一、自動伸縮
1、啟動
[root@k8s-master ~]# kubectl autoscale deployment nginx-deployment --max=8 --min=2 --cpu-percent=80 deployment "nginx-deployment" autoscaled
2、查看創建
[root@k8s-master ~]# kubectl get all NAME DESIRED CURRENT UP-TO-DATE AVAILABLE AGE deploy/nginx-deployment 2 2 2 2 13h NAME REFERENCE TARGET CURRENT MINPODS MAXPODS AGE hpa/nginx-deployment Deployment/nginx-deployment 80% 0% 2 8 17s NAME CLUSTER-IP EXTERNAL-IP PORT(S) AGE svc/kubernetes 10.254.0.1 <none> 443/TCP 2d svc/nginx 10.254.145.15 <nodes> 80:32000/TCP 1d NAME DESIRED CURRENT READY AGE rs/nginx-deployment-2950479891 0 0 0 13h rs/nginx-deployment-3113009173 2 2 2 13h NAME READY STATUS RESTARTS AGE po/nginx-deployment-3113009173-h5plc 1/1 Running 0 17s po/nginx-deployment-3113009173-vckhg 1/1 Running 1 13h
3、修改副本數為1
[root@k8s-master ~]# kubectl edit deployment nginx-deployment 修改為1 replicas: 1 deployment "nginx-deployment" edited [root@k8s-master ~]# kubectl get all NAME DESIRED CURRENT UP-TO-DATE AVAILABLE AGE deploy/nginx-deployment 1 1 1 1 13h NAME CLUSTER-IP EXTERNAL-IP PORT(S) AGE svc/kubernetes 10.254.0.1 <none> 443/TCP 2d svc/nginx 10.254.145.15 <nodes> 80:32000/TCP 1d NAME DESIRED CURRENT READY AGE rs/nginx-deployment-2950479891 0 0 0 13h rs/nginx-deployment-3113009173 1 1 1 13h NAME READY STATUS RESTARTS AGE po/nginx-deployment-3113009173-vckhg 1/1 Running 1 13h
二、hpa伸縮
1、編輯horizontalpodautoscaler 文件
[root@k8s-master ~]# kubectl get horizontalpodautoscaler
NAME REFERENCE TARGET CURRENT MINPODS MAXPODS AGE
nginx-deployment Deployment/nginx-deployment 80% 0% 2 8 1m
# Please edit the object below. Lines beginning with a '#' will be ignored,
# and an empty file will abort the edit. If an error occurs while saving this file will be
# reopened with the relevant failures.
#
apiVersion: autoscaling/v1
kind: HorizontalPodAutoscaler
metadata:
creationTimestamp: 2019-01-22T01:00:02Z
name: nginx-deployment
namespace: default
resourceVersion: "41194"
selfLink: /apis/autoscaling/v1/namespaces/default/horizontalpodautoscalers/nginx-deployment
uid: 0c897472-1de1-11e9-9773-000c292bd9e1
spec:
maxReplicas: 8
minReplicas: 2
scaleTargetRef:
apiVersion: extensions/v1beta1
kind: Deployment
name: nginx-deployment
targetCPUUtilizationPercentage: 80
status:
currentCPUUtilizationPercentage: 0
currentReplicas: 2
desiredReplicas: 2
lastScaleTime: 2019-01-22T01:00:02Z
2、修改副本為1
[root@k8s-master ~]# kubectl edit deployment nginx-deployment replicas: 1 deployment "nginx-deployment" edited [root@k8s-master ~]# kubectl get all NAME DESIRED CURRENT UP-TO-DATE AVAILABLE AGE deploy/nginx-deployment 2 2 2 2 13h NAME REFERENCE TARGET CURRENT MINPODS MAXPODS AGE hpa/nginx-deployment Deployment/nginx-deployment 80% 0% 2 8 6m NAME CLUSTER-IP EXTERNAL-IP PORT(S) AGE svc/kubernetes 10.254.0.1 <none> 443/TCP 2d svc/nginx 10.254.145.15 <nodes> 80:32000/TCP 1d NAME DESIRED CURRENT READY AGE rs/nginx-deployment-2950479891 0 0 0 13h rs/nginx-deployment-3113009173 2 2 2 13h NAME READY STATUS RESTARTS AGE po/nginx-deployment-3113009173-9hlq1 1/1 Running 0 2s po/nginx-deployment-3113009173-vckhg 1/1 Running 1 13h
明明修改為1，怎么還有2個？是因為hpa如下配置
spec: maxReplicas: 8 minReplicas: 2
3、修改副本數為5
[root@k8s-master ~]# kubectl edit hpa nginx-deployment 修改： spec: maxReplicas: 8 minReplicas: 5 horizontalpodautoscaler "nginx-deployment" edited [root@k8s-master ~]# kubectl get all NAME DESIRED CURRENT UP-TO-DATE AVAILABLE AGE deploy/nginx-deployment 5 5 5 2 13h NAME REFERENCE TARGET CURRENT MINPODS MAXPODS AGE hpa/nginx-deployment Deployment/nginx-deployment 80% 0% 5 8 8m NAME CLUSTER-IP EXTERNAL-IP PORT(S) AGE svc/kubernetes 10.254.0.1 <none> 443/TCP 2d svc/nginx 10.254.145.15 <nodes> 80:32000/TCP 1d NAME DESIRED CURRENT READY AGE rs/nginx-deployment-2950479891 0 0 0 13h rs/nginx-deployment-3113009173 5 5 2 13h NAME READY STATUS RESTARTS AGE po/nginx-deployment-3113009173-97l9c 0/1 ContainerCreating 0 2s po/nginx-deployment-3113009173-9hlq1 1/1 Running 0 2m po/nginx-deployment-3113009173-qq4h8 0/1 ContainerCreating 0 2s po/nginx-deployment-3113009173-sfp8z 0/1 ContainerCreating 0 2s po/nginx-deployment-3113009173-vckhg 1/1 Running 1 13h
看到自動伸縮的過程了吧！
三、HAP-deployment-rs-rc-pod調用原理
1、什么是hpa
Horizontal Pod Autoscaling可以根據CPU使用率或應用自定義metrics自動擴展Pod數量（支持replication controller、deployment和replica set）。
- 控制管理器每隔30s（可以通過–horizontal-pod-autoscaler-sync-period修改）查詢metrics的資源使用情況
- 支持三種metrics類型
- 預定義metrics（比如Pod的CPU）以利用率的方式計算
- 自定義的Pod metrics，以原始值（raw value）的方式計算
- 自定義的object metrics
- 支持兩種metrics查詢方式：Heapster和自定義的REST API
- 支持多metrics
客戶端;
通過kubectl創建一個horizontalPodAutoscaler對象，並存儲到etcd中
服務端：
api server:負責接受創建hpa對象，然后存入etcd
hpa controler和其他的controler類似，每30s同步一次，將已經創建的hpa進行一次管理（從heapster獲取監控數據，查看是否需要scale, controler的store中就保存着從始至終創建出來的hpa，當做一個緩存），watch hpa有變化也會運行。從heapster中獲取scale數據，和hpa對比，計算cup利用率等信息，然后重新調整scale。根據hpa.Spec.ScaleTargetRef.Kind（例如Deployment，然后deployment控制器在調整pod數量），調整其值，發送到apiserver存儲到etcd，然后更新hpa到etcd.
2、示例
# 創建pod和service $ kubectl run php-apache --image=gcr.io/google_containers/hpa-example --requests=cpu=200m --expose --port=80 service "php-apache" created deployment "php-apache" created # 創建autoscaler $ kubectl autoscale deployment php-apache --cpu-percent=50 --min=1 --max=10 deployment "php-apache" autoscaled $ kubectl get hpa NAME REFERENCE TARGET CURRENT MINPODS MAXPODS AGE php-apache Deployment/php-apache/scale 50% 0% 1 10 18s # 增加負載 $ kubectl run -i --tty load-generator --image=busybox /bin/sh Hit enter for command prompt $ while true; do wget -q -O- http://php-apache.default.svc.cluster.local; done # 過一會就可以看到負載升高了 $ kubectl get hpa NAME REFERENCE TARGET CURRENT MINPODS MAXPODS AGE php-apache Deployment/php-apache/scale 50% 305% 1 10 3m # autoscaler將這個deployment擴展為7個pod $ kubectl get deployment php-apache NAME DESIRED CURRENT UP-TO-DATE AVAILABLE AGE php-apache 7 7 7 7 19m # 刪除剛才創建的負載增加pod后會發現負載降低，並且pod數量也自動降回1個 $ kubectl get hpa NAME REFERENCE TARGET CURRENT MINPODS MAXPODS AGE php-apache Deployment/php-apache/scale 50% 0% 1 10 11m $ kubectl get deployment php-apache NAME DESIRED CURRENT UP-TO-DATE AVAILABLE AGE php-apache 1 1 1 1 27m
3、關系原理圖
4、小結
1、首先最底層的資源永遠都是pod
2、比pod高級一點的資源叫rc副本控制器
3、在pod上面有一個高級的rs
4、那誰來管理rs呢？是deployment
5、HPA自動管理deployment，deployment設置為1，HPA最低設置為3，deployment這就會被自動設計為3個