6.1.2 Configuring server start arguments

You must configure the server start arguments on each WebLogic Server instance of the LiveCycle cluster to add LiveCycle options.

Before you start this procedure, you must know if your cluster uses a 32-bit or 64-bit JVM. See Preparing to Install LiveCycle (Server Cluster)to determine the JVM that is required for your cluster configuration.

Before you start this procedure, determine how your LiveCycle cluster implements cluster caching so that you can correctly configure a server start argument for it. You can implement cluster caching by using UDP or TCP but not both. Choose the implementation appropriate for your cluster:

UDP can be used only if your cluster is IPv4-based.
Use TCP if your cluster is either IPv4-based or IPv6-based. On an IPv6-based cluster, use TCP to be IPv6-compliant.

If you implement cluster caching by using TCP, ensure that you configure the TCP locators correctly. (See “Configuring the caching locators (caching using TCP only)” ).

It is recommended to use TCP instead of UDP multicasting for production systems because of the inherent reliability of the TCP protocol.

To configure the server start arguments:

In the WebLogic Server Administration Console, under Domain Structure, click Environment > Servers and, in the right pane, click the name of a server in the LiveCycle cluster.
Click the Configuration tab > Server Start.
Under Change Center, click Lock & Edit.
In the Arguments box, add one of the following sets of JVM arguments:
- (32-bit JVM only) Add -XX:MaxPermSize=256m -Xms256m -Xmx1024m
- (64 -bit JVM only) Add -XX:MaxPermSize=512m -Xms256m -Xmx2048m
Note: Copy the appropriate text block above to a text editor, and ensure that all line breaks are removed:
(IPv4 only) In the Arguments box, add the following JVM arguments:

-Dadobeidp.RootDirectory=<appserver domain> -Djava.net.preferIPv4Stack=true -Dfile.encoding=utf8

-Djava.security.policy=<appserver root>/server/lib/weblogic.policy
(IPv6 only) In the Arguments box, add the following JVM arguments:

-Dadobeidp.RootDirectory=<appserver domain> -Djava.net.preferIPv6Stack=true -Djava.net.preferIPv6Addresses=true -Dfile.encoding=utf8

-Djava.security.policy=<appserver root>/server/lib/weblogic.policy

Note: If you're using a 64-bit UNIX platform, add the following JVM argument:
```
-d64
```
Note: The adobeidp.RootDirectory must be created in exactly the same location on all nodes of the cluster.

Copy the above text block to a text editor, ensure that all line breaks are removed, and replace all of these occurrences:
- <appserver domain> with your application server domain path
- <appserver root> with the application server root directory
Configure a JVM argument for cluster caching. In the JAVA_OPTS line, add or change one of the following arguments:

Caching using UDP discovery
- Configure the multicast port argument in the following format:
```
-Dadobe.cache.multicast-port=<port number>
```
  Note: The value for <port number> can be any available port between 1025 and 65535. The multicast port must be unique to the LiveCycle cluster (that is, the port must not be used by any other cluster on the same network, any attempt to use the same port by any other cluster on the same network would result in bootstrap failure). It is recommended that you configure the same <port number> on all nodes in the LiveCycle cluster, as in this example:
```
-Dadobe.cache.multicast-port=33456
```
- Setting multicast address argument is optional. Default muticast addresses for IPv4 and IPv6 are as following:
```
IPv6 - FF38::1234 
IPv4 - 239.192.81.1
```
  If you have restriction on multicast addresses in your network, use following argument to set multicast addresses:
```
-Dadobe.cache.multicast-address=<ip address>
```
  The value for <ip address> is the IP address used for multicast networking. The IP address is ignored if adobe.cache.multicast-port is zero.
  
  The multicast address must be unique to the LiveCycle cluster and must not be used by any other cluster on the same network. It is recommended that you configure the same <ip address> on all nodes in the LiveCycle cluster. For example:
```
-Dadobe.cache.multicast-address=239.192.81.1
```
- For machines with multiple Network Interfaces
  
  Some machines may be connected to multiple networks via multiple Network Interface Cards (NICs). For such machines, set the JVM property -Dadobe.cache.bind-address to the IP address of the network interface card that you are using for LiveCycle Server.
```
-Dadobe.cache.bind-address=<IP Address>
```
  Note: It is recommended to set JVM property -Dadobe.cache.bind-address for machines with one Network Interface Card, also.
Caching using TCP only
- For IPv4, configure the cluster locators argument in the following format:
```
-Dadobe.cache.cluster-locators=<IPaddress>[<port number>],<IPaddress> [<port number>]
```
  For IPv6, configure the cluster locators argument in the following format:
```
-Dadobe.cache.cluster-locators=<hostname>@<IPv6 address>[<port number>],<hostname>@<IPv6 address>[<port number>]
```
  Note: Configure, as a comma-separated list, the locators for all nodes of the cluster. The value for <IPaddress> is the IP address of the computer that is running the locator. The value for <port number> is any unused port between 1025 and 65535. It is recommended that you configure the same <port number> on all nodes in the LiveCycle cluster, as in this example:
```
-Dadobe.cache.cluster-locators=10.20.30.5[22345],10.20.30.6[22345]
```
  Note: Do not run TCP locators on all the nodes of the LiveCycle Cluster. It is good practice to run minimum two locators. One TCP locator serves as a primary locator and other TCP locator serves as a secondary locator to handle failover issues. You can add more than two TCP locators as backup locators but it is not mandatory. For information on configuring TCP locators, see 4.4 Configuring the caching locators in clusters (caching using TCP only).
To prevent application server from Denial of Service attacks configure the following JVM argument:
```
-DentityExpansionLimit=10000
```
Click Save and then click Activate Changes.
Repeat steps 2 to 8 for each server in your cluster.