NonStop Server Guide for BEA WebLogic Server 8.1

Introduction
HP NonStop Server Guide for BEA WebLogic Server 8.1—529891-001
1-8
Differences Between SMP Systems and HP
NonStop Servers
The same domain (without any specific direction from the WebLogic Server domain
administrator) can be deployed transparently:
The WebLogic Server instances are distributed across all the available CPUs (which
are loosely coupled). A JVM and all its threads run within the CPU where the JVM was
started, so, when a CPU is lost, only WebLogic Server instances running on that CPU
are affected. WebLogic Server instances on other CPUs are not affected. The
WebLogic Server instances on the failed CPU can be automatically restarted on other
CPUs without reconfiguration and migration requirements.
Fault-tolerance is provided by every component (such as TCP/IP, TMF, and SQL/MX)
so the WebLogic Server application on a NonStop server is always available.
Moreover, the Node Manager and Administration Server are monitored to restart
automatically if the processor they run on fails (or the process itself fails for other
reasons). This method provides continued availability of monitoring and management
services.
The deployment benefits of the NonStop server, such as tight integration with the fault-
tolerant transaction manager, massively scalable relational database, and TCP/IP are
transparently provided to the WebLogic Server application when deployed on a
NonStop server.
Figure 1-5. WebLogic Server Domain Representation on a NonStop Server
VST002.vsd
Administrative
Server
Managed Server
(independent of
cluster)
CPU 0
Managed Server
(part of a cluster)
Node Manager
Managed Server
(part of a cluster)
Managed Server
(independent of
cluster)
Managed Server
(part of a cluster)
Managed Server
(part of a cluster)
CPU 1 CPU 2 CPU 3