High-availability monitoring and management device and redundant switching method for high-density blade server

Abstract

The invention discloses a high-availability monitoring and management device and a redundant switching method for high-density blade servers. The device of the invention includes a master CMU and a slave CMU, a total of two redundantly arranged frame management units, the master CMU and the slave CMU. There are two communication links between them, and the two communication links include a first communication link for sending a heartbeat message containing equipment status information and a second communication link for sending a remedial heartbeat message containing equipment status information. road. The invention can ensure the ability of remote monitoring and management of all components, ensure the normal operation and maintenance of high-density blade servers, can greatly reduce the risk of operation and maintenance, can solve the phenomenon of "brain split", and can transmit equipment status through heartbeat messages and remedial heartbeat messages. It can comprehensively judge the status of the master CMU and the slave CMU to avoid the problem of switching without switching or switching without switching, which can greatly improve the availability of the high-density blade server monitoring and management system.

Description

technical field [0001] The invention relates to a high-availability technology for servers, in particular to a high-availability monitoring and management device and a redundancy switching method for high-density blade servers. Background technique [0002] A supercomputing center or data center is generally deployed with a large number of high-density blade servers. The chassis of each high-density blade server contains dozens of computing motherboards, several switching motherboards (service data network), one monitoring motherboard, and several chassis. Components such as power modules and chassis cooling modules (fans). Computing motherboards and switching motherboards usually integrate a board-level management unit (BMU, Base Management Unit) to implement board monitoring and management in the form of daughter cards. The monitoring motherboard acts as a chassis management unit (CMU Chassis Management Unit) to collect BMU monitoring and management information and impleme...

AI Technical Summary

Problems solved by technology

In addition, if a single heartbeat line implements redundant detection, the disconnection of the heartbeat line will cause the master and slave servers to think that they should undertake the service work and compete for shared resources, which will cause system chaos, that is, the phenomenon of "split brain", which is also redundant. Problems to avoid when switching

Images