OKX review

List of contents of this article:

• 1,OKX review Distributed system infrastructure?
• 2, distributed log side Case
• 3, Plumelog distributed log component instruction manual
• 4, Flume Quick Start

Distributed system infrastructure?

1. Big data ecological technology system Hadoop is a distributed system infrastructure developed by the Apache Foundation. The core design of the Hadoop framework is HDFS and MapReduce. HDFS provides the storage of massive data, and MapReduce provides the calculation of massive data.

2. Distributed system For users, what they face is a server that provides the services users need. In fact, these services are a distributed system composed of many servers behind them, so the distributed system looks like a supercomputer.

3. Building a complete distributed system requires six necessary components: input node, output node, network switch, management node, control software and operation and maintenance module.

Distributed log scheme

1. Our project is a distributed system, but there is no distributed log system. It is extremely painful to check the log every time it is declassed. When N terminals are opened, the shell knocks off, which is extremely inefficient and ELK is decisively introduced.

2. If you want to diagnose complex operations, the usual solution is to pass the unique ID to each method in the request to identify the log. Sleuth can be easily integrated with the log framework Logback and SLF4J, and use log tracking and diagnostic problems by adding unique identifiers.

3. After the Hadoop Security mechanism and NodeMagager log aggregation functionThe analysis of the energy code explores two solutions: 1) Independent authentication by individual users in each computing framework; 2) Unified authentication by Yarn users in the log aggregation function module, and the advantages and disadvantages of the two solutions are compared.

4. Kafka is usually used to run monitoring data. This involves aggregating statistical information from distributed applications to generate a centralized operational data summary. Many people use Kafka as an alternative to log aggregation solutions.

5. Java intermediate: collaborative development and maintenance of enterprise team projects, modular foundation and application of commercial projects, software project testing and implementation, and application and optimization of enterprise mainstream development framework, etc.

Plumelog Distributed Log Component User Manual

1. Introduce Maven Dependency Configuration Introduce Maven Dependency Configuration Note: If this item is not configured, no link information will be displayed on the interface. The principle of this module is to use the springAOP tangent to generate a link log. The core is to configure springAOP. If you are not familiar with springAOP before configuration, please familiarize yourself with the suggestions.

2. Our project is a distributed system, but there is no distributed log system. It is extremely painful to check the log every time it is declassed. When N terminals are opened, the shell knocks off, which is extremely inefficient and ELK is decisively introduced.

3. Both are more efficient than expressJS. We also used Red.Is as a cache, instead of doing analysis tasks directly here, is to improve the docking efficiency with Pusher as much as possible. After all, the production speed of logs is very fast, but network transmission is relatively inefficient.

Flume Quick Start

1. Flume writes the Event order to the end of the File Channel file, and sets maxFileS in the configuration file The ize parameter configures the size of the data file. When the size of the written file reaches the upper limit, Flume will recreate a new file to store the written Event.

2. Offline log collection tool: Flume Flume introduction core component introduction Flume instance: log collection, suitable scenarios, frequently asked questions.

3. Of course, we can also use this tool to store online real-time data or enter HDFS. At this time, you can use it with a tool called Flume, which is specially used to provide simple processing of data and write to various data recipients (such as Kafka) .

4. In terms of big data development, it mainly involves big data application development, which requires certain programming ability. In the learning stage, it is mainly necessary to learn to master the big data technical framework, including Hadoop, hive, oozie, flume, hbase, k Afka, scala, spark and so on.

5. Big data architecture design stage: Flume distributed, Zookeeper, Kafka.Big data real-time self-calculation stage: Mahout, Spark, storm. Big data zd data acquisition stage: Python, Scala.