Kubernetes中无法通过Service名称连接MongoDB问题求助

问题描述 #

在Kubernetes部署MongoDB实例，已创建对应的Pod和Service，但通过Service名称auth-mongo-srv连接MongoDB时出现连接超时错误，使用Service的Cluster IP（10.105.37.75）则可以正常连接。

Service信息 #

NAME             TYPE        CLUSTER-IP       EXTERNAL-IP   PORT(S)     AGE
auth-mongo-srv   ClusterIP   10.105.37.75     <none>        27017/TCP   19s

MongoDB连接代码 #

try {
    await mongoose.connect('mongodb://auth-mongo-srv:27017/auth');
    
    console.log("Connected with MongoDB !!!");
} catch (err) {
    console.log(err);
}

控制台错误信息 #

[INFO] 17:23:25 ts-node-dev ver. 2.0.0 (using ts-node ver. 10.9.1, typescript ver. 
4.8.4)
[auth-service] MongooseServerSelectionError: connection timed out
[auth-service]     at NativeConnection.Connection.openUri 
(/app/node_modules/mongoose/lib/connection.js:824:32)
[auth-service]     at /app/node_modules/mongoose/lib/index.js:381:10
[auth-service]     at 
/app/node_modules/mongoose/lib/helpers/promiseOrCallback.js:41:5
[auth-service]     at new Promise (<anonymous>)
[auth-service]     at promiseOrCallback 
(/app/node_modules/mongoose/lib/helpers/promiseOrCallback.js:40:10)
[auth-service]     at Mongoose._promiseOrCallback 
(/app/node_modules/mongoose/lib/index.js:1234:10)
[auth-service]     at Mongoose.connect 
(/app/node_modules/mongoose/lib/index.js:380:20)
[auth-service]     at /app/src/index.ts:13:24
[auth-service]     at Generator.next (<anonymous>)
[auth-service]     at /app/src/index.ts:8:71 {
[auth-service]   reason: TopologyDescription {
[auth-service]     type: 'Unknown',
[auth-service]     servers: Map(1) { 'auth-mongo-srv:27017' => [ServerDescription] },
[auth-service]     stale: false,
[auth-service]     compatible: true,
[auth-service]     heartbeatFrequencyMS: 10000,
[auth-service]     localThresholdMS: 15,
[auth-service]     setName: null,
[auth-service]     maxElectionId: null,
[auth-service]     maxSetVersion: null,
[auth-service]     commonWireVersion: 0,
[auth-service]     logicalSessionTimeoutMinutes: null
[auth-service]   },
[auth-service]   code: undefined
[auth-service] }

排查与解决步骤 #

1. 验证DNS解析 #

• 在发起连接的客户端Pod内执行DNS查询，确认Service名称能解析到正确的Cluster IP：

  kubectl exec -it <client-pod-name> -- nslookup auth-mongo-srv
  

• 如果解析失败，检查CoreDNS组件状态：

  kubectl get pods -n kube-system | grep coredns
  kubectl logs <coredns-pod-name> -n kube-system
  

• 若客户端与MongoDB Service不在同一命名空间，需使用完整域名格式：auth-mongo-srv.<mongo-namespace>.svc.cluster.local

2. 检查Service与Pod的关联 #

• 查看Service的selector配置，确认与MongoDB Pod的labels匹配：

  kubectl describe service auth-mongo-srv
  kubectl describe pod <mongo-pod-name>
  

• 检查Service的Endpoints，确认有活跃的Pod IP：

  kubectl get endpoints auth-mongo-srv
  

  如果Endpoints为空，说明selector不匹配，需调整Service的selector或Pod的labels。

3. 检查网络策略限制 #

• 确认当前命名空间没有网络策略阻止客户端访问MongoDB的27017端口：

  kubectl get networkpolicy
  

  若存在网络策略，需添加规则允许客户端所在命名空间访问MongoDB Service。

4. 优化连接配置 #

• 使用完整的Service域名进行连接：

  await mongoose.connect('mongodb://auth-mongo-srv.<namespace>.svc.cluster.local:27017/auth');
  

• 增加连接超时与重试参数：

  await mongoose.connect('mongodb://auth-mongo-srv:27017/auth', {
    connectTimeoutMS: 30000,
    retryWrites: true,
    w: 'majority'
  });
  

内容的提问来源于stack exchange，提问作者Nahid Hasan