☰

Kubernetes Tiller pod 死亡问题

tag-icon • tag-icon linux docker aws kubernetes
Kubernetes Tiller pod 死亡问题

我的 kubernetes 运行在 AWS 之上，我使用 helm 进行模板化，问题是tiller pod 即使集群上没有太多负载，我每隔几个小时就会死机。我从日志中不知道什么

ubuntu@kops:~$ kubectl get pods -owide -n kube-system | grep tiller tiller-deploy-6b985bb7b-88ssk 0/1 CrashLoopBackOff 71 19h 100.96.4.3 ip-172-20-46-194.us-west-2.compute.internal ubuntu@kops:~$ ubuntu@kops:~$ ubuntu@kops:~$ kubectl describe pod tiller-deploy-6b985bb7b-88ssk Error from server (NotFound): pods "tiller-deploy-6b985bb7b-88ssk" not found ubuntu@kops:~$ ubuntu@kops:~$ ubuntu@kops:~$ ubuntu@kops:~$ ubuntu@kops:~$ kubectl logs tiller-deploy-6b985bb7b-88ssk Error from server (NotFound): pods "tiller-deploy-6b985bb7b-88ssk" not found ubuntu@kops:~$

其运行的 EC2 有大量可用内存，CPU 为 8 核负载... admin@ip-172-20-46-194:~$ free -h total used free shared buffers cached Mem: 31G 5.1G 26G 1.1M 1.5G 2.4G -/+ buffers/cache: 1.2G 30G Swap: 0B 0B 0B 

top - 03:08:29 up 19:20,  1 user,  load average: 79.51, 78.59, 77.98
Tasks: 176 total,   4 running, 172 sleeping,   0 stopped,   0 zombie
%Cpu(s): 28.2 us,  0.5 sy,  0.1 ni,  0.0 id,  0.0 wa,  0.0 hi,  0.0 si, 71.3 st
KiB Mem:  32950672 total,  5371456 used, 27579216 free,  1524240 buffers
KiB Swap:        0 total,        0 used,        0 free.  2561276 cached Mem

  PID USER      PR  NI    VIRT    RES    SHR S  %CPU %MEM     TIME+ COMMAND
12203 root      20   0  906520  16756   1492 S 235.0  0.1   1427:46 docker
23974 root      20   0  906520  16796   1492 S 221.1  0.1   1054:34 docker
24072 root      20   0  906520  16796   1492 S 132.0  0.1 669:17.68 docker
12318 root      20   0  906520  16792   1492 S 130.7  0.1 900:54.95 docker
17543 nobody    20   0  906520  16796   1492 S  29.1  0.1  84:01.18 docker
23865 nobody    20   0  906520  16796   1492 R  15.9  0.1  69:02.31 docker
12112 nobody    20   0  906520  16792   1492 S  14.6  0.1  91:27.51 docker
 3013 root      20   0 6753392 124600  50028 S   7.6  0.4  73:59.01 kubelet
 6378 nobody    20   0  683644 432120  29772 R   6.0  1.3  21:09.54 prometheus

答案1

为了让您更进一步，您需要添加--namespace kube-system到描述命令中：

kubectl --namespace kube-system describe pod tiller-deploy-6b985bb7b-88ssk

然后，您可以更有效地根本原因。如果您想尝试删除 pod 以暂时绕过它，则相同。

相关内容