Question

我已根据https://github.com/CiscoCloud/kubernetes-ansible在debian上部署了我的群集。一切正常，除了一些插件，我怀疑名称解析。

我是这样定义的rc

apiVersion: v1
kind: ReplicationController
metadata:
  name: monitoring-influx-grafana-v1
  namespace: kube-system
  labels: 
    k8s-app: influxGrafana
    version: v1
    kubernetes.io/cluster-service: "true"
spec: 
  replicas: 1
  selector: 
    k8s-app: influxGrafana
    version: v1
  template: 
    metadata: 
      labels: 
        k8s-app: influxGrafana
        version: v1
        kubernetes.io/cluster-service: "true"
    spec: 
      containers: 
        - image: gcr.io/google_containers/heapster_influxdb:v0.3
          name: influxdb
          resources:
            limits:
              cpu: 100m
              memory: 200Mi
          ports: 
            - containerPort: 8083
              hostPort: 8083
            - containerPort: 8086
              hostPort: 8086
          volumeMounts:
          - name: influxdb-persistent-storage
            mountPath: /data
        - image: gcr.io/google_containers/heapster_grafana:v0.7
          name: grafana
          resources:
            limits:
              cpu: 100m
              memory: 100Mi
          env: 
            - name: INFLUXDB_EXTERNAL_URL
              value: /api/v1/proxy/namespaces/kube-system/services/monitoring-influxdb:api/db/
            - name: INFLUXDB_HOST
              value: monitoring-influxdb
            - name: INFLUXDB_PORT
              value: "8086"
      volumes:
      - name: influxdb-persistent-storage
        emptyDir: {}

因此容器连接到名为＆＃34; monitoring-influxdb＆＃34;的主机。实际上，当我使用＆＃34; busy-box＆＃34;时，dns可以很好地解析这个名称。荚：

kubectl exec busybox --namespace=kube-system -- nslookup monitoring-influxdb
Server:    10.233.0.10
Address 1: 10.233.0.10

名称：monitoring-influxdb 地址1：10.233.83.114

但是在monitoring-influx-grafana-v1 pod中，名称未解析：

kubectl exec monitoring-influx-grafana-v1-obu8h --namespace=kube-system -- ping -c 1 monitoring-influxdb
PING monitoring-influxdb.kube-system.svc.osm.local (10.233.83.114) 56(84) bytes of data.

--- monitoring-influxdb.kube-system.svc.osm.local ping statistics ---
1 packets transmitted, 0 received, 100% packet loss, time 0ms

注意＆＃34; kubernetes＆＃34;服务已解决

kubectl exec monitoring-influx-grafana-v1-obu8h --namespace=kube-system -- ping -c 1 kubernetes
PING kubernetes.default.svc.osm.local (10.233.0.1) 56(84) bytes of data.
64 bytes from 10.233.0.1: icmp_seq=1 ttl=63 time=0.441 ms

--- kubernetes.default.svc.osm.local ping statistics ---
1 packets transmitted, 1 received, 0% packet loss, time 0ms
```

服务在etcd

中定义如下

{'apiVersion': 'v1',
 'kind': 'Endpoints',
 'metadata': {'creationTimestamp': '2015-08-31T16:11:47Z',
              'labels': {'kubernetes.io/cluster-service': 'true',
                         'kubernetes.io/name': 'InfluxDB'},
              'name': 'monitoring-influxdb',
              'namespace': 'kube-system',
              'resourceVersion': '219',
              'selfLink': '/api/v1/namespaces/kube-system/endpoints/monitoring-influxdb',
              'uid': 'fab1bd0d-4ffa-11e5-bad3-52540011ed64'},
 'subsets': [{'addresses': [{'ip': '10.233.3.5',
                             'targetRef': {'kind': 'Pod',
                                           'name': 'monitoring-influx-grafana-v1-obu8h',
                                           'namespace': 'kube-system',
                                           'resourceVersion': '391',
                                           'uid': 'fa6fb734-4ffa-11e5-bad3-52540011ed64'}}],
              'ports': [{'name': 'http', 'port': 8083, 'protocol': 'TCP'},
                        {'name': 'api', 'port': 8086, 'protocol': 'TCP'}]}]}

我认为这个问题也会导致其他pod的错误

kubectl logs kibana-logging-v1-ifztn --namespace=kube-system
...
{"@timestamp":"2015-08-31T16:13:29.082Z","level":"info","message":"Unable to connect to elasticsearch at http://elasticsearch-logging:9200. Retrying in 2.5 seconds.","node_env":"production"}
...

kubectl logs monitoring-heapster-v8-9cy06 --namespace=kube-system
...
E0831 16:12:27.367871       1 driver.go:326] Database creation failed: Post http://monitoring-influxdb:8086/db?u=root&p=root: dial tcp: lookup monitoring-influxdb: no such host. Retrying after 30 seconds
...

此致 Smana

Answer 1

Smana， 1）就像你再次猜到的那样，它不是DNS解析问题 PING monitoring-influxdb.kube-system.svc.osm.local（10.233.83.114）56（84）字节的数据。 ping命令表示正在ping IP 10.233.83.144

2）群集ips不支持ICMP。目前，仅支持tcp和udp。因此ping失败。

无法在命名空间kube-system上解析服务名称

1 个答案: