Introduction to Zookeeper and pseudo-distributed cluster construction

Introduction to Zookeeper

Distributed system definition and imminent problems
ZooKeeper's main use scenario is as a distributed collaborative service for distributed systems. We define a distributed system as: a distributed system is a system composed of multiple software running independently across multiple physical hosts. By analogy, a distributed system is a group of people working together. There are many people and the power of each server is limited. However, through a distributed system, a cluster composed of n servers, the computing power can be expanded infinitely. .

The advantages are obvious, more people work fast and back up each other, but the disadvantages are also obvious. We can imagine taking a small R&D team as an example of software development. Suppose we have a 5-person project team. To start a system development, the project team faces the following problems:

Listed in the figure are the problems that the project team will face. These problems also occur every day in our daily work. It doesn't feel how complicated, but this is because our human brain is a supercomputer and can deal with these problems flexibly. Moreover, the information exchange in reality does not rely on the network, and there will be no information imbalance due to network delays or interruptions, and the actual handling of the above problems is not rigorous, which causes many problems. Think about it, have there been ambiguities in the assignment of tasks due to poor communication in the project? Is the task unable to proceed due to the resignation of the personnel, or even contact the resigning personnel for assistance? Have there been unreasonable task assignments? Various problems like this must happen in your project. In the real world, we can artificially coordinate, even if something goes wrong, artificially make up for it and work overtime. But it doesn't work in the computer world. Everything must be rigorous, and the above problems should be avoided as much as possible. Therefore, the distributed system must adopt a reasonable way to solve the above problems.

Pseudo-distributed cluster construction

Unfinished,,, to be continued,,,