| commit | 43d1c25234d4d8ee9baa6deb9b0717d017f16ec3 | [log] [tgz] |
|---|---|---|
| author | Sixh-PrFor <chenbing399@163.com> | Sun Oct 23 15:51:04 2022 +0800 |
| committer | GitHub <noreply@github.com> | Sun Oct 23 15:51:04 2022 +0800 |
| tree | 57459c055def4517bbd79f8c8198f929b8ebf159 | |
| parent | fbd0b4af24ca6f27b275e1e308f26c25fbdc20dd [diff] | |
| parent | e7de1ec1f4989536f80dc771c065f61f89edf2a1 [diff] |
Merge branch 'main' into impactCn-patch-1
This module provided SDK to watch available ShenYu instance list as upstream nodes by Service Register Center for OpenResty.
In the cluster mode, Apache ShenYu supports the deployment of multiple ShenYu instances, which may have new instances joining or leaving at any time. Hence, Apache ShenYu introduces Service Discovery modules to help client to detect the available instances. Currently, Apache ShenYu Bootstrap already supports Apache Zookeeper, Nacos, Etcd, and consul. Client or LoadBalancer can get the available ShenYu instances by those Service register center.
The first, clone the source from GitHub.
git clone https://github.com/apache/incubator-shenyu-nginx
Then, build from source and install.
cd incubator-shenyu-nginx luarocks make rockspec/shenyu-nginx-main-0.rockspec
Modify the Nginx configure, create and initialize the ShenYu Register to connect to the target register center. The module will fetch the all of ShenYu instances which are registered to Etcd in the same cluster. It works like Etcd client to watch(based on long polling) ShenYu instance lists.
Here is an example for Etcd.
init_worker_by_lua_block {
local register = require("shenyu.register.etcd")
register.init({
balancer_type = "chash",
etcd_base_url = "http://127.0.0.1:2379",
})
}
balancer_type specify the balancer. It has supported chash and round robin.etcd_base_url specify the Etcd server.(Currently, authentication is not supported.)Add an upstream block for ShenYu and enable to update upstream servers dynamically. This case will synchronize the ShenYu instance list with register center. And then pick one up for handling the request.
upstream shenyu {
server 0.0.0.1; -- bad
balancer_by_lua_block {
require("shenyu.register.etcd").pick_and_set_peer()
}
}
Finally, restart OpenResty.
openresty -s reload
Here is a completed example working with ETCD.
Modify the Nginx configure, create and initialize the ShenYu Register to connect to target register center. Here is an example for Nacos.
init_worker_by_lua_block {
local register = require("shenyu.register.nacos")
register.init({
shenyu_storage = ngx.shared.shenyu_storage,
balancer_type = "chash",
nacos_base_url = "http://127.0.0.1:8848",
username = "nacos",
password = "naocs",
})
}
balancer_type specify the balancer. It has supported chash and round robin.nacos_base_url specify the Nacos server address.username specify the username to log in Nacos. (it is only required when Nacos auth enable)password specify the password to log in Nacos.Modify the upstream to enable to update upstream servers dynamically. This case will synchronize the ShenYu instance list with register center. And then pick one up for handling the request.
upstream shenyu {
server 0.0.0.1; -- bad
balancer_by_lua_block {
require("shenyu.register.nacos").pick_and_set_peer()
}
}
Finally, restart OpenResty.
openresty -s reload
Here is a completed example working with Nacos.
Modify the Nginx configure, create and initialize the ShenYu register to connect to target register center. Listen for changes to the node via the zookeeper watch event. Here is an example of the zookeeper configuration.
init_worker_by_lua_block { local register = require("shenyu.register.zookeeper") register.init({ servers = {"127.0.0.1:2181","127.0.0.1:2182"}, shenyu_storage = ngx.shared.shenyu_storage, balancer_type = "roundrobin" }); }
servers zookeeper cluster address.balancer_type specify the balancer. It has supported chash and round robin.Modify the upstream to enable to update upstream servers dynamically. This case will synchronize the ShenYu instance list with register center. And then pick one up for handling the request.
upstream shenyu { server 0.0.0.1; balancer_by_lua_block { require("shenyu.register.zookeeper").pick_and_set_peer() } }
Finally, restart OpenResty.
openresty -s reload
Here is a completed example working with Zookeeper.