Availability Guide for Application Design

Instrumenting an Application for Availability
Availability Guide for Application Design525637-004
8-38
User-Written Management Applications
The backup process of a process pair has encountered the same failure condition
encountered in the original primary.
A Pathway server process, in spite of numerous retry attempts, repeatedly returns
errors to a requester.
A communications link with a remote server process has gone down.
A hardware component and its backup have simultaneously failed.
Your management application can help to make your business application available
again in the shortest possible time by responding to critical event messages that
highlight these kinds of problems. An automated or human operator is then responsible
for switching processing to an alternate server class, switching to stand-in processing,
bringing communications lines back online, and so on.
User-Written Management Applications
While HP products such as NonStop TMF provide you with a wide range of
management capabilities, you can develop your own management application, if you
need to, by accessing the SPI programming interface.
The SPI programming interface enables you to write management applications that
can read event messages from EMS consumer distributor processes and issue
commands to applications and subsystems that have an SPI management interface.
The management application might also call Measure procedures to gather
performance statistics on system or application resources. Such an application might
also include an operator interface and logic to automate operations tasks by generating
commands from event messages.
A brief overview of the major functions of a user-written management application is
given here. For complete details, refer to the SPI Programming Manual.
When to Use SPI Programming for Management
Applications
If HP products or third-party products are not suitable for managing your applications,
you can design and write management applications yourself.
Using SPI to Send Commands and Receive Responses
To send commands to the application interface, your management application must do
the following:
1. Establish a connection with the subsystem’s management interface. If necessary,
you must start the interface and then open it. The subsystem is then ready to
receive commands.
2. Use SPI procedures to build an SPI message buffer containing a command. You
must first initialize the buffer and then add the command tokens.