kubernetes-一些常见的异常

1. kuber-apiserver: Getting a bunch of 404 when Initializers option is activated

# systemctl status kube-apiserver

Dec 19 12:43:30 rmxk8sctl01 kube-apiserver[7012]: I1219 12:43:30.665129 7012 wrap.go:42] GET /apis/admissionregistration.k8s.io/v1alpha1/initializerconfigurations: (167.967µs) 404 [[kube-apiserver/v1.8.5 (linux/amd64) kubernetes/cce11c6] 127.0.0.1:45026]

在kube-apiserver 增加参数：

--runtime-config=api/all=true,admissionregistration.k8s.io/v1alpha1=true

# kubectl api-versions | grep admissionregistration
admissionregistration.k8s.io/v1alpha1
admissionregistration.k8s.io/v1beta1

参考： https://github.com/kelseyhightower/kubernetes-the-hard-way/issues/276

2. etcdserver: Fix txn request 'took too long' warnings to use loggable request stringer

升级etcd后(3.1.7->3.3.7)：
systemctl status etcd -l 查看服务状态，有以下错误，但是并不影响kubernetes的正常运行：

6月 26 11:22:19 etcd-02 etcd[22147]: failed to send out heartbeat on time (exceeded the 100ms timeout for 274.333787ms)
6月 26 11:22:19 etcd-02 etcd[22147]: server is likely overloaded
6月 26 11:22:19 etcd-02 etcd[22147]: failed to send out heartbeat on time (exceeded the 100ms timeout for 274.400312ms)
6月 26 11:22:19 etcd-02 etcd[22147]: server is likely overloaded
6月 26 11:23:28 etcd-02 etcd[22147]: read-only range request "key:\"/registry/mutatingwebhookconfigurations/\" range_end:\"/registry/mutatingwebhookconfigurations0\" " with result "range_response_count:0 size:7" took too long (131.675812ms) to execute
6月 26 11:23:28 etcd-02 etcd[22147]: read-only range request "key:\"/registry/validatingwebhookconfigurations/\" range_end:\"/registry/validatingwebhookconfigurations0\" " with result "range_response_count:0 size:7" took too long (132.999236ms) to execute
6月 26 11:23:28 etcd-02 etcd[22147]: read-only range request "key:\"/registry/validatingwebhookconfigurations/\" range_end:\"/registry/validatingwebhookconfigurations0\" " with result "range_response_count:0 size:7" took too long (131.950418ms) to execute
6月 26 11:24:28 etcd-02 etcd[22147]: read-only range request "key:\"/registry/mutatingwebhookconfigurations/\" range_end:\"/registry/mutatingwebhookconfigurations0\" " with result "range_response_count:0 size:7" took too long (156.340818ms) to execute
6月 26 11:24:34 etcd-02 etcd[22147]: store.index: compact 3990430
6月 26 11:24:34 etcd-02 etcd[22147]: finished scheduled compaction at 3990430 (took 665.068µs)

解决参考(等下一个版本，升级到3.3.8也有这个错误提示)：https://github.com/coreos/etcd/pull/9835

3. istio-sidecar-injector.istio-system.svc:443 Client.Timeout

升级istio到 0.8.0后，一直不能使用自动嵌入sidecar.

master1 kube-controller-manager: I0405 21:40:29.377013 1538 event.go:218] Event(v1.ObjectReference{Kind:"ReplicaSet", Namespace:"default", Name:"sleep-6bc9d848fc", UID:"322f0a8a-38d5-11e8-aad2-005056846055", APIVersion:"extensions", ResourceVersion:"2633", FieldPath:""}): type: 'Warning' reason: 'FailedCreate' Error creating: Internal error occurred: failed calling admission webhook "sidecar-injector.istio.io": Post https://istio-sidecar-injector.istio-system.svc:443/inject: net/http: request canceled while waiting for connection (Client.Timeout exceeded while awaiting headers)

继续跟进此案例：https://github.com/istio/old_issues_repo/issues/271