Poor Man's Distributed Network
Author: youerning
Introduction: Foreword: The poor refer to no money or do not want to spend too much money (I belong to both the former and the latter TT), distributed network mainly refers to the host network environment distributed in different geographical environments such as different provinces or different countries (Who doesn't have a cheaper foreign VPS is not \~\~)
Summary of a virtualization project: failure caused by a fiber
Author: Wang Chunhai
Introduction: Abstract: In a virtualization project in September this year, everything was normal in the early stage of the project. After adding and replacing memory for the server, the ESXi host storage was disconnected, the virtual machine system was slow, and the ESXi host was slow to start. After multiple inspections, the fault was finally rectified. The cause of the eventual failure was simple: a problem with the fiber connecting the ESXi host to the storage caused the failure. However, the whole project process involves a series of events such as memory replacement, motherboard replacement, firmware upgrade, etc. Therefore, the failure point was not correctly located in the early failure analysis, which made things more and more complicated. Below I will restore the whole process once, I hope this will be helpful to other friends who often do projects.
Realization of smart home based on Raspberry Pi (Raspberry Pi3) platform interacting through WeChat (WeChat module) (1)
Author: Dongdongjiang's k8s
introduction: In fact, I made this smart home system because of the graduation project of the school. It has been more than 20 days since the last article was published. I thought about just making a smoke alarm, and then using Zabbix to alarm , but through the design of more than 20 days, I found that the function of realizing the alarm is actually not only email, but also SMS, WeChat, and even phone calls. But due to various reasons, such as.....money 0.0, hahaha, so I plan to use the corporate WeChat to make an alarm, and then it seems that a simple interaction through ordinary WeChat is still a good choice, and make it The effect is also great.
【Enterprise new clothes】--How to quickly design a highly available Alibaba Cloud architecture for enterprises
Author: Gan Bing
Introduction: Some friends may also know about Alibaba Cloud and have used some of Alibaba Cloud's products. If a company asks you to design an Alibaba Cloud environment, Alibaba Cloud has so many products (ECS, CDN, RDS) etc.), how do we blend these products and services together beautifully? This is also a topic that many friends are more concerned about, and the author will discuss it with you.
Oracle HA dual-host active and standby is based on shared storage mode and uses keepalived management to achieve high availability
Author: HMLinux
Introduction: HA (High Available) consists of two computers and provides a set of the same services to the outside world, also known as a master and a backup mode. Under normal circumstances, the main server provides services, and the standby server is in standby mode, and the standby server can take over the work of the main server at any time. That is, when the primary server goes down or the services provided are unavailable, the standby server will actively activate and replace the primary server to continue to provide services. At this time, the service resources on the primary server include network (IP), storage, and services (Web/database). It will be transferred to the standby machine to take over, thereby providing uninterrupted service. Based on this, the downtime of the server can be reduced to a minimum, and the business can be uninterrupted or temporarily interrupted.
Geometric shape recognition and measurement in OpenCV
Author: gloomyfish
Introduction: I often see people who have recently learned OpenCV asking questions about how to identify some simple geometric shapes and their colors. In fact, functions related to geometric analysis are found through the contours of OpenCV, and it only takes less than 100 lines of code. These simple geometric shape recognition and object measurement related operations can be well implemented. This article will demonstrate how to use OpenCV contour discovery and geometric analysis related functions to achieve the following functions
Flume+Kafka+Storm+Redis builds big data real-time processing system: real-time statistics website PV, UV+ display
Author: xpleaf
Introduction: The article given above is an offline data analysis case based on MapReduce. It processes the user access logs generated by the website and analyzes the PV, UV and other data of the website on a certain day, corresponding to the above As shown in the figure, it takes the data processing method of offline processing, and here is the data processing method of another route, that is, online processing based on Storm. In the complete case given below, we will complete the following several jobs
10-year-old station vomiting blood migration record
Author: sery
Introduction: Due to some content censorship reasons, a forum that has been running for a long time was asked to migrate to the machine room within a time limit. The control is very strict, and it is required to be completed within 1-2 days. The files of the forum (mainly pictures and attachments) are more than 100G, and the database has 80G. I evaluated it, exporting data, deploying a new environment, importing data, debugging until normal access, 1-2 days may be a bit stressful, and if the migration does not go smoothly, it will take more time. But there is no way to refuse, first agree to say it.
keepalived + nginx polling method for high availability and load balancing access to backend apache web service
Author: Dockerhadoop
Introduction: keepalived+nginx high availability has two modes: active-standby and active-active. In the active-standby mode, one VIP is provided externally, and only one server is working at the same time, and the other is a standby machine; in the active-active mode, two VIPs are provided externally, and the two machines are mutually backup. The following describes the test steps in detail.
Analysis of VGG Convolutional Neural Network Model
Author: gloomyfish
Introduction: The full name of GG is Visual Geometry Group, which belongs to the Department of Science and Engineering of Oxford University. It has released a series of convolutional network models starting with VGG, which can be used in face recognition, image classification, etc., from VGG16 to VGG19. The original intention of VGG to study the depth of convolutional networks is to find out how the depth of convolutional networks affects the accuracy and accuracy of large-scale image classification and recognition. Originally, VGG-16 is known as the very deep convolutional network (GG-Very -Deep-16 CNN), VGG is deepening the number of network layers and in order to avoid too many parameters, 3x3 small convolution kernels are used in all layers, and the convolution layer step size is set to 1.