k8s 组件之coredns

coredns configmap默认配置

apiVersion: v1
data:
  Corefile: |
    .:53 {
        errors
        health {
          lameduck 5s
        }
        ready
        kubernetes cluster.local in-addr.arpa ip6.arpa {   # 内部域名解析配置
           pods insecure
          fallthrough in-addr.arpa ip6.arpa
        }
        prometheus :9153
        forward . /etc/resolv.conf {       # 对于外部域名采用node节点的配置文件
          max_concurrent 1000
        }
        cache 30
        loop
        reload
        loadbalance
    }

如何让集群pod解析外部域名(非集群内部svc域名)

1. 单个域名解析

apiVersion: v1
data:
  Corefile: |
    .:53 {
        errors
        health {
          lameduck 5s
        }
        ready
        kubernetes cluster.local in-addr.arpa ip6.arpa {
           pods insecure
          fallthrough in-addr.arpa ip6.arpa
        }
        prometheus :9153
        forward . /etc/resolv.conf {
          max_concurrent 1000
        }
        cache 30
        loop
        reload
        loadbalance
        hosts {
            129.226.170.70 www.tb.com   # 强制解析
           }
    }

2. 采用DNS服务器解析

[root@master-1 ~]# kubectl  get cm -n kube-system coredns -oyaml
apiVersion: v1
data:
  Corefile: |
    .:53 {
        errors
        health {
          lameduck 5s
        }
        ready
        kubernetes cluster.local in-addr.arpa ip6.arpa {  # 内部解析
           pods insecure
          fallthrough in-addr.arpa ip6.arpa
        }
        prometheus :9153
        cache 30
        loop
        reload
        loadbalance
        forward . 8.8.8.8  223.5.5.5 {   # 对应集群内部无法解析的，转发给dns服务器
           max_concurrent 1000
        }
    }
...

注意：如果有内部dns服务器，建议写公司内部DNS服务器地址。

3. 重启coredns控制器

rollout: 与资源的发布（rollout）相关的操作子命令，包括查看、暂停、恢复、重启等功能。

[root@master-1 ~]# kubectl  rollout restart deploy coredns -n kube-system
deployment.apps/coredns restarted

4. 监测pod状态

滚动更新策略保证始终有一部分 Pod 在服务中，所以不会完全中断解析服务。

coredns-64ccdcbccb-qtvk7                   0/1     Pending   0          0s
coredns-64ccdcbccb-qtvk7                   0/1     Pending   0          0s
coredns-6cdc74485f-cm9pm                   1/1     Terminating   0          37m
coredns-64ccdcbccb-qtvk7                   0/1     ContainerCreating   0          0s
coredns-64ccdcbccb-pr74x                   0/1     Pending             0          0s
coredns-64ccdcbccb-pr74x                   0/1     Pending             0          0s
coredns-64ccdcbccb-qtvk7                   0/1     ContainerCreating   0          2s
coredns-64ccdcbccb-qtvk7                   0/1     Running             0          3s
coredns-6cdc74485f-cm9pm                   1/1     Terminating         0          37m
coredns-6cdc74485f-cm9pm                   0/1     Terminating         0          37m
coredns-64ccdcbccb-qtvk7                   1/1     Running             0          7s
coredns-6cdc74485f-mk925                   1/1     Terminating         0          37m
coredns-6cdc74485f-cm9pm                   0/1     Terminating         0          37m
coredns-6cdc74485f-cm9pm                   0/1     Terminating         0          37m
coredns-64ccdcbccb-pr74x                   0/1     Pending             0          9s
coredns-64ccdcbccb-pr74x                   0/1     ContainerCreating   0          9s
coredns-64ccdcbccb-pr74x                   0/1     ContainerCreating   0          10s
coredns-64ccdcbccb-pr74x                   0/1     Running             0          11s
coredns-6cdc74485f-mk925                   1/1     Terminating         0          37m
coredns-6cdc74485f-mk925                   0/1     Terminating         0          37m
coredns-64ccdcbccb-pr74x                   1/1     Running             0          14s
coredns-6cdc74485f-mk925                   0/1     Terminating         0          37m
coredns-6cdc74485f-mk925                   0/1     Terminating         0          37m

5. 测试解析

[root@master-1 ~]# kubectl  exec busybox -- cat /etc/resolv.conf
nameserver 10.0.0.2
search default.svc.cluster.local svc.cluster.local cluster.local
options ndots:5

[root@master-1 ~]# kubectl  exec busybox -- ping www.jd.com
PING www.jd.com (58.220.65.19): 56 data bytes
64 bytes from 58.220.65.19: seq=0 ttl=52 time=30.665 ms
64 bytes from 58.220.65.19: seq=1 ttl=52 time=26.216 ms
64 bytes from 58.220.65.19: seq=2 ttl=52 time=25.405 ms
^C
[root@master-1 ~]# kubectl  exec busybox -- ping www.tb.com
PING www.tb.com (129.226.170.70): 56 data bytes
64 bytes from 129.226.170.70: seq=0 ttl=48 time=46.721 ms
64 bytes from 129.226.170.70: seq=1 ttl=48 time=65.689 ms
64 bytes from 129.226.170.70: seq=2 ttl=48 time=85.008 ms