agent process that might be running on the
machine. For example, issue the following command from the
profile_root/bin directory
of a federated node on a Linux machine to stop the node agent process:
./stopNode.sh
If servers are running and security is enabled,
use the following command:
./stopNode.sh -user user_ID -password password
- Stop the deployment manager dmgr process
with the stopManager command.
Stop
all dmgr processes that are running on the machine. For example, issue
this command on a Linux machine from the
profile_root/bin directory
of the deployment manager profile:
./stopManager.sh -user user_ID -password password
- Stop each running application server
with the stopServer command.
Stop
all server processes in all profiles on the machine. For example,
issue the following command from the
profile_root/bin directory
to stop the server1 process in the application server profile:
./stopServer.sh server1
If
a server is running and security is enabled, use the following command:
./stopServer.sh server1 -user user_ID -password password
If
you have multiple servers, you can use the
serverStatus command
to find running application servers. Issue the following command from
the
profile_root/bin directory
to determine which servers, if any, are running:
./serverStatus.sh -all
- Optional: Back up configuration
files and log files to refer to them later if necessary.
Use
the AdminTask command scripting interface to create
a configuration archive file of an existing WebSphere Application
Server profile, for example.
The
uninstaller program removes all profiles by default, including all
of the configuration data and applications in each profile. Before
you start the uninstall procedure, back up the config folder,
the installableApps folder, and the installedApps folder
of each profile, if necessary, or use the -OPT removeProfilesOnUninstall="false"
parameter on the uninstall command.
The uninstaller program leaves
all profiles, including all of the configuration data, applications,
and other user data in each profile. Use the -OPT removeProfilesOnUninstall="true"
parameter on the uninstall command to delete all
of the data, or delete it manually at a later time, if you decide
the data is unnecessary.
Back up the config folder
and the logs folder of each profile to refer
to it later if necessary. You cannot reuse profiles so there is no
need to back up an entire profile.
- Issue the uninstall command.
Run
the
uninstall command:
The uninstaller wizard begins and displays the Welcome
panel.
You can also issue the uninstall command with a silent parameter
to use the wizard without the graphical user interface.
Issue the following
command to start the uninstaller wizard in silent mode, without the graphical
user interface, and to remove all profiles:
- If you are using the wizard, click Next to
begin uninstalling the product.
The uninstaller wizard displays
a confirmation panel that lists a summary of the components that you are uninstalling.
- Click Next to continue uninstalling the
product.
When
using the wizard, a panel allows you to choose whether or not the uninstaller
deletes all profiles before it deletes the core product files. By default,
all profiles will be deleted, but this option can be deselected on the panel.
To change the default
behavior, start the wizard with this command:
uninstall -OPT removeProfilesOnUninstall="false"
After
uninstalling profiles, the uninstaller program deletes the core product files
in component order.
- Click Finish to close the wizard after
the wizard removes the product.
- Remove any configuration entries in
the managed node that describe a deleted deployment manager.
A
common topology is to install the core product files on multiple machines.
One machine has the deployment manager and other machines have managed nodes
created from custom profiles or federated application server profiles. If
you delete a Network Deployment installation where you created an application
server profile or a custom profile and federated the node into a deployment
manager cell in another installation, you must remove the configuration from
the deployment manager.
The official statement of support for a node
configuration problem in the managed node is that you use the backupConfig command
after the initial installation. Use the command again whenever you make significant
changes to the configuration that you must save. With a valid backup of the
configuration, you can always use the restoreConfig command
to get back to a previously existing state in the configuration.
If you must manually clean up the configuration on the managed
node, you can attempt the following unsupported procedure:
- Rename the cell_name directory for the
node to the original name if the current name is not the original name.
Go to the profile_root/config/cells/ directory.
Rename the cell_name directory to the original name.
- Delete the dmgr_node_name directory if
it exists.
Go to the profile_root/config/cells/original_cell_name/nodes directory to look
for the dmgr_node_name directory that you must delete.
- Edit the setupCmdLine.sh file on an operating
system such as AIX or Linux, or the setupCmdLine.sh file
on a Windows system and change the cell name to the original cell name.
The file is in the profile_root/bin directory.
Change the value of the WAS_CELL variable to the original
cell name.
- Remove any configuration entries in
the deployment manager that describe a deleted managed node.
Open
the administrative console of the deployment manager and click System
administration > Nodes > node_name >
Remove node.
If the administrative console cannot successfully
remove the node, run the following command with the deployment manager running:
The official statement of support for a node configuration problem
in the deployment manager is that you use the backupConfig command
after the initial installation. Use the command again whenever you make significant
changes to the configuration that you must save. With a valid backup of the
configuration, you can always use the restoreConfig command
to get back to a previously existing state in the configuration.
If
you must manually clean up the configuration, you can attempt the following
unsupported procedure:
- Within the nodes directory of the deployment
manager, remove the configuration directory for the node that you deleted.
Go to the profile_root/config/cells/cell_name/nodes directory to find the deleted_node_name file.
- Within the buses directory of the deployment
manager, remove the configuration directory for the node that you deleted.
Go to the profile_root/config/cells/cell_name/buses directory to find the deleted_node_name file.
- Edit the coregroup.xml file in each subdirectory
of the coregroups directory of the deployment manager.
Look for elements of type coreGroupServers. Remove any
coreGroupServers elements that have a reference to the node that you deleted.
Go to the profile_root/config/cells/cell_name/coregroups/deleted_node_name directory
to find the file.
- Edit the nodegroup.xml file in each subdirectory
of the nodegroups directory of the deployment manager.
Look for elements of type members. Remove any members elements
that have a reference to the node that you deleted.
Go to the profile_root/config/cells/cell_name /coregroups/deleted_node_name directory to find the file.
- Review the log file.
Review
the app_server_root/logs/uninstlog.txt file.
Review the app_server_root/logs/uninstall/log.txt file.
The log file records file system or other unusual errors.
Look for the INSTCONFSUCCESS indicator of success in the log:
(date_time),
Uninstall, com.ibm.ws.install.ni.ismp.actions.
SetExitCodeAction, msg1,
CWUPI0000I: EXITCODE=0
(date_time),
Uninstall, com.ibm.ws.install.ni.ismp.actions.
ISMPLogSuccessMessageAction, msg1,
INSTCONFSUCCESS
- If any product files remain, uninstall manually before reinstalling.
The uninstaller program leaves some
log files, including the app_server_root/logs/uninstall/log.txt file.
Manually uninstall the product to remove all artifacts
of the product so that you can reinstall into the same installation
root directory. If you do not plan to reinstall, you do not need to
manually uninstall.
See Uninstalling manually for more information.
What to do next
After uninstalling the product and cleaning the system with the
manual uninstall procedure, you can reinstall.
See Task overview: Installing for
an overview of installing the product and creating a functional e-business
environment.