K8s 部署
大约 4 分钟
环境需求
- minikube v1.23.0 +
- mysql 8.0 + | MariaDB 10.7 + | Postgres 14 + (Postgres 15 + 推荐)
- redis 6.0 +
- docker
Minikube 配置
K8s配置
API 服务
api/etc/core.yaml
Name: core.api
Host: 0.0.0.0 # 需要 0.0.0.0 以便外部访问
Port: 9100
Timeout: 30000
Auth:
AccessSecret: jS6VKDtsJf3z1n2VKDtsJf3z1n2 # JWT的加密密钥，各个API应保持一致才能解析
AccessExpire: 259200 # 秒，过期时间
Log:
ServiceName: coreApiLogger
Mode: file # 日志模式
Path: /home/ryan/logs/core/api # log 保存路径，使用filebeat同步
Level: info # 日志等级
Compress: false # 日志压缩
KeepDays: 7 # 保存时长（天）
StackCoolDownMillis: 100 # 多少毫秒后再次写入堆栈跟踪。用来避免堆栈跟踪日志过多
RedisConf:
Host: 127.0.0.1:6379 # 改成自己的redis地址
Type: node
# Pass: xxx # 也可以设置密码
CoreRpc:
Target: k8s://simple-admin/corerpc-svc:9101 # 格式 k8s://namespace/service-name:port
Captcha:
KeyLong: 5 # 验证码长度
ImgWidth: 240 # 验证码图片宽度
ImgHeight: 80 # 验证码图片高度
DatabaseConf:
Type: mysql
Path: "127.0.0.1" # 修改成自己的mysql地址
Port: 3306
Config: charset=utf8mb4&parseTime=True&loc=Local # gorm时间转换需要如下配置
DBName: simple_admin # 数据库名，可以自定义
Username: root # 用户名
Password: "123456" # 密码
MaxIdleConn: 10 # 最大空闲连接
MaxOpenConn: 100 # 最大连接数
LogMode: error # log 级别
LogZap: false # 目前log zap还未实现
# 服务监控
Prometheus:
Host: 0.0.0.0
Port: 4000
Path: /metricsrpc/etc/core.yaml
Name: core.rpc
ListenOn: 0.0.0.0:9101
DatabaseConf:
Type: mysql
Path: "127.0.0.1" # 修改成自己的mysql地址
Port: 3306
Config: charset=utf8mb4&parseTime=True&loc=Local
DBName: simple_admin
Username: root # 用户名
Password: "123456" # 密码
MaxIdleConn: 10 # 最大空闲连接
MaxOpenConn: 100 # 最大连接数
LogMode: error
LogZap: false
Log:
ServiceName: coreRpcLogger
Mode: file
Path: /home/ryan/logs/core/rpc # log 保存路径，使用filebeat同步
Encoding: json
Level: info
Compress: false
KeepDays: 7 # 保存时长（天）
StackCoolDownMillis: 100
RedisConf:
Host: 192.168.50.216:6379 # 改成自己的redis地址
Type: node
# Pass: xxx # 也可以设置密码
# 服务监控
Prometheus:
Host: 0.0.0.0
Port: 4001
Path: /metricsDocker镜像编译发布
手动编译
# 设置环境变量
export VERSION=0.0.1 # 版本号
export DOCKER_USERNAME=xxx # docker仓库用户名
export DOCKER_PASSWORD=xxx # docker仓库密码
export REPO=docker.io # docker仓库地址
# 生成镜像
make docker
# 发布镜像
make publish-docker建议使用 gitlab-ci， 项目已默认提供， 需要在 gitlab runner 设置 variable ： $DOCKER_USERNAME 和 $DOCKER_PASSWORD
variables:
VERSION: 0.0.19
REPO: docker.io
stages:
- info
- build
- publish
- clean
info-job:
stage: info
script:
- echo "Start build version $VERSION"
- export VERSION=$VERSION
- export DOCKER_USERNAME=$DOCKER_USERNAME
- export DOCKER_PASSWORD=$DOCKER_PASSWORD
- export REPO=$REPO
build-job:
stage: build
script:
- echo "Compiling the code and build docker image..."
- make docker
- echo "Compile complete."
deploy-job:
stage: publish
environment: production
script:
- echo "Publish docker images..."
- make publish-docker
- echo "Docker images successfully published."
clean-job:
stage: clean
script:
# 删除所有 none 镜像 | delete all none images
- docker images |grep none|awk '{print $3}'|xargs docker rmi
- echo "Delete all none images successfully."部署流程
- 生成docker镜像
- 上传docker仓库
- 在k8s集群使用命令 kubectl apply -f deploy/k8s/coreapi.yaml 等直接部署
注意
生成镜像和上传仓库建议直接使用gitlab-ci自动发布
coreapi 部署文件详解
coreapi 是所有服务的label和metadata:name
命名空间默认是 default, 可自行修改
apiVersion: apps/v1
kind: Deployment
metadata:
name: core-api
labels:
app: core-api
spec:
replicas: 3
revisionHistoryLimit: 5
selector:
matchLabels:
app: core-api
template:
metadata:
labels:
app: core-api
spec:
serviceAccountName: endpoints-finder
containers:
- name: core-api
image: ryanpower/core-api:0.0.19 # 主要修改此处镜像
ports:
- containerPort: 9100 # 端口， 与 core.yaml 内配置端口相同
readinessProbe:
tcpSocket:
port: 9100
initialDelaySeconds: 5
periodSeconds: 10
livenessProbe:
tcpSocket:
port: 9100
initialDelaySeconds: 15
periodSeconds: 20
resources:
requests:
cpu: 100m # 最低 cpu 需求， 1000m 为一个cpu,测试环境建议小一些
memory: 100Mi # 本地调试我设置了 100 mb, 正式环境可以为 512Mi
limits:
cpu: 1000m # 最高占用 cpu
memory: 1024Mi # 最高占用的内存
volumeMounts:
- name: timezone
mountPath: /etc/localtime
- mountPath: /home/data
name: simple-admin-pv
volumes:
- name: timezone
hostPath:
path: /usr/share/zoneinfo/Asia/Shanghai
- name: simple-admin-pv # log 持久化卷
persistentVolumeClaim:
claimName: simple-admin-pv-claim
---
apiVersion: v1
kind: Service
metadata:
name: core-api-svc
labels:
app: core-api-svc
spec:
type: NodePort
ports:
- port: 9100
targetPort: 9100
name: api
protocol: TCP
selector:
app: core-api
---
apiVersion: v1
kind: Service
metadata:
name: core-api-svc
labels:
app: core-api-svc
spec:
ports:
- port: 4000
name: prometheus
targetPort: 4000
selector:
app: core-api
---
# 服务监控
apiVersion: monitoring.coreos.com/v1
kind: ServiceMonitor
metadata:
name: core-rpc
labels:
serviceMonitor: prometheus
spec:
selector:
matchLabels:
app: core-rpc-svc
endpoints:
- port: prometheus
---
# autoscaling 用于动态增加负载，通过 metric-server 获取使用率，目前获取 metric 还有些问题，近期会修复
apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
name: core-api-hpa-c
labels:
app: core-api-hpa-c
spec:
scaleTargetRef:
apiVersion: apps/v1
kind: Deployment
name: core-api
minReplicas: 3 # 最小副本
maxReplicas: 10 # 最大副本
metrics:
- type: Resource
resource:
name: cpu
target:
type: Utilization
averageUtilization: 80 # 平均使用率 80%
---
apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
name: core-api-hpa-m
labels:
app: core-api-hpa-m
spec:
scaleTargetRef:
apiVersion: apps/v1
kind: Deployment
name: core-api
minReplicas: 3
maxReplicas: 10
metrics:
- type: Resource
resource:
name: memory
target:
type: Utilization
averageUtilization: 80core rpc 和 backend ui 相似
前端 nginx 请求设置
simple-admin-backend-ui/deploy/default.conf
server {
listen 80;
listen [::]:80;
server_name localhost;
location / {
root /usr/share/nginx/html;
index index.html index.htm;
try_files $uri $uri/ /index.html;
}
location /sys-api/ {
proxy_pass http://core-api-svc.default.svc.cluster.local:9100/;
}
# location /file-manager/ {
# proxy_pass http://file-api-svc.default.svc.cluster.local:9102/;
# }
}注意
proxy_pass 格式 http://{service-name}.{namespace}.svc.cluster.local:{port}/
快速部署
执行 deploy/k8s/setup.sh