Linux and Docker: How to implement a highly available container cluster?

Linux and Docker: How to implement a highly available container cluster?

Abstract: With the development of container technology, more and more enterprises are gradually deploying applications into containers. In a production environment, achieving high availability for a container cluster is crucial. This article will introduce how to use Linux and Docker to build a highly available container cluster, and demonstrate the specific implementation method through code examples.

1. Build a Docker Swarm cluster
   Docker Swarm is a native container cluster management tool provided by Docker. It allows us to form multiple Docker hosts into a cluster and uniformly manage the deployment and operation of containers.

First, install Docker on each Docker host that you want to join the cluster. Then, select a host as the management node of the Swarm cluster and run the following command to initialize the cluster:

$ docker swarm init --advertise-addr <MANAGER-IP>

Where, 50c92af3965d072a01ca8b69c30c4169 is the IP address of the management node. Through the above command, we will get a token for other hosts to join the cluster. Next, run the following command on other hosts that want to join the cluster:

$ docker swarm join --token <TOKEN> <MANAGER-IP>:<PORT>

Among them, fe44cc81ab4c20272595cc1edf562112 is the token obtained in the previous step, 50c92af3965d072a01ca8b69c30c4169 and a3b0c87895079be75e30be94102cc20b are management respectively. The IP address and port number of the node.

2. Configuring Service Discovery
   Each node in the container cluster needs to be able to discover and access other nodes. To implement service discovery, we can use tools such as Consul or Etcd. Here we take Consul as an example.

First, install and start Consul on all Docker hosts. Then, run the following command to create a Consul service:

$ docker service create --name consul --publish 8500:8500 --constraint 'node.role == manager' gliderlabs/consul-server -bootstrap

This command creates a service named consul on the management node of the Swarm cluster and maps the container's 8500 port to the host's 8500 port.

Next, run the following command on other nodes to join the Consul service:

$ docker service create --name consul --network <NETWORK> gliderlabs/consul-agent -join consul

Where, 1d4d001a3682e7356cdcdf6423e29e0f is the network name of the Swarm cluster.

3. Create container service
   In a highly available container cluster, we need to deploy applications to multiple container instances and distribute requests through load balancing. Docker Swarm provides a concept called service for managing container services.

First, prepare a Docker image that contains the application we want to deploy. Then, run the following command to create a service:

$ docker service create --name <SERVICE-NAME> --replicas <REPLICAS> --publish <PORT> <IMAGE>

where 1f69d38ff7f38c32e75a871d13a954c2 is the name of the service, f9137ba0285552c9339eaf6686d35256 is the number of container instances to be deployed, a3b0c87895079be75e30be94102cc20b is the number to be mapped The port number, 5a0f7277d940330d201ee28c00ca2724 is the Docker image where the application is located.

4. Run Container Service
   The service created in the previous step will automatically deploy container instances on multiple nodes in the cluster.

You can use the following command to view the running status of the service and the distribution of container instances:

$ docker service ls
$ docker service ps <SERVICE-NAME>

5. Achieve load balancing
   In order to achieve load balancing, we can use Docker Swarm's built-in load balancing function. Through the name and port number of the service, you can access multiple container instances behind the service.
6. Realize automatic expansion
   When the load in the container cluster becomes larger, we can achieve automatic expansion by increasing the number of container instances.

You can use the following command to expand the number of instances of the service:

$ docker service scale <SERVICE-NAME>=<REPLICAS>

Where, 1f69d38ff7f38c32e75a871d13a954c2 is the name of the service, f9137ba0285552c9339eaf6686d35256 is the container to be expanded Number of instances.

Summary: Using Linux and Docker to build a highly available container cluster not only improves the availability of applications, but also provides elastic scaling and load balancing capabilities. Through reasonable configuration and management, we can achieve efficient and stable container services.

Code example:

1. Initialize Swarm cluster:

   $ docker swarm init --advertise-addr 192.168.0.1

2. Join Swarm cluster:

   $ docker swarm join --token <TOKEN> 192.168.0.1:2377

3. Create Consul service:

   $ docker service create --name consul --publish 8500:8500 --constraint 'node.role == manager' gliderlabs/consul-server -bootstrap

4. Join Consul service:

   $ docker service create --name consul --network my-network gliderlabs/consul-agent -join consul

5. Create application service:

   $ docker service create --name my-service --replicas 3 --publish 8080:8080 my-app

6. Number of extended service instances:

   $ docker service scale my-service=5

The above is the detailed content of Linux and Docker: How to implement a highly available container cluster?. For more information, please follow other related articles on the PHP Chinese website!