High availability provides several benefits, including
load balancing and failover. High availability with WebSphere® Premises Server consists
of setting up a server cluster and then configuring those servers
for load balancing.
- Make sure that you have completed all the prerequisite steps necessary for
high availability.
- Launch the high availability post-installation script located
at the root of the High Availability for IBM® WebSphere Premises Server Central
Site Server disk.
setupwin32.exe
setuplinux.bin
The Welcome panel displays.
- Click Next.
- This panel shows the installation directory for the WebSphere Premises Server high
availability system.
The directory is:
IBM_RFID_HOME\HA
IBM_RFID_HOME/HA
Click Next to continue.
- Enter the host name and port for WebSphere Application Server Network Deployment,
and click Next.
Tip: Make
sure that WebSphere Application Server Network Deployment is
running. The installer verifies that it can connect to WebSphere Application Server Network Deployment using
the port and host name you have provided before continuing. If it
cannot connect, you will be asked to go Back and
edit the values on the previous panel, or you can Cancel out
of the installer.
- Create the cluster members. Create at least
one member on this panel in order to proceed with the installation.
Use
the Add Member button to add cluster members.
The created member's name, node, and weight appear in the box at the
bottom of the installer panel.
To delete a cluster member, select
the member name from the list of created members and click Delete
Member.
For more information on creating cluster
members, see Adding members to a cluster.
- Click Next.
- A summary panel displays your installation selections.
Click Install to continue the installation
process.
When the installation is complete, another
summary panel displays the installation status and prompts you to
check the log file for any errors.
IBM_RFID_HOME\HA\logs\install.log
IBM_RFID_HOME/HA/logs/install.log
If you do see errors or exceptions in the installation log
file, try uninstalling and reinstalling the high availability topology.
Also check the Troubleshooting tips documentation for
possible resolutions to the problem. If you are unable to resolve
the errors, contact IBM Support.
- If you see exceptions in the WebSphere Application Server SystemOut.log file
on the central and node servers, follow the procedure in this technote.
- Restart the central server and the cluster.
- If you are using WebSphere Application Server security,
enable it, and then restart the deployment manager, all node agents,
and all servers.
- Enable dynamic cache replication for all servers in the
cluster.
- In the WebSphere Application Server administrative console,
go to and check Enable service
at server startup for each server in the cluster.
- Define a new replication domain by going to . Choose Entire
domain when creating the new replication domain.
- Navigate to and add the new replication to all object cache components.
- Check Enable cache replication.
- Choose your cluster name for Full group replication
domain.
- Choose Push only for Replication
type.
- Set Push frequency to 1 seconds.
- Configure the Data Capture and Delivery controllers
for high availability.
- Make sure you are using Java™ 1.4.2
on your Data Capture and Delivery controllers.
- Set the appropriate MQ user name for your operating
system in the controller's Equinox script.
-Duser.name=MUSR_MQADMIN
-Duser.name=mqm
If you used the sample
files provided with the IBM Data Capture and Delivery Toolkit for WebSphere Premises Server to
set up your remote Data Capture and Delivery controllers,
modify the remoteDC script with the MQ user name.
For example:
%JAVA_HOME%\bin\java" -Duser.name=MUSR_MQADMIN -Xmx256M -Xms256M
"$JAVA_HOME/bin/java" -Duser.name=mqm -Xmx256M -Xms256M
- Edit the config.ini file in the
controller's Equinox configuration directory
make sure the configuration is set to the dc_core4dts.txt file
for the bundle list and E4 for the edge controller.
com.ibm.rfid.bundle.list.url=http://IP_address:port_number/bundleadmin/GetBundle?name=http://IBM_HTTP_Server_IP_address/bundles/bundlelists/dc_core4dts.txt
com.ibm.rfid.edge.config.url=http://IP_address:port_number/ibmrfidadmin/premises.sl?action=getconfig&edge=E4&version=6.1
The
values for IP_address and IBM_HTTP_Server_IP_address are
the name of the server that is hosting the Bundle Repository Server.
The
second line of code points to the E4
controller, which is installed with WebSphere Premises Server specifically
for high availability.
- Check to see if Data Transformation is running
(started as a service) on your central server, and if so, stop it.
- Start the Equinox runtime on the Data Capture and Delivery controllers.
- Start the bundle loader on the Data Capture and Delivery controllers.
- Find the ID of the bundle loader bundle by running the
OSGi ss command.
- Start the bundle loader bundle by entering start bundle_ID at
the OSGi prompt.
- Test the clustered configuration
using the Simulated Reader in the WebSphere Premises Server Administrative Console. Choose R4 as your simulated test reader.
Optionally,
you can test with a real reader.
- Create a new remote Data Capture and Delivery controller based
on the E4 sample to use with your real reader.