nacos_k8s

使用helm3 安装nacos-server:v2.2.1集群到k8s1.24集群的完整步骤
使用mysql：

root@k8s-master51:/igo/soft/mysql8.4.5# kg svc -n mysql
NAME                               TYPE        CLUSTER-IP    EXTERNAL-IP   PORT(S)          AGE
mysql-cluster-primary              NodePort    10.96.2.79    <none>        3306:30306/TCP   90m
mysql-cluster-primary-headless     ClusterIP   None          <none>        3306/TCP         90m
mysql-cluster-secondary            NodePort    10.96.1.149   <none>        3306:30307/TCP   90m
mysql-cluster-secondary-headless   ClusterIP   None          <none>        3306/TCP         90m

env

k8s 1.24
helm v3.19.2
nacos-server:v2.2.1
https://hub.docker.com/r/nacos/nacos-server/tags

数据库建表，初始化

建库
kubectl exec -it -n mysql mysql-cluster-primary-0 — bash
mysql -uroot -p

CREATE DATABASE nacos_config DEFAULT CHARACTER SET utf8mb4;
CREATE USER 'nacos'@'%' IDENTIFIED BY 'Nacos@2026';
GRANT ALL PRIVILEGES ON nacos_config.* TO 'nacos'@'%';
FLUSH PRIVILEGES;

下载建表语句：
wget -O mysql-schema-2.2.1.sql https://raw.githubusercontent.com/alibaba/nacos/2.2.1/distribution/conf/mysql-schema.sql
拷贝进容器
kubectl -n mysql cp /igo/soft/nacos/mysql-schema-2.2.1.sql mysql-cluster-primary-0:/tmp/mysql-schema-2.2.1.sql

导入表结构
mysql> use nacos_config;
Database changed
mysql> source /tmp/mysql-schema-2.2.1.sql;
Query OK, 0 rows affected, 3 warnings (0.03 sec)

随时测一把连通性：

kubectl -n nacos run mysql-client --rm -it --restart=Never \
  --image=mysql:8.0 \
  --command -- bash

mysql -h mysql-cluster-primary.mysql.svc.cluster.local -P 3306 -u nacos -p
mysql> use nacos_config
mysql> show tables;

安装

git clone https://github.com/nacos-group/nacos-k8s.git
cd nacos-k8s/helm
helm show values . > values-default.yaml

！！！由于新版的chart和2.2.1app不匹配，探针路径问题导致POD永远无法ready，需要关闭探针：
sed -i.bak '/startupProbe:/,/path: \/v3\/console\/health\/liveness/d' templates/statefulset.yaml


JWT_TOKEN=$(openssl rand -base64 32 | tr -d '\n')
helm upgrade --install nacos . -n nacos \
  --set global.mode=cluster \
  --set nacos.replicaCount=3 \
  --set nacos.image.tag=v2.2.1 \
  --set service.type=NodePort \
  --set service.port=8848 \
  --set service.nodePort=30000 \
  --set nacos.storage.type=mysql \
  --set nacos.storage.db.host=mysql-cluster-primary.mysql.svc.cluster.local \
  --set nacos.storage.db.port=3306 \
  --set nacos.storage.db.name=nacos_config \
  --set nacos.storage.db.username=nacos \
  --set nacos.storage.db.password='Nacos@2026' \
  --set-string nacos.authToken="$JWT_TOKEN" \
  --set-string nacos.identityKey='serverIdentity' \
  --set-string nacos.identityValue='security'

后续密钥可以找回：
kubectl -n nacos exec sts/nacos -- env | grep NACOS_AUTH

会出现 POD 无法ready，log没有报错的情况，原因：探针配错版本（3.x chart 的探针规则（/v3/... + 8080），而镜像是 2.2.1）
    Liveness:  http-get http://:8080/v3/console/health/liveness delay=10s timeout=10s period=5s #success=1 #failure=3
    Startup:   http-get http://:8080/v3/console/health/readiness delay=180s timeout=10s period=5s #success=1 #failure=3

正确探针：kubectl exec -it nacos-0 -n nacos -- curl -i http://127.0.0.1:8848/nacos/actuator/health


文件写死： templates/statefulset.yaml
解决：（关闭探针）
sed -i.bak '/startupProbe:/,/path: \/v3\/console\/health\/liveness/d' templates/statefulset.yaml


kubectl -n nacos logs nacos-0 --tail=50
kubectl -n nacos logs nacos-1 --tail=50
kubectl -n nacos logs nacos-2 --tail=50

output

NAME: nacos
LAST DEPLOYED: Fri Apr 10 09:08:10 2026
NAMESPACE: nacos
STATUS: deployed
REVISION: 1
TEST SUITE: None
NOTES:

Get the application URL by running these commands:
export NODE_PORT=$(kubectl get svc nacos-cs –namespace nacos -o jsonpath=”{.spec.ports[3].nodePort}” )
export NODE_IP=$(kubectl get nodes –namespace nacos -o jsonpath=”{.items[0].status.addresses[0].address}”)
echo http://$NODE_IP:$NODE_PORT/index.html
MODE:
standalone: you need to modify replicaCount in the values.yaml, .Values.replicaCount=1
cluster: kubectl scale sts nacos-nacos –replicas=3

默认账密： nacos/nacos

回退重装命令

kubectl delete sts nacos -n nacos
helm unintstall nacos -n nacos
kubectl delete all -l app.kubernetes.io/instance=nacos -n nacos --force --grace-period=0
kubectl delete sts nacos -n nacos --force --grace-period=0
kubectl delete configmap nacos-cm -n nacos
kubectl delete svc -n nacos --all
kubectl get all -n nacos

离线安装

docker pull nacos/nacos-server:v2.2.1
docker pull nacos/nacos-peer-finder-plugin:1.1
 nacos-peer-finder用来在 Kubernetes 集群模式下自动发现同组 Nacos 节点
docker save -o nacos-server-v2.2.1.tar nacos/nacos-server:v2.2.1
docker save -o nacos-peer-finder-plugin-1.1.tar nacos/nacos-peer-finder-plugin:1.1

docker load -i nacos-server-v2.2.1.tar
docker load -i nacos-peer-finder-plugin-1.1.tar
docker images | grep -E 'nacos-server|peer-finder'

Post Views: 8

env

数据库建表，初始化

安装

output

回退重装命令

离线安装

igozhang