A weekly technical update on distributed technology 2016-09-04

Searcher Distributed portraitThe main areas covered by the direction include the theory of distributed systems and design practices,Resource Management and virtualization technologies, large-scale service stability technology Devops,  "distributed direction one week technology dynamic " is a featured technical article on distributed direction that I summarize and organize weekly hope to enable you to follow the industry-related technology dynamics , ,  Learning Distributed system theory and design ideas ,  support everyone's daily work .  Weekly technical developments will be released hi welcome everyone to read .   distributed technology dynamic   have any comments or suggestions ,  You can always contact me

Hi Group : 1478429

Mail group : [email protected]

Previous list: Distributed direction one week technical news

Distributed System Practice

1. Open source Phxsql: Highly available, consistent MySQL cluster

http://mp.weixin.qq.com/s?__biz=MzI4NDMyNTU2Mw==&mid=2247483783&idx=1&sn= A2d6e589f1f591ded7703eb74aefccbe&scene=1&srcid=083090unfb8jmdsgvoe7g0lm#rd

Key points: The two-day open source Phxsql News has been in the circle of friends, from the open source phxpaxos that moment, I have a premonition of the arrival of this day, hehe. Although the selection of the main or strong consistent data in the case of the use of ZK, but ZK is a bypass system, and can not support the log class data strong consistent replication, so based on Paxos or raft to achieve a strong consistency of data replication lib ( Formally, a state machine based on the Paxos or raft protocol can be applied to the data consistency-preserving scenario of multiple replicas in a distributed storage system. and Phxsql formally used the Paxos protocol to realize the Binlog replication of MySQL, which solves the problem of weak consistency between master and slave MySQL. (however, MySQL 5.7 contains the group replication function can also achieve strong consistency, but it is still GA stage, hehe)

2. Cassandra Fault detection principle--accrual Failure Detector

http://mp.weixin.qq.com/s?__biz=MzAxNjc1MTk5Nw==&mid=2649157371&idx=1&sn= Cccc27cd1bfaada57344f3d2a9f9731e&scene=0#rd

Key points: When it comes to service failure detection, we think of the method is nothing but the heartbeat, but how long or how many times the seizure of the heart failure? Or how often the heartbeat is sent? This article describes a fault detection algorithm that helps you answer these questions.

Service and resource management technology

1. Container and virtual machine Twin Cities

Http://blog.daocloud.io/dockervsvm/?hmsr=toutiao.io&utm_medium=toutiao.io&utm_source=toutiao.io

Important: Containers and virtual machines are fundamentally different, and many of the students who are just beginning to learn Docker tend to think of Docker as a lightweight virtual machine (in fact, many beginners ' tutorials are written just to help you understand Docker quickly), This article explains the differences and relationships between containers and virtual machines.

2. Deep analysis of the "state" of the container

http://mp.weixin.qq.com/s?__biz=MzIwNzA1MTA5OA==&mid=2888434879&idx=1&sn= 4bbc3656c2f4b102bfed1b649fee86b7&scene=4#wechat_redirect

Important: Containers cannot be persisted, but many applications are in fact saved state, including data, configuration, logs, keys, and other information, this article describes the different "state" solutions

Service High Availability Technology

1. From 5 servers to the three centers: the evolution of the operation and maintenance architecture of Meizu system

http://mp.weixin.qq.com/s?__biz=MzAwMDU1MTE1OQ==&mid=2653547732&idx=1&sn= C41362ec93692d9d8e0792e978f1a269&scene=1&srcid=0830xrjrfiek5dpu7broapd6#rd

Important: This article introduces the evolution of the Meizu system architecture, which allows us to glimpse the whole system architecture of Meizu.

2. Analysis of client timeout problems caused by linux-2.6.32 version kernel TCP bug

http://mp.weixin.qq.com/s?__biz=MzI0NzAxMTgxNg==&mid=2456160301&idx=1&sn= 2ed8df42e40fa141ac4d9ce5cb0e1afa&scene=1&srcid=0831rvoxaavh5kmqecunbjzn#rd

Important: This article describes the problem from the surface until the process of tracking the nature of the problem, the Linux kernel, although the stability is very high, but still have bugs, encounter strange problems, the last way is to look at the Linux source code.

DevOps Technology

1. Nine best practices that DevOps must understand

http://mp.weixin.qq.com/s?__biz=MzA4MjkxMzMyNg==&mid=2654068995&idx=1&sn= 1f2f4a803ebfbca887b76178c6a23d02&scene=0#rd

Important: There is an article summarizing DevOps practices that provides 9 tips that are very pertinent.

2. What? In the network operation, the mature company must eliminate the cto?!

http://mp.weixin.qq.com/s?__biz=MzA4Nzg5Nzc5OA==&mid=2651661867&idx=1&sn= 8a16e8c7b0d54ecb50862f758911ac51&scene=0#rd

Main points: Online often occur on the network jitter, switch exceptions brought about by the refusal or downgrade, we try to complain about the quality of the network, but the network operation is really very complex, this article introduces the network operation of various pain points, although a lot of pain points there is no solution, but it contains a senior network workers in-depth thinking.

Hodgepodge

1. Inter-process communication for Linux-semaphores

http://liwei.life/2016/08/29/semaphores/

Important: Semaphores are one of the most common means of interprocess communication, and this article explains in depth the use of semaphores.

2. Start understanding the CFS scheduler from a few questions

http://ju.outofmemory.cn/entry/105407

IMPORTANT: CFS is a very good process scheduling algorithm, the goal is to ensure the fairness of the process, this article through a few questions of the solution let us more in-depth understanding of CFS.

A weekly technical update on distributed technology 2016-09-04