What does docker cluster mean?

In docker, a cluster is a group of multiple machines running docker and joined in a group. It is a group of service entities that work together. A cluster is used to provide a service platform that is more scalable and available than a single service entity. After joining the cluster, you can continue to run your own docker commands, but now these machines are executed on the cluster by the cluster manager, and the cluster manager can use different Policies to manage running containers.

The operating environment of this tutorial: linux7.3 system, docker version 19.03, Dell G3 computer.

What does docker cluster mean

A cluster is a group of multiple machines running docker and joined in a group.

After joining the cluster, you can continue to run your own docker commands, but now these machines are executed on the cluster by the cluster manager. The machines in the cluster can be physical or virtual. After joining the group, each machine is called a node.

The cluster manager can use different strategies to manage running containers, for example: the emptiest node command is used to fill the least utilized containers; global ensures that each machine can only get one instance of the specified container. You can write these policies to a file to enforce the cluster manager's policies.

The cluster manager is the only machine in the cluster that can execute commands. You can also authorize other machines to join the cluster management work.

A cluster is a group of service entities (can be understood as servers) that work together to provide a service platform that is more scalable and available than a single service entity. From the client's perspective, a cluster looks like a service entity, but in fact a cluster consists of a set of service entities.

Extended knowledge

In the docker cluster service, the following concepts must be understood.

Swarm

Swarm is a cluster of multiple hosts running Docker Engine.

Starting from v1.12, cluster management and orchestration functions have been integrated into Docker Engine. When Docker Engine initializes a Swarm or joins an existing Swarm, it starts Swarm Mode.

When Swarm Mode is not started, Docker executes container commands; after running Swarm Mode, Docker increases the ability to orchestrate services. Docker allows running both Swarm Service and separate containers on the same Docker host.

node

Each Docker Engine in Swarm is a node, and there are two types of nodes: manager and worker.

In order to deploy applications to Swarm, we need to execute the deployment command on the manager node. The manager node will disassemble the deployment task and assign it to one or more worker nodes to complete the deployment.

The manager node is responsible for performing orchestration and cluster management work, keeping and maintaining Swarm in the desired state. If there are multiple manager nodes in Swarm, they will automatically negotiate and elect a leader to perform orchestration tasks.

Woker node accepts and executes tasks dispatched by manager node. In the default configuration, the manager node is also a worker node, but it can be configured as a manager-only node to be responsible for orchestration and cluster management.

The work node will regularly report its own status and the status of the tasks it is executing to the manager node, so that the manager can maintain the status of the entire cluster.

service

service defines the tasks to be performed on the worker node. The main orchestration task of swarm is to ensure that the service is in the desired state.

Give a service example: start an nginx service in swarm, use the image nginx:latest, and the number of copies is 3.

The manager node is responsible for creating this service. After analysis, it is known that three nginx containers need to be started. The task of running the containers is assigned according to the current status of each worker node. For example, two containers are run on worker1 and one is run on worker2. container.

After running for a period of time, worker2 suddenly crashed. The manager monitored this failure and immediately started a new nginx container on worker3. This ensures that the service is in the desired three replica states.

In short, swarm organizes clusters in the form of nodes; at the same time, one or more services can be deployed on each node, and each service can include one or more containers ( container).

Recommended learning: "docker video tutorial"

The above is the detailed content of What does docker cluster mean?. For more information, please follow other related articles on the PHP Chinese website!