7. Swarm builds clusters

5 Line code , Can let wechat small program on the shelf to their own APP in | Register and send it to Dajiang 、 Huawei 、 Cherry keyboard ！>>>

Use dokcer Problems that will come up ：

• How to manage so many containers ？
• How to easily expand horizontally
• If the container down 了 , How can I resume the boot ？
• How to update the container without affecting the business ？
• How to monitor and track these containers ？
• How to schedule container creation ？
• Protect private data ？

Docker It comes with a container arrangement tool ：Swarm Mode

Manager Node as manager , To maintain high availability , You can't deploy just one , If there are multiple deployment, there is a problem of how to synchronize the state .docker Built in distributed database , adopt raft Agreement assurance Manager Information is synchronized , There's no cleft brain .

Worker Nodes are work nodes , It also needs to be with other Worker Node data synchronization , Use Gossip The protocol synchronizes the information .

We go through swarm manager Node to deploy a service, We don't know about this service Which container nodes will eventually run on .swarm manager According to some strategies , For example, memory. , Hard disk resources, etc , take service Deployed in some container node in .

3 nodes swarm cluster setup

1. On the first machine , establish docker swarm manager command

docker swarm init --advertise-addr=192.168.8.117      This machine IP

Create success , Returns the docker swarm join --token SWMTKN-1-1aud6ztpiwgujisqn20gb9lfldziidj0dz7lww3puod36t5j5d-c6azyl8d0rhhzd8c25u9laxxp 192.168.8.117:2377    For others docker swarm join This machine , Build clusters .

2. On the second machine , establish docker swarm And add the first machine , command

docker swarm join --token SWMTKN-1-1aud6ztpiwgujisqn20gb9lfldziidj0dz7lww3puod36t5j5d-c6azyl8d0rhhzd8c25u9laxxp 192.168.8.117:2377

Something goes wrong

Error response from daemon: rpc error: code = Unavailable desc = all SubConns are in TransientFailure, latest connection error: connection error: desc = "transport: Error while dialing dial tcp 192.168.8.117:2377: connect: no route to host"

reason

 If you want to use swarm function , Need at all manager node Turn on the node 2377 port .

terms of settlement

Open designated port  firewall-cmd --zone=public --add-port=2377/tcp --permanent

Restart the firewall to take effect systemctl restart firewalld

see swarm node

docker node ls

ID                            HOSTNAME                STATUS              AVAILABILITY        MANAGER STATUS      ENGINE VERSION
5xeakkex0klrezwb1p5y0pcja     localhost.localdomain   Ready               Active                                  19.03.13
rk1pu6zgxkh278262d5yprzqi     localhost.localdomain   Ready               Active                                  19.03.13
ulms2qb8gvnehsmgx7d7dcx9a *   localhost.localdomain   Ready               Active              Leader              19.03.13

Steps to build ：

1、 Environmental preparation ：
1.1、 Three are ready to be installed docker engine Of centos/Ubuntu System host （docker Version must be in 
1.12 Version above , The old version does not support swarm）
1.2、docker Container host ip Fixed address , All work nodes in the cluster must have access to the management node 
1.3、 The cluster management node must use the appropriate protocol and ensure that the port is available 
 Cluster management communication ：TCP, port 2377    
 Node communication ：TCP and UDP, port 7946    
 Overlay networks (docker The Internet )：UDP, port 4789   overlay drive     
 explain ： Three container hosts ip The addresses are ：
192.168.200.162（ The management node ）
192.168.200.163（ Work node ）
192.168.200.158（ Work node ）
 The host names are ：manager1、work1 as well as work2
vim /etc/hostname  ( You need to restart after the modification )
2、 establish docker swarm
2.1、 stay manager1 Create... On the machine docker swarm colony 
docker swarm init ‐‐advertise‐addr 192.168.200.162
（‐‐advertise‐addr Will be IP The address of the machine is set as the cluster management node ; If it's a single node , There is no need to refer to 
 Count ）
2.2、 View management node cluster information ：
docker node ls
3、 towards docker swarm Add a work node in the ： Execute the following commands in the two work nodes respectively ,ip The address is 
manager Node 
3.1、 Add two work node 
docker swarm join ‐‐token xxx 192.168.200.138:2377  （worker1）
docker swarm join ‐‐token xxx 192.168.200.138:2377  （worker2）
（‐‐token xxx: Add authentication information of work nodes to the specified cluster ,xxx Authentication information is creating docker
swarm When the ）
3.2、 Continue to view the differences between the management node cluster information and the previous one 
docker node ls
4、 stay docker swarm Deploy services in 
 stay Docker Swarm When deploying services in a cluster , Both available Docker Hub Start the service with a self-contained image on the , also 
 You can use self access Dockerfile Build the image to start the service . If you use yourself through Dockerfile Built 
 To start a service, you must first push the image to Docker Hub Central warehouse . In order to facilitate the study of readers , here 
 To use the Docker Hub Bring your own alpine Image as an example to deploy cluster services 
4.1、 Deployment Services 
docker service create ‐‐replicas 1 ‐‐name helloworld alpine ping
docker.com
docker service create Instructions ： Used in Swarm Create a cluster based on alpine Mirror service 
‐‐replicas Parameters ： Specifies that the service has only one replica instance 
‐‐name Parameters ： Specify the service name after successful creation as helloworld
ping docker.com Instructions ： Represents the command to be executed after the service is started 
5. see docker swarm Services in the cluster 
 View the list of services ：docker service ls
 View details of deploying specific services ：docker service inspect  The service name 
 View the allocation and operation of services on cluster nodes ：docker service ps  The service name 
6、 Change the number of copies 
 stay manager1 On , Change the number of service copies （ The created copies are randomly assigned to different nodes ）
docker service scale helloworld=5
7、 Delete service （ At the management node ）
docker service rm  The service name 
8、 Access the service 
8.1、 View the network list in the cluster environment ：docker network ls
8.2、 stay manager1 Create a overlay For driven networks （ Network connection used by default ingress）
docker network create ‐d=overlay my‐multi‐host‐network
8.3、 In the cluster management node manager1 Put a nginx service 
docker service create \
  ‐‐network my‐multi‐host‐network \
  ‐‐name my‐web \
  ‐p 8080:80 \
  ‐‐replicas 2 \
  nginx
8.3、 View the running status of the service in the management node ：
docker service ps my‐web
8.4、 Access test