NonStop Server Guide for BEA WebLogic Server 8.1

Configuring Persistent WebLogic Server Processes
HP NonStop Server Guide for BEA WebLogic Server 8.1529891-001
4-11
no_load
case because no servers will have been started so all the CPUs will have the same
load (that is, 0). If a different CPU is listed first for each Managed Server, the servers
will still start in different CPUs. However, managed_load still relies heavily on the pid
files remaining current. If Managed Servers are started or restarted frequently, the
likelihood of the pid files containing stale data increases.
no_load
no_load does not use a load-balancing algorithm. Instead, every CPU listed in
WLS_CPUS that is currently up receives a load of 1. Thus, when the CPUs are scanned
in WLS_CPUS-order, the first CPU that is currently up is used. The remaining CPUs in
the list are examined when all the earlier CPUs in the list are down. In effect, the first
CPU in WLS_CPUS is the primary CPU for the Managed Server, the second CPU is the
backup, and so on.
Using no_load allows the most control for assigning Managed Servers to specific
CPUs. Therefore, when working with replication groups, using no_load allows you to
avoid having the primary and secondary server being started in the same CPU.
However, in a 2-CPU system, the backup CPU for a Managed Server needs to be the
primary CPU of another Managed Server. By using no_load, you can avoid any issue
with the multi-start case because every server will start in its assigned CPU regardless
of where the other servers are started.
Configuring the Node Manager
This section provides step-by-step instructions about how to configure the Node
Manager, including settings specific to the NonStop server platform. For further
information, refer to the Overview of Node Manager and Configuring, Starting, and
Stopping Node Manager sections in the BEA document titled Configuring and
Managing WebLogic Server and the Troubleshooting and Notes section.
Settings
These instructions use and assume these settings:
The server host where admin, Node Manager, and managed servers are running is
called: icebat4.txn.cpqcorp.net
A domain with servers in it already exists and is called rldomain. There are two
managed servers called rlserver1 and rlserver2.
Step 1. Create a Machine and Associate Servers To It
1. Start an admin server for your domain:
. ./my_env (loading your usual WL_HOME, JAVA_HOME, and so on)
cd $WL_HOME/../user_projects/domains/rldomain
./startWebLogic.sh
Access the GUI. For example: http://icebat4.txn.cpqcorp.net:7001/console