System Manager for Himalaya S-Series Non-Stop Kernel Systems

Compaq Computer Corp.

Client: Jennifer Dunvan (Manager, Supportability Products)

Summary

Compaq Computers' Himalaya S-Series Non-Stop Kernel (NSK) systems are used for very demanding, mission-critical applications. These systems are redundant, so a single failure will not stop the system. However, after a failure, the failing component should be quickly replaced to restore fault-tolerance margins to the system. Therefore the NonStop Himalaya Server requires continuous systems management. A sufficient definition of "management" in this context means that all hardware component bad health states (such as a disk controller failure) will be diagnosed immediately by the manager, and all service operations (such as a tape drive replacement) will be intelligently aided by the manager.

The manager resides on an NT workstation that communicates to the NonStop server via a TCP/IP connection.

The manager uses a Java-based user interface on the NT workstation to ultimately manage the NonStop server via the TCP/IP connection.

Objective

Develop a System Manager Prototype that fulfills the requirements described above.

Required Components

In addition to the general requirements in the summary, the System Manager Prototype will fulfill the following:

1. Receive hardware system state data from the NonStop server
2. Provide a real-time display of the NonStop system state to the user
3. Immediately alert the user of any NonStop system component failure
4. Suggest appropriate recovery actions based on the type of system component failure
5. Guide the user through such recommended recovery actions in an interactive and automated fashion
6. Allow the user to safely execute system commands on the NonStop server and view the results of the executed code