This is a simple nginx config example for load balancing made by me, which aims to split traffic to different servers. Nginx is from RU!!!
demo the demo server will rotate you to the site or other site I hosted by refresh the page.
install nginx.
Polling: The polling method is the default method of Nginx load. As the name suggests, all requests are assigned to different services in chronological order. If the service is down, it can be automatically eliminated. After the configuration is as follows, the 8.8.0.1 service, 8.8.0.2 service and 8.8.0.3 service are rotated.
upstream polling {
server 8.8.0.1;
server 8.8.0.2;
server 8.8.0.3;
}
...
server{
...
location / {
proxy_pass http://polling;
}
...
}Weights: Specify the weight ratio of each service. The weight is proportional to the access ratio. It is usually used when the performance of the backend service machine is not uniform. The weight of the good performance is allocated to maximize the performance of the server. After the following configuration, the access ratio of the 8.8.0.2 service will be 8.8.0.1 Serves twice as much.
upstream weights {
server 8.8.0.1 weight=1;
server 8.8.0.2 weight=2;
server 8.8.0.3 weight=3;
}
...
server{
...
location / {
proxy_pass http://weights;
}
...
}Iphash: Each request is allocated according to the hash result of the access ip. After such processing, each visitor accesses a backend service fixedly, as follows (ip_hash can be used in conjunction with weight).
upstream iphash {
ip_hash;
server 8.8.0.1 weight=1;
server 8.8.0.2 weight=2;
server 8.8.0.3 weight=3;
}
...
server {
...
location / {
proxy_pass http://iphash;
}
...
}Least connection: Distribute requests to the service with the fewest connections.(least_conn can be used in conjunction with weight).
upstream least_conn {
least_conn;
server 8.8.0.1 weight=1;
server 8.8.0.2 weight=2;
server 8.8.0.3 weight=3;
}
...
server {
...
location / {
proxy_pass http://least_conn;
}
...
}Fair: Requests are distributed according to the response time of the backend server, and those with short response times are given priority.
upstream fair {
fair;
server 8.8.0.1 weight=1;
server 8.8.0.2 weight=2;
server 8.8.0.3 weight=3;
}
...
server {
...
location / {
proxy_pass http://fair;
}
...
}Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.
Please make sure to update tests as appropriate.