etcd operator creates/configures/manages etcd clusters atop Kubernetes
Перейти к файлу
Hongchao Deng ee750d3c56 remove tpr example 2016-08-23 14:26:19 -07:00
design add resize doc 2016-08-17 13:50:21 -07:00
example remove tpr example 2016-08-23 14:26:19 -07:00
hack comments 2016-08-10 15:46:09 -07:00
.gitignore Initial commit 2016-08-04 16:54:59 -07:00
Dockerfile implement recovery of one etcd node 2016-08-09 17:45:33 -07:00
LICENSE Initial commit 2016-08-04 16:54:59 -07:00
README.md update readme 2016-08-23 14:26:04 -07:00
cluster.go refactor; create TPR on init 2016-08-23 14:24:21 -07:00
controller.go refactor; create TPR on init 2016-08-23 14:24:21 -07:00
main.go refactor; create TPR on init 2016-08-23 14:24:21 -07:00
member_set.go fix build 2016-08-21 20:33:02 -07:00
recovery.go etcd pods bootstrap could take > 5s 2016-08-23 13:31:23 -07:00
spec.go add spec 2016-08-21 18:32:53 -07:00
util.go refactor; create TPR on init 2016-08-23 14:24:21 -07:00

README.md

kube-etcd-controller

Managed etcd clusters on Kubernetes:

  • creation
  • destroy
  • resize
  • recovery
  • backup
  • rolling upgrade

Requirements

  • Kubernetes 1.4+
  • etcd 3.0+

Limitations

  • Backup only works for data in etcd3 storage, not etcd2 storage.

Deploy kube-etcd-controller

$ kubectl create -f example/etcd-controller.yaml
pod "kubeetcdctrl" created

kube-etcd-controller will create a TPR automatically.

$ kubectl get thirdpartyresources
NAME                      DESCRIPTION             VERSION(S)
etcd-cluster.coreos.com   Managed etcd clusters   v1

Create an etcd cluster

$ cat example/example-etcd-cluster.yaml
apiVersion: "coreos.com/v1"
kind: "EtcdCluster"
metadata:
  name: "example-etcd-cluster"
size: 3
$ kubectl create -f example/example-etcd-cluster.yaml
$ kubectl get pods
NAME                READY     STATUS    RESTARTS   AGE
etcd-cluster-0000   1/1       Running   0          11s
etcd-cluster-0001   1/1       Running   0          11s
etcd-cluster-0002   1/1       Running   0          11s
$ kubectl get services
NAME                CLUSTER-IP     EXTERNAL-IP   PORT(S)             AGE
etcd-cluster-0000   10.0.104.18    <none>        2380/TCP,2379/TCP   45s
etcd-cluster-0001   10.0.243.108   <none>        2380/TCP,2379/TCP   45s
etcd-cluster-0002   10.0.45.68     <none>        2380/TCP,2379/TCP   45s
kubernetes          10.0.0.1       <none>        443/TCP             8m
$ kubectl log etcd-cluster-0000
...
2016-08-05 00:33:32.453768 I | api: enabled capabilities for version 3.0
2016-08-05 00:33:32.454178 N | etcdmain: serving insecure client requests on 0.0.0.0:2379, this is strongly discouraged!

Destroy an existing etcd cluster

$ kubectl delete -f example/example-etcd-cluster.yaml
$ kubectl get pods
NAME                       READY     STATUS    RESTARTS   AGE

Try cluster recovery

Simulate a pod failure by simply delete it

$ kubectl delete pod etcd-cluster-0000

etcd controller will recover the failure by creating a new pod etcd-cluster-0003

$ kubectl get pods
NAME                       READY     STATUS    RESTARTS   AGE
etcd-cluster-0001   1/1       Running   0          5s
etcd-cluster-0002   1/1       Running   0          5s
etcd-cluster-0003   1/1       Running   0          5s