Method for automatically locating internal storage faults

Abstract

The invention provides a method for automatically locating internal storage faults. The method comprises the following steps: reading a value of a register of an MCA, and analyzing the value of the register so as to obtain internal storage fault information, wherein the internal storage fault information comprises internal storage fault types and the number of fault times corresponding to the internal storage fault types; comparing the number of fault times corresponding to each internal storage fault type with a preset threshold value of the number of corresponding fault times; locating the internal storage fault of which the number of fault times exceeds the threshold value of the number of corresponding fault times so as to obtain the logical position of a fault internal storage. By means of the method, the internal storage faults of a server can be located, and the internal storage is repaired according to the internal storage fault types.

Description

technical field [0001] The invention relates to the technical field of server maintenance, in particular to a method for automatically locating memory faults. Background technique [0002] In the data center, the number of servers is very large, and there are multiple roots of memory in each machine. If a memory module fails, you need to determine which memory module is faulty. In terms of server memory fault location, out-of-band memory monitoring has become a means that most enterprises rely on. Server manufacturers often provide users with server hardware information through BMC (Baseboard Management Controller, baseboard management controller) + sensors. health status. The memory fault is intercepted by the SMI (structure of management information, management information structure) supported by the BIOS (Basic Input Output System, basic input and output system), and the information is stored in the Sensor of the BMC system. Users can use the IPMI (Intelligent Platform ...

AI Technical Summary

Problems solved by technology

[0004] First, most machine models do not support accurate memory monitoring. Therefore, the use of out-of-band memory monitoring is limited by manufacturers and server models.

[0005] Second, the alarm mechanism of out-of-band memory monitoring often does not meet the actual needs of enterprises

When manufacturers determine the alarm threshold, they usually use the comprehensive value of the research results from various customer units, and it cannot be adjusted dynamically.

However, this value cannot fully adapt to the needs of every enterprise

Images