infiniband network detection method

Abstract

The invention discloses an InfiniBand network detection method. The method includes the steps of obtaining a first corresponding relationship between a device name of a device in the InfiniBand network and an LID (link interface device) number, a second corresponding relationship of a physical port number and a logical port number of each port in the device, the LID number of the device with a wrong port in the InfiniBand network and a logical port number of the wrong port and then obtaining the device name of the device with the wrong port and the physical port number of the wrong port according to the first corresponding relationship, the second corresponding relationship, the LID number of the device with the wrong port and the logical port number of the wrong port. The InfiniBand network detection method can be widely applied to clusters of different sizes, and a link of a problem in the entire IB (InfiniBand) network can be found out only through one operation of the method. Furthermore, the detection method can be integrated with other debugging methods, and thus, human error can be minimized, debugging efficiency of a large IB network can be greatly improved, and time and manpower are saved.

Description

technical field [0001] The present invention basically relates to the field of network communication, and more specifically, relates to an InfiniBand network detection method. Background technique [0002] InfiniBand (abbreviated as IB) is a high-speed network technology and standard. It has the advantages of high bandwidth and low latency. It is widely used in high-performance computing, big data processing and other fields. High-performance computing clusters usually use InfiniBand as a high-speed computing network or storage network. . Before the nodes of the cluster use the InfiniBand network, it is necessary to ensure that each device connected to the network can send and receive data normally. In addition, due to the electrical performance of the hardware, the quality of the cable, and the physical contact during device installation, some ports of some devices may fail to send data. The zero error rate cannot be maintained. At this time, it is necessary to debug the I...

AI Technical Summary

Problems solved by technology

At this time, it will be a very time-consuming task to manually compare and find the problematic devices one by one. In a large supercomputing center, such a debugging will even consume the staff for half a day to a day; Search and compare, it is also very easy to miss in the search process

Images