Availability Guide for Application Design

Increasing the Availability of Tuxedo Applications
Availability Guide for Application Design—525637-004
5-9
Availability of NonStop Tuxedo Applications
processor failure, then the supervisor process restarts it in a different processor. If only
one processor is configured, then the WSL process is not restarted until the processor
comes back online.
When the WSL process restarts, it adopts any WSH processes that it previously
orphaned.
The configuration shown in Figure 5-3 on page 5-9 ensures that a WSL process is
always available when the workstation client attempts to connect. In this configuration,
two WSL processes are available, each listening for connection requests through
different NonStop TCP/IP process pairs, on different TCP/IP addresses, and through
different communications controllers. The workstation client can be configured with two
WSL network addresses; if it fails to make a connection through the first address, then
the second address is automatically tried.
Recovery From NonStop TCP/IP Process Failure
A TCP/IP connection can be maintained during periods of inactivity by specifying use
of the keep-alive option. Such connections normally are not broken unless the TCP/IP
process fails.
If the TCP/IP process is configured as a process pair, then the backup process takes
over if the primary TCP/IP process fails, and WSL processes continue to listen for
incoming connection requests. If the TCP/IP process has no backup process, then
WSL processes stop their associated WSH processes and then stop themselves.
Subsequently, the WSL process must be restarted manually.
Figure 5-3. Fault-Tolerant WSL Process Configuration
Workstation
/WS
Client
HP NonStop Server
NonStop Tuxedo
Transaction Monitor
(System/T)
NonStop
TCP/IP
NonStop
TCP/IP
NonStop
TCP/IP
TCP/IP
WSH
WSL backup
process
WSL primary
process
VST403.vdd