From this chapter we'll focus on RPC usage in zero system, because they are micro environment, we need to prepare more service nodes here.
Demo Projects and environment
| Http Port | Ipc Port | Ipc Service Name | Project | Role |
|---|---|---|---|---|
| 6100 | -- | -- | up-athena | Api Gateway |
| 6201 | -- | -- | up-atlas | Common Service |
| 6301 | 6311 | ipc-epimetheus | up-epimetheus | Originator |
| 6401 | 6411 | ipc-coeus | up-coeus | Coordinator A |
| 6402 | 6412 | ipc-crius | up-crius | Coordinator B |
| 6403 | 6413 | ipc-cronus | up-cronus | Coordinator C |
| 6501 | 6511 | ipc-hecate | up-hecate | Terminator |
Why we need so many projects ? Because in micro service tutorials we'll focus on Service Registry, Discovery and Communication, we consider use more projects to describe different scenarios and let developers know how to develop micro services in zero framework.
You can refer D10074 - Configuration, vertx-etcd3.yml tutorials for etcd configuration for zero node. Because of our examples here we listed following different configurations
- Api Gateway Configuration
- Common Service ( None Rpc Server )
- Service Node ( With Rpc Server, Originator/Coordinator/Terminator )
As we introduced in previous tutorials, here all the configuration for etcd3 should be the same as following, in my demo
we used around 3 nodes of etcd 6181, 6180, 6179, you can correct the configuration info based on your own environment
setting. The last thing is that the micro for all of our zero instances are zero-istio:
vertx-etcd3.yml
etcd:
micro: zero-istio
nodes:
- host: localhost
port: 6181
- host: localhost
port: 6180
- host: localhost
port: 6179
timeout: 2The first thing is that to configure api gateway, the whole src/main/resources folder structure is as following:
Api gateway will scan etcd service center to see the registered services here,
vertx.yml
zero:
lime: etcd3
vertx:
instance:
- name: zero-micro
options:
# Fix block 2000 limit issue.
maxEventLoopExecuteTime: 30000000000vertx-server.yml
server:
- name: up-athena
type: api
config:
port: 6100
host: 0.0.0.0Be careful of the
typeproperty in vertx-server.yml, this property in api gateway must beapiinstead ofhttp. When your zero instance will run as Api Gateway, you must set uniqueapiconfiguration invertx-server.ymlup.god.file.
When you start zero api gateway, you should see following output message.
......
[ ZERO ] ( node = server, type = API ) filtered configuration port set = [6100].
......
[ ZERO ] ( Etcd Center ) Etcd configuration center is enabled !
[ ZERO ] ( node = server, type = API ) filtered configuration port set = [6100].
[ ZERO ] ( Etcd Center: zero-istio ) Etcd Client timeout = "2s" with nodes = 3
[ ZERO ] ( Etcd Center ) Etcd network checking has been passed successfully !
Setting up Etcd4j Netty client
[ ZERO ] The raw data ( node = circuit, type = Circuit ) before validation is {}.
......
[ ZERO ] ( Api Gateway ) ZeroApiAgent has been started successfully. Endpoint: http://0.0.0.0:6100/.
......Then you can wait for some seconds, there should be some additional logs output as following every 3 seconds.
vert.x-worker-thread-2 ...... \
[ ZERO ] ( Discovery ) Records ( added = 0, updated = 5, deleted = 0 ) have been refreshed!
vert.x-worker-thread-3 ...... \
[ ZERO ] ( Discovery ) Records ( added = 0, updated = 5, deleted = 0 ) have been refreshed!In zero api gateway there are two threads scanned etcd service center every 2 seconds, when you start new service and registry into etcd service center, these two threads will scan the added new threads and put into service discovery in api gateway. In this kind of situation you can implement service up/down at any time.
In the example we also started a common service, this service does not expose Rpc port but it will use RpcClient
instead.
vertx.yml
zero:
lime: etcd3, rpc
vertx:
instance:
- name: zero-micro
options:
# Fix block 2000 limit issue.
maxEventLoopExecuteTime: 30000000000vertx-server.yml
server:
- name: up-atlas
type: http
config:
port: 6201
host: 0.0.0.0There are some new configuration that should be mentioned is that if you want to use RpcClient to call remote rpc
services, you must do some additional configuration.
From vert.x 3.5.1, the api of Rpc
useSslhas been changed and Rpc Server could useHttpServerOptionsinstead of the old configuration, be careful.
vertx-rpc.yml
rpc:
ssl: false
uniform:
type: PEMvertx-inject.yml
rpc: io.vertx.mod.plugin.rpc.RpcInfixAbove Rpc configuration is required.
When you started this service, you should see more logs output that you haven't seen before:
......
[ ZERO ] ( Etcd Center ) Etcd configuration center is enabled !
[ ZERO ] ( Etcd Center: zero-istio ) Etcd Client timeout = "2s" with nodes = 3
[ ZERO ] ( Etcd Center ) Etcd network checking has been passed successfully !
Setting up Etcd4j Netty client
HV000001: Hibernate Validator 6.0.7.Final
......
[ ZERO ] ---> ZERO://MICRO/REGISTRY/START ( Http Server ) ZeroHttpAgent ( name = up-atlas ) is sending data to internal address.
[ ZERO ] ( Etcd Center ) The status up-atlas of service RUNNING has been registried to /zero/zero-istio/endpoint/services/up-atlas:10.0.0.7:6201
[ ZERO ] ( Etcd Catalog ) The following routes has been push to:
[ Up Micro ] <Application Name> = "zero-istio",
[ Up Micro ] Configuration Path = /zero/zero-istio/endpoint/routes/up-atlas:10.0.0.7:6201,
[ Up Micro ] Service Name = up-atlas,
[ Up Micro ] EndPoint = http://10.0.0.7:6201
[ Up Micro ] Route Uris =
[ Up Micro ]
[ Up Micro ] √ Successfully to registered Routes, wait for discovery......SUCCESS √
[ ZERO ] ZERO://MICRO/REGISTRY/START <--- ( Micro Worker ) ZeroHttpRegistry ( name = up-atlas ) get data from internal address.From above logs we could see the configuration path on etcd .
The last configuration is service node configuration: up-coeus, up-crius, up-cronus, up-epimetheus, up-hecate , except
service name, port, other configuration is the same.
Here we only list the project up-crius , the folder structure of this project is as following:
vertx.yml
zero:
lime: etcd3, rpc
vertx:
instance:
- name: zero-micro
options:
# Fix block 2000 limit issue.
maxEventLoopExecuteTime: 30000000000vertx-rpc.yml
rpc:
ssl: false
uniform:
type: PEMvertx-server.yml
server:
- name: up-crius
type: http
config:
port: 6402
host: 0.0.0.0
- name: ipc-crius
type: ipc
config:
port: 6412
host: 0.0.0.0Then when you start this node here, you should see following message:
......
[ ZERO ] ( Etcd Center ) Etcd configuration center is enabled !
[ ZERO ] ( Etcd Center: zero-istio ) Etcd Client timeout = "2s" with nodes = 3
[ ZERO ] ( Etcd Center ) Etcd network checking has been passed successfully !
Setting up Etcd4j Netty client
......
# Http Endpoint part, it's the same as Common
[ ZERO ] ---> ZERO://MICRO/REGISTRY/START ( Http Server ) ZeroHttpAgent ( name = up-crius ) is sending data to internal address.
[ ZERO ] ( Etcd Center ) The status up-crius of service RUNNING has been registried to /zero/zero-istio/endpoint/services/up-crius:10.0.0.7:6402
[ ZERO ] ( Etcd Catalog ) The following routes has been push to:
[ Up Micro ] <Application Name> = "zero-istio",
[ Up Micro ] Configuration Path = /zero/zero-istio/endpoint/routes/up-crius:10.0.0.7:6402,
[ Up Micro ] Service Name = up-crius,
[ Up Micro ] EndPoint = http://10.0.0.7:6402
[ Up Micro ] Route Uris =
[ Up Micro ]
[ Up Micro ] √ Successfully to registered Routes, wait for discovery......SUCCESS √
[ ZERO ] ZERO://MICRO/REGISTRY/START <--- ( Micro Worker ) ZeroHttpRegistry ( name = up-crius ) get data from internal address.
# Rpc Service part
[ ZERO ] ---> ZERO://MICRO/IPC/START ( Rpc Server ) ZeroRpcAgent ( name = ipc-crius ) is sending data to internal address.
[ ZERO ] ( Etcd Center ) The status ipc-crius of service RUNNING has been registried to /zero/zero-istio/ipc/services/ipc-crius:10.0.0.7:6412
[ ZERO ] ( Etcd Catalog ) The following routes has been push to:
[ Up Rpc ] <Application Name> = "zero-istio",
[ Up Rpc ] Configuration Rpc Point = /zero/zero-istio/ipc/routes/ipc-crius:10.0.0.7:6412,
[ Up Rpc ] Service Name = ipc-crius,
[ Up Rpc ] Ipc Channel = grpc://10.0.0.7:6412
[ Up Rpc ] Ipc Address =
[ Up Rpc ] √ Successfully to registered IPCs, wait for community......SUCCESS √
[ ZERO ] ZERO://MICRO/IPC/START <--- ( Micro Worker ) ZeroRpcRegistry ( name = ipc-crius ) get data from internal address.You should repeat the configuration steps of 4 on
up-coeus, up-crius, up-cronus, up-epimetheus, up-hecate
When you have finished all the configuration, your zero system micro environment has been finished, in this kind of situation you can continue the further tutorial for all the micro service demos ( Especially for Rpc demos ).