Server relevant failure prediction method and system

Abstract

The invention discloses a server relevant failure prediction method. The method comprises steps as follows: a failure type is determined after a failure happens to a server; a failure relevant to the failure is acquired from a relevant failure set obtained in advance according to an Apriori algorithm and taken as a follow-up prediction failure; a worker performs corresponding compensation operation according to the follow-up prediction failure. After the failure happens to the server, possible follow-up failure types are predicted according to the relevance of the failure types in the relevant failure set, so that the worker can perform certain compensation operation in advance, the follow-up failure is avoided in time, influence on the server is reduced as far as possible after the follow-up failure happens, and the RAS (remote access service) performance of the server is improved. The invention further discloses a server relevant failure prediction system. The system also has the advantages, and unnecessary details are not described here anymore.

Description

technical field [0001] The present invention relates to the technical field of server fault management, in particular to a server related fault prediction method and system thereof. Background technique [0002] With the rapid development of computer technology, high-performance server systems have increasingly become the needs of economic and social development, and the requirements for server security have also increased, and security is mainly reflected in RAS performance. RAS performance refers to machine reliability (Reliability), availability (Availability) and serviceability (Serviceability). The RAS capability mainly depends on several aspects: system-level RAS features, CPU RAS features, memory RAS features, I / O RAS features, etc. [0003] These different levels of RAS features require that when a server fails, it can be quickly diagnosed and repaired, so that it will affect the stable operation of the server as little as possible. However, due to the possible corr...

AI Technical Summary

Problems solved by technology

[0003] These different levels of RAS features require that when a server fails, it can be quickly diagnosed and repaired, so that it will affect the stable operation of the server as little as possible. However, due to the possible correlation between different failures, that is, One fault may cause another fault, but the current server can only detect and deal with the fault after the fault occurs, and the diagnosis and processing of subsequent faults are often not timely enough, which easily affects the RAS performance of the server

Images