HP NetServer AA 4000 HP Netserver AA Solution Administrator's Guide v - Page 27

Endurance Software Endurance software integrates and synchronizes processing across four industry-standard Pentium class systems. All components are capable of multiprocessing, forming a logical, fully redundant, fault tolerant server. This results in the Windows operating system, all applications, and the local area network logically perceiving the integrated systems in the Endurance server as a standard Windows server. Endurance server software provides fault tolerant functionality, including: • Redundant processing environment. Using redundant CEs and memory, the Endurance software performs CPU and memory functions in a tightly coupled manner. In the event of a CPU or memory failure, the operating system remains continuously available, providing constant computing without any loss of context. • Redundant I/O subsystem. The Endurance software redirects I/O requests from the CE to the IOPs. The Endurance software coordinates all I/O requests, as follows: - Because all I/O devices in the Endurance server physically reside on the IOPs, Endurance software redirects all I/O requests from the CEs to the IOPs. As a result, these I/O devices are logically associated with the CE. For example, the Endurance software can redirect a CD-ROM drive physically located in an IOP so that it is logically assigned to the CE (the CE has complete control of that CD-ROM drive). - For mirrored disk devices and redundant Ethernet connections, the Endurance software coordinates all I/O processing so that these redundant components appear as a single logical device. For example, in a mirrored disk set, the two physical disks (one on each IOP) appear to the CE operating system and applications as one logical disk. If a disk failure occurs, the disk operations can transparently continue operating without affecting any server applications or clients. • Automatic fault detection and reporting. The Endurance software continuously monitors the server. If the software detects a fault, the Endurance server isolates the fault and deconfigures any affected components. When a fault occurs, Endurance software ensures that the server remains operational by automatically using the redundant components. The fault detection and reporting process occurs seamlessly, and in a fully operational Endurance server, it maintains complete server availability. The Endurance software reports all events, including general status, state changes, and faults, to the Windows event log or displays them in application windows as they occur. For a detailed description of all Endurance messages, refer to HP Netserver AA Solution Messages. Introduction 1-9