Implementation of load Balancing configuration
1. Reverse Proxy
2, implementation agent (192.168.1.161 Agent to 192.168.1.62)
Worker_processes 1;
events{
worker_connections 1024;
}
http{
include mine.types;
Default_type Application/octet-stream;
Senfile on;
Keepalive_timeout;
server{
Listen;
server_name locahost;
Location/{
#root html;
Index index.html index.htm;
Proxy_pass http://192.168.1.62:8080; # Agent to http://192.168.1.62:8080
}
error_page 502 503 504/50x.html;
Location =/50x.html {
root html;
}
}} Restart Nginx:./nginx-s Reload
Second, the processing strategy of session during load balancing
1. Achieve Basic load Balancing
# Set up the server group, the policy is the default policy, the average polling (1) upstream Tomcats {server 192.168.1.62:8080;
Server 192.168.1.63:8080;
# Set up the server group, the policy is Ip_hash policy, the user is locked to a server (2) upstream Tomcats {ip_hash;
Server 192.168.1.62:8080;
Server 192.168.1.63:8080;
# Configure server group information for load Balancing (5) Upstream backends{server 192.168.1.62:8080 backup;
Server 192.168.1.63;
# Weight: Weight, the higher the value the larger the load (by weight decision polling) # server 192.168.1.4 weight=5;
# Backup: back-up machine, only if not the backup machine is dead to enable; server 192.168.1.4 backup;
# Down: Stop sign, will not be accessed, mainly for temporary downtime maintenance server 192.168.1.65 down;
# Max_fails: To the specified number of times that the server hangs up # Fail_timeout: How long before you hang up to test if the server 192.168.1.66 max_fails=2 fail_timeout=60s has been restored;
} worker_processes 1;
events{worker_connections 1024; http{include mine.types;
Default_type Application/octet-stream;
Senfile on;
Keepalive_timeout 65;
server{Listen 80;
server_name locahost;
Location/{#root html;
Index index.html index.htm; ProXy_pass Http://tomcats;
# Send current access to server group} Error_page 502 503 504/50x.html;
Location =/50x.html {root html; }
}
}
Each time you refresh, the server being proxied is different, the default average polling. The balance strategy includes
(1) None polling, weight determined by weight
(2) Ip_hash, a user bound to a server
(3) Fair strategy, according to the speed of the server response dynamic allocation, response fast load
(4) Url_hash,url binding on a server
Three, the problem that the application code should pay attention to in the cluster environment