溫馨提示×
您好，登錄后才能下訂單哦！
點擊 登錄注冊 即表示同意《億速云用戶服務條款》
您好，登錄后才能下訂單哦！
kubernetes中涉及很多概念，包含云生態社區中各類技術，學習成本比較高，k8s中通常以編寫yaml文件完成資源的部署，對于較多入門的人來說是個較高的門坎，本文以命令行的形式代理大家快速入門，俯瞰kubernetes核心概念，快速入門。
kubernetes是一個開源的容器引擎管理平臺，實現容器化應用的自動化部署，任務調度，彈性伸縮，負載均衡等功能，cluster是由master和node兩種角色組成，其中master負責管理集群，master節點由kube-apiserver，kube-controller-manager，kube-scheduler，etcd角色組成，node節點運行實際的應用，由Container Runtime，kubelet和kube-proxy組成，其中Container Runtime可能是Docker，rke，containerd，node節點可由物理機或者虛擬機組成。
1、查看master組件角色
[root@node-1 ~]# kubectl get componentstatuses
NAME STATUS MESSAGE ERROR
scheduler Healthy ok
controller-manager Healthy ok
etcd-0 Healthy {"health":"true"} 2、 查看node節點列表
[root@node-1 ~]# kubectl get nodes
NAME STATUS ROLES AGE VERSION
node-1 Ready master 26h v1.14.1
node-2 Ready <none> 26h v1.14.1
node-3 Ready <none> 26h v1.14.13、查看node節點詳情
[root@node-1 ~]# kubectl describe node node-3
Name: node-3
Roles: <none>
Labels: beta.kubernetes.io/arch=amd64。#標簽和Annotations
beta.kubernetes.io/os=linux
kubernetes.io/arch=amd64
kubernetes.io/hostname=node-3
kubernetes.io/os=linux
Annotations: flannel.alpha.coreos.com/backend-data: {"VtepMAC":"22:f8:75:bb:da:4e"}
flannel.alpha.coreos.com/backend-type: vxlan
flannel.alpha.coreos.com/kube-subnet-manager: true
flannel.alpha.coreos.com/public-ip: 10.254.100.103
kubeadm.alpha.kubernetes.io/cri-socket: /var/run/dockershim.sock
node.alpha.kubernetes.io/ttl: 0
volumes.kubernetes.io/controller-managed-attach-detach: true
CreationTimestamp: Sat, 10 Aug 2019 17:50:00 +0800
Taints: <none>
Unschedulable: false。#是否禁用調度，cordon命令控制的標識位。
Conditions: #資源調度能力，MemoryPressure內存是否有壓力（即內存不足）
#DiskPressure磁盤壓力
#PIDPressure磁盤壓力
#Ready，是否就緒，表明節點是否處于正常工作狀態，表示資源充足+相關進程狀態正常
Type Status LastHeartbeatTime LastTransitionTime Reason Message
---- ------ ----------------- ------------------ ------ -------
MemoryPressure False Sun, 11 Aug 2019 20:32:07 +0800 Sat, 10 Aug 2019 17:50:00 +0800 KubeletHasSufficientMemory kubelet has sufficient memory available
DiskPressure False Sun, 11 Aug 2019 20:32:07 +0800 Sat, 10 Aug 2019 17:50:00 +0800 KubeletHasNoDiskPressure kubelet has no disk pressure
PIDPressure False Sun, 11 Aug 2019 20:32:07 +0800 Sat, 10 Aug 2019 17:50:00 +0800 KubeletHasSufficientPID kubelet has sufficient PID available
Ready True Sun, 11 Aug 2019 20:32:07 +0800 Sat, 10 Aug 2019 18:04:20 +0800 KubeletReady kubelet is posting ready status
Addresses: #地址和主機名
InternalIP: 10.254.100.103
Hostname: node-3
Capacity: #容器的資源容量
cpu: 2
ephemeral-storage: 51473868Ki
hugepages-2Mi: 0
memory: 3880524Ki
pods: 110
Allocatable: #已分配資源情況
cpu: 2
ephemeral-storage: 47438316671
hugepages-2Mi: 0
memory: 3778124Ki
pods: 110
System Info: #系統信息，如內核版本，操作系統版本，cpu架構，node節點軟件版本
Machine ID: 0ea734564f9a4e2881b866b82d679dfc
System UUID: D98ECAB1-2D9E-41CC-9A5E-51A44DC5BB97
Boot ID: 6ec81f5b-cb05-4322-b47a-a8e046d9bf79
Kernel Version: 3.10.0-957.el7.x86_64
OS Image: CentOS Linux 7 (Core)
Operating System: linux
Architecture: amd64
Container Runtime Version: docker://18.3.1 . #Container Runtime為docker，版本為18.3.1
Kubelet Version: v1.14.1 #kubelet版本
Kube-Proxy Version: v1.14.1 #kube-proxy版本
PodCIDR: 10.244.2.0/24 #pod使用的網絡
Non-terminated Pods: (4 in total)。 #下面是每個pod資源占用情況
Namespace Name CPU Requests CPU Limits Memory Requests Memory Limits AGE
--------- ---- ------------ ---------- --------------- ------------- ---
kube-system coredns-fb8b8dccf-hrqm8 100m (5%) 0 (0%) 70Mi (1%) 170Mi (4%) 26h
kube-system coredns-fb8b8dccf-qwwks 100m (5%) 0 (0%) 70Mi (1%) 170Mi (4%) 26h
kube-system kube-flannel-ds-amd64-zzm2g 100m (5%) 100m (5%) 50Mi (1%) 50Mi (1%) 26h
kube-system kube-proxy-x8zqh 0 (0%) 0 (0%) 0 (0%) 0 (0%) 26h
Allocated resources: #已分配資源情況
(Total limits may be over 100 percent, i.e., overcommitted.)
Resource Requests Limits
-------- -------- ------
cpu 300m (15%) 100m (5%)
memory 190Mi (5%) 390Mi (10%)
ephemeral-storage 0 (0%) 0 (0%)
Events: <none>kubernetes是容器編排引擎，其負責容器的調度，管理和容器的運行，但kubernetes調度最小單位并非是container，而是pod，pod中可包含多個container，通常集群中不會直接運行pod，而是通過各種控制器如Deployments，ReplicaSets，DaemonSets的方式運行，為啥？因為控制器能夠保證pod狀態的一致性，正如官方所描述的一樣“make sure the current state match to the desire state”，確保當前狀態和預期的一致，簡單來說就是pod異常了，控制器會在其他節點重建，確保集群當前運行的pod和預期設定的一致。
kubernetes中pod是實際運行的載體，pod依附于node中，node可能會出現故障，kubernetes的控制器如replicasets會在其他node上重新拉起一個pod，新的pod會分配一個新的IP；再者，應用部署時會包含多個副本replicas，如同個應用deployments部署了3個pod副本，pod相當于后端的Real Server，如何實現這三個應用訪問呢？對于這種情況，我們一般會在Real Server前面加一個負載均衡Load Balancer，service就是pod的負載均衡調度器，service將動態的pod抽象為一個服務，應用程序直接訪問service即可，service會自動將請求轉發到后端的pod。負責service轉發規則有兩種機制：iptables和ipvs，iptables通過設置DNAT等規則實現負載均衡，ipvs通過ipvsadm設置轉發規。
根據服務不同的訪問方式，service分為如下幾種類型：ClusterIP，NodePort，LoadBalancer和_ExternalName，可通過type設置。
ExternalName，通過服務名字暴露服務名，當前可由ingress實現，將外部的請求以域名轉發的形式轉發到集群，需要依附具體的外部實現，如nginx，traefik,各大云計算廠商實現接入細節。
pod是動態變化的，ip地址可能會變化（如node故障），副本數可能會變化，如應用擴展scale up，應用鎖容scale down等，service如何識別到pod的動態變化呢？答案是labels，通過labels自動會過濾出某個應用的Endpoints，當pod變化時會自動更新Endpoints，不同的應用會有由不同的label組成。labels相關可以參考下https://kubernetes.io/docs/concepts/overview/working-with-objects/labels/
我們開始部署一個應用即deployments，kubernetes中包含各種workload如無狀態話的Deployments，有狀態化的StatefulSets，守護進程的DaemonSets，美中workload對應不同的應用場景，我們先以Deployments為例入門，其他workload均以此類似，一般而言，在kubernetes中部署應用均以yaml文件方式部署，對于初學者而言，編寫yaml文件太冗長，不適合初學，我們先kubectl命令行方式實現API的接入。
1、部署nginx應用，部署三個副本
[root@node-1 ~]# kubectl run nginx-app-demo --image=nginx:1.7.9 --port=80 --replicas=3
kubectl run --generator=deployment/apps.v1 is DEPRECATED and will be removed in a future version. Use kubectl run --generator=run-pod/v1 or kubectl create instead.
deployment.apps/nginx-app-demo created2、查看應用列表,可以看到當前pod的狀態均已正常，Ready是當前狀態，AVAILABLE是目標狀態
[root@node-1 ~]# kubectl get deployments
NAME READY UP-TO-DATE AVAILABLE AGE
nginx-app-demo 3/3 3 3 72s3、查看應用的詳細信息,如下我們可以知道Deployments是通過ReplicaSets控制副本數的，由Replicaset控制pod數
[root@node-1 ~]# kubectl describe deployments nginx-app-demo
Name: nginx-app-demo #應用名稱
Namespace: default #命名空間
CreationTimestamp: Sun, 11 Aug 2019 21:52:32 +0800
Labels: run=nginx-app-demo #labels，很重要，后續service通過labels實現訪問
Annotations: deployment.kubernetes.io/revision: 1 #滾動升級版本號
Selector: run=nginx-app-demo #labels的選擇器selector
Replicas: 3 desired | 3 updated | 3 total | 3 available | 0 unavailable #副本控制器
StrategyType: RollingUpdate #升級策略為RollingUpdate
MinReadySeconds: 0
RollingUpdateStrategy: 25% max unavailable, 25% max surge #RollingUpdate升級策略，即最大不超過25%的pod
Pod Template: #容器應用模版，包含鏡像，port，存儲等
Labels: run=nginx-app-demo
Containers:
nginx-app-demo:
Image: nginx:1.7.9
Port: 80/TCP
Host Port: 0/TCP
Environment: <none>
Mounts: <none>
Volumes: <none>
Conditions: #當前狀態
Type Status Reason
---- ------ ------
Available True MinimumReplicasAvailable
Progressing True NewReplicaSetAvailable
OldReplicaSets: <none>
NewReplicaSet: nginx-app-demo-7bdfd97dcd (3/3 replicas created) #ReplicaSets控制器名稱
Events: #運行事件
Type Reason Age From Message
---- ------ ---- ---- -------
Normal ScalingReplicaSet 3m24s deployment-controller Scaled up replica set nginx-app-demo-7bdfd97dcd to 34、查看replicasets情況，通過查看可知replicasets副本控制器生成了三個pod
1. 查看replicasets列表
[root@node-1 ~]# kubectl get replicasets
NAME DESIRED CURRENT READY AGE
nginx-app-demo-7bdfd97dcd 3 3 3 9m9s
2. 查看replicasets詳情
[root@node-1 ~]# kubectl describe replicasets nginx-app-demo-7bdfd97dcd
Name: nginx-app-demo-7bdfd97dcd
Namespace: default
Selector: pod-template-hash=7bdfd97dcd,run=nginx-app-demo
Labels: pod-template-hash=7bdfd97dcd #labels，增加了一個hash的label識別replicasets
run=nginx-app-demo
Annotations: deployment.kubernetes.io/desired-replicas: 3 #滾動升級的信息，副本樹，最大數，應用版本
deployment.kubernetes.io/max-replicas: 4
deployment.kubernetes.io/revision: 1
Controlled By: Deployment/nginx-app-demo #副本的父控制，為nginx-app-demo這個Deployments
Replicas: 3 current / 3 desired
Pods Status: 3 Running / 0 Waiting / 0 Succeeded / 0 Failed
Pod Template: #容器模板，繼承于deployments
Labels: pod-template-hash=7bdfd97dcd
run=nginx-app-demo
Containers:
nginx-app-demo:
Image: nginx:1.7.9
Port: 80/TCP
Host Port: 0/TCP
Environment: <none>
Mounts: <none>
Volumes: <none>
Events: #事件日志，生成了三個不同的pod
Type Reason Age From Message
---- ------ ---- ---- -------
Normal SuccessfulCreate 9m25s replicaset-controller Created pod: nginx-app-demo-7bdfd97dcd-hsrft
Normal SuccessfulCreate 9m25s replicaset-controller Created pod: nginx-app-demo-7bdfd97dcd-qtbzd
Normal SuccessfulCreate 9m25s replicaset-controller Created pod: nginx-app-demo-7bdfd97dcd-7t72x5、查看pod的情況,實際應用部署的載體，pod中部署了一個nginx的容器并分配了一個ip，可通過該ip直接訪問應用
1. 查看pod的列表，和replicasets生成的名稱一致
[root@node-1 ~]# kubectl get pods
NAME READY STATUS RESTARTS AGE
nginx-app-demo-7bdfd97dcd-7t72x 1/1 Running 0 13m
nginx-app-demo-7bdfd97dcd-hsrft 1/1 Running 0 13m
nginx-app-demo-7bdfd97dcd-qtbzd 1/1 Running 0 13m
查看pod的詳情
[root@node-1 ~]# kubectl describe pods nginx-app-demo-7bdfd97dcd-7t72x
Name: nginx-app-demo-7bdfd97dcd-7t72x
Namespace: default
Priority: 0
PriorityClassName: <none>
Node: node-3/10.254.100.103
Start Time: Sun, 11 Aug 2019 21:52:32 +0800
Labels: pod-template-hash=7bdfd97dcd #labels名稱
run=nginx-app-demo
Annotations: <none>
Status: Running
IP: 10.244.2.4 #pod的ip地址
Controlled By: ReplicaSet/nginx-app-demo-7bdfd97dcd #副本控制器為replicasets
Containers: #容器的信息，包括容器id，鏡像，丟按扣，狀態，環境變量等信息
nginx-app-demo:
Container ID: docker://5a0e5560583c5929e9768487cef43b045af4c6d3b7b927d9daf181cb28867766
Image: nginx:1.7.9
Image ID: docker-pullable://nginx@sha256:e3456c851a152494c3e4ff5fcc26f240206abac0c9d794affb40e0714846c451
Port: 80/TCP
Host Port: 0/TCP
State: Running
Started: Sun, 11 Aug 2019 21:52:40 +0800
Ready: True
Restart Count: 0
Environment: <none>
Mounts:
/var/run/secrets/kubernetes.io/serviceaccount from default-token-txhkc (ro)
Conditions: #容器的狀態條件
Type Status
Initialized True
Ready True
ContainersReady True
PodScheduled True
Volumes: #容器卷
default-token-txhkc:
Type: Secret (a volume populated by a Secret)
SecretName: default-token-txhkc
Optional: false
QoS Class: BestEffort #QOS類型
Node-Selectors: <none> #污點類型
Tolerations: node.kubernetes.io/not-ready:NoExecute for 300s
node.kubernetes.io/unreachable:NoExecute for 300s
Events: #事件狀態，拉鏡像，啟動容器
Type Reason Age From Message
---- ------ ---- ---- -------
Normal Scheduled 14m default-scheduler Successfully assigned default/nginx-app-demo-7bdfd97dcd-7t72x to node-3
Normal Pulling 14m kubelet, node-3 Pulling image "nginx:1.7.9"
Normal Pulled 14m kubelet, node-3 Successfully pulled image "nginx:1.7.9"
Normal Created 14m kubelet, node-3 Created container nginx-app-demo
Normal Started 14m kubelet, node-3 Started container nginx-app-demokubernetes為每個pod都分配了一個ip地址，可通過該地址直接訪問應用，相當于訪問RS，但一個應用是一個整體，由多個副本數組成，需要依賴于service來實現應用的負載均衡，service我們探討ClusterIP和NodePort的訪問方式。
1、設置pod的內容，為了方便區分，我們將三個pod的nginx站點內容設置為不同，以觀察負載均衡的效果
查看pod列表
[root@node-1 ~]# kubectl get pods
NAME READY STATUS RESTARTS AGE
nginx-app-demo-7bdfd97dcd-7t72x 1/1 Running 0 28m
nginx-app-demo-7bdfd97dcd-hsrft 1/1 Running 0 28m
nginx-app-demo-7bdfd97dcd-qtbzd 1/1 Running 0 28m
進入pod容器中
[root@node-1 ~]# kubectl exec -it nginx-app-demo-7bdfd97dcd-7t72x /bin/bash
設置站點內容
[root@nginx-app-demo-7bdfd97dcd-7t72x:/# echo "web1" >/usr/share/nginx/html/index.html
以此類推設置另外兩個pod的內容為web2和web3
[root@nginx-app-demo-7bdfd97dcd-hsrft:/# echo web2 >/usr/share/nginx/html/index.html
[root@nginx-app-demo-7bdfd97dcd-qtbzd:/# echo web3 >/usr/share/nginx/html/index.html2、獲取pod的ip地址，如何快速獲取pod的ip地址呢，可以通過-o wide參數顯示更多的內容,會包含pod所屬node和ip
[root@node-1 ~]# kubectl get pods -o wide
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
nginx-app-demo-7bdfd97dcd-7t72x 1/1 Running 0 34m 10.244.2.4 node-3 <none> <none>
nginx-app-demo-7bdfd97dcd-hsrft 1/1 Running 0 34m 10.244.1.2 node-2 <none> <none>
nginx-app-demo-7bdfd97dcd-qtbzd 1/1 Running 0 34m 10.244.1.3 node-2 <none> <none>3、訪問pod的ip，查看站點內容,不同的pod站點內容和上述步驟設置一致。
[root@node-1 ~]# curl http://10.244.2.4
web1
[root@node-1 ~]# curl http://10.244.1.2
web2
[root@node-1 ~]# curl http://10.244.1.3
web3通過pod的ip直接訪問應用，對于單個pod的應用可以實現，對于多個副本replicas的應用則不符合要求，需要通過service來實現負載均衡，service需要設置不同的type，默認為ClusterIP即集群內部訪問，如下通過expose子命令將服務暴露到service。
1、暴露service,其中port表示代理監聽端口，target-port代表是容器的端口，type設置的是service的類型
[root@node-1 ~]# kubectl expose deployment nginx-app-demo --name nginx-service-demo \
--port=80 \
--protocol=TCP \
--target-port=80 \
--type ClusterIP
service/nginx-service-demo exposed2、查看service的詳情，可以看到service通過labels選擇器selector自動將pod的ip生成endpoints
查看service列表，顯示有兩個，kubernetes為默認集群創建的service
[root@node-1 ~]# kubectl get services
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
kubernetes ClusterIP 10.96.0.1 <none> 443/TCP 29h
nginx-service-demo ClusterIP 10.102.1.1 <none> 80/TCP 2m54s
查看service詳情，可以看到Labels的Seletor和前面Deployments設置一致，Endpoints將pod組成一個列表
[root@node-1 ~]# kubectl describe services nginx-service-demo
Name: nginx-service-demo #名稱
Namespace: default #命名空間
Labels: run=nginx-app-demo #標簽名稱
Annotations: <none>
Selector: run=nginx-app-demo #標簽選擇器
Type: ClusterIP #service類型為ClusterIP
IP: 10.102.1.1 #服務的ip，即vip，集群內部會自動分配一個
Port: <unset> 80/TCP #服務端口，即ClusterIP對外訪問的端口
TargetPort: 80/TCP #容器端口
Endpoints: 10.244.1.2:80,10.244.1.3:80,10.244.2.4:80 #訪問地址列表
Session Affinity: None #負載均衡調度算法
Events: <none>3、訪問service的地址，可以訪問的內容可知，service自動實現了pods的負載均衡，調度策略為輪詢，為何？因為service默認的調度策略Session Affinity為None，即是輪訓，可以設置為ClientIP，實現會話保持，相同客戶端IP的請求會調度到相同的pod上。
[root@node-1 ~]# curl http://10.102.1.1
web3
[root@node-1 ~]# curl http://10.102.1.1
web1
[root@node-1 ~]# curl http://10.102.1.1
web2
[root@node-1 ~]# curl http://10.102.1.14、ClusterIP原理深入剖析，service后端實現有兩種機制：iptables和ipvs，環境安裝采用iptables，iptables通過nat的鏈生成訪問規則，KUBE-SVC-R5Y5DZHD7Q6DDTFZ為入站DNAT轉發規則，KUBE-MARK-MASQ為出站轉發
[root@node-1 ~]# iptables -t nat -L -n
Chain KUBE-SERVICES (2 references)
target prot opt source destination
KUBE-MARK-MASQ tcp -- !10.244.0.0/16 10.102.1.1 /* default/nginx-service-demo: cluster IP */ tcp dpt:80
KUBE-SVC-R5Y5DZHD7Q6DDTFZ tcp -- 0.0.0.0/0 10.102.1.1 /* default/nginx-service-demo: cluster IP */ tcp dpt:80
出站：KUBE-MARK-MASQ源地址段不是10.244.0.0/16訪問10.102.1.1的目標端口80時，將請求轉發給KUBE-MARK-MASQ鏈
入站：KUBE-SVC-R5Y5DZHD7Q6DDTFZ任意原地址訪問目標10.102.1.1的目標端口80時將請求轉發給KUBE-SVC-R5Y5DZHD7Q6DDTFZ鏈5、查看入站請求規則，入站請求規則將會映射到不同的鏈，不同鏈將會轉發到不同pod的ip上。
1. 查看入站規則KUBE-SVC-R5Y5DZHD7Q6DDTFZ，請求將轉發至三條鏈
[root@node-1 ~]# iptables -t nat -L KUBE-SVC-R5Y5DZHD7Q6DDTFZ -n
Chain KUBE-SVC-R5Y5DZHD7Q6DDTFZ (1 references)
target prot opt source destination
KUBE-SEP-DSWLUQNR4UPH24AX all -- 0.0.0.0/0 0.0.0.0/0 statistic mode random probability 0.33332999982
KUBE-SEP-56SLMGHHOILJT36K all -- 0.0.0.0/0 0.0.0.0/0 statistic mode random probability 0.50000000000
KUBE-SEP-K6G4Z74HQYF6X7SI all -- 0.0.0.0/0 0.0.0.0/0
2. 查看實際轉發的三條鏈的規則,實際映射到不同的pod的ip地址上
[root@node-1 ~]# iptables -t nat -L KUBE-SEP-DSWLUQNR4UPH24AX -n
Chain KUBE-SEP-DSWLUQNR4UPH24AX (1 references)
target prot opt source destination
KUBE-MARK-MASQ all -- 10.244.1.2 0.0.0.0/0
DNAT tcp -- 0.0.0.0/0 0.0.0.0/0 tcp to:10.244.1.2:80
[root@node-1 ~]# iptables -t nat -L KUBE-SEP-56SLMGHHOILJT36K -n
Chain KUBE-SEP-56SLMGHHOILJT36K (1 references)
target prot opt source destination
KUBE-MARK-MASQ all -- 10.244.1.3 0.0.0.0/0
DNAT tcp -- 0.0.0.0/0 0.0.0.0/0 tcp to:10.244.1.3:80
[root@node-1 ~]# iptables -t nat -L KUBE-SEP-K6G4Z74HQYF6X7SI -n
Chain KUBE-SEP-K6G4Z74HQYF6X7SI (1 references)
target prot opt source destination
KUBE-MARK-MASQ all -- 10.244.2.4 0.0.0.0/0
DNAT tcp -- 0.0.0.0/0 0.0.0.0/0 tcp to:10.244.2.4:80 Service通過ClusterIP只能提供集群內部的應用訪問，外部無法直接訪問應用，如果需要外部訪問有如下幾種方式：NodePort，LoadBalancer和Ingress，其中LoadBalancer需要由云服務提供商實現，Ingress需要安裝單獨的Ingress Controller，日常測試可以通過NodePort的方式實現，NodePort可以將node的某個端口暴露給外部網絡訪問。
1、修改type的類型由ClusterIP修改為NodePort類型（或者重新創建，指定type的類型為NodePort）
1. 通過patch修改type的類型
[root@node-1 ~]# kubectl patch services nginx-service-demo -p '{"spec":{"type": "NodePort"}}'
service/nginx-service-demo patched
2. 確認yaml文件配置，分配了一個NodePort端口，即每個node上都會監聽該端口
[root@node-1 ~]# kubectl get services nginx-service-demo -o yaml
apiVersion: v1
kind: Service
metadata:
creationTimestamp: "2019-08-11T14:35:59Z"
labels:
run: nginx-app-demo
name: nginx-service-demo
namespace: default
resourceVersion: "157676"
selfLink: /api/v1/namespaces/default/services/nginx-service-demo
uid: 55e29b78-bc45-11e9-b073-525400490421
spec:
clusterIP: 10.102.1.1
externalTrafficPolicy: Cluster
ports:
- nodePort: 32416 #自動分配了一個NodePort端口
port: 80
protocol: TCP
targetPort: 80
selector:
run: nginx-app-demo
sessionAffinity: None
type: NodePort #類型修改為NodePort
status:
loadBalancer: {}
3. 查看service列表，可以知道service的type已經修改為NodePort,同時還保留ClusterIP的訪問IP
[root@node-1 ~]# kubectl get services
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
kubernetes ClusterIP 10.96.0.1 <none> 443/TCP 30h
nginx-service-demo NodePort 10.102.1.1 <none> 80:32416/TCP 68m2、通過NodePort訪問應用程序,每個node的地址相當于vip，可以實現相同的負載均衡效果，同時CluserIP功能依可用
1. NodePort的負載均衡
[root@node-1 ~]# curl http://node-1:32416
web1
[root@node-1 ~]# curl http://node-2:32416
web1
[root@node-1 ~]# curl http://node-3:32416
web1
[root@node-1 ~]# curl http://node-3:32416
web3
[root@node-1 ~]# curl http://node-3:32416
web2
2. ClusterIP的負載均衡
[root@node-1 ~]# curl http://10.102.1.1
web2
[root@node-1 ~]# curl http://10.102.1.1
web1
[root@node-1 ~]# curl http://10.102.1.1
web1
[root@node-1 ~]# curl http://10.102.1.1
web33、NodePort轉發原理，每個node上通過kube-proxy監聽NodePort的端口，由后端的iptables實現端口的轉發
1. NodePort監聽端口
[root@node-1 ~]# netstat -antupl |grep 32416
tcp6 0 0 :::32416 :::* LISTEN 32052/kube-proxy
2. 查看nat表的轉發規則，有兩條規則KUBE-MARK-MASQ出口和KUBE-SVC-R5Y5DZHD7Q6DDTFZ入站方向。
Chain KUBE-NODEPORTS (1 references)
target prot opt source destination
KUBE-MARK-MASQ tcp -- 0.0.0.0/0 0.0.0.0/0 /* default/nginx-service-demo: */ tcp dpt:32416
KUBE-SVC-R5Y5DZHD7Q6DDTFZ tcp -- 0.0.0.0/0 0.0.0.0/0 /* default/nginx-service-demo: */ tcp dpt:32416
3. 查看入站的請求規則鏈KUBE-SVC-R5Y5DZHD7Q6DDTFZ
[root@node-1 ~]# iptables -t nat -L KUBE-SVC-R5Y5DZHD7Q6DDTFZ -n
Chain KUBE-SVC-R5Y5DZHD7Q6DDTFZ (2 references)
target prot opt source destination
KUBE-SEP-DSWLUQNR4UPH24AX all -- 0.0.0.0/0 0.0.0.0/0 statistic mode random probability 0.33332999982
KUBE-SEP-56SLMGHHOILJT36K all -- 0.0.0.0/0 0.0.0.0/0 statistic mode random probability 0.50000000000
KUBE-SEP-K6G4Z74HQYF6X7SI all -- 0.0.0.0/0 0.0.0.0/0
4. 繼續查看轉發鏈，包含有DNAT轉發和KUBE-MARK-MASQ和出站返回的規則
[root@node-1 ~]# iptables -t nat -L KUBE-SEP-DSWLUQNR4UPH24AX -n
Chain KUBE-SEP-DSWLUQNR4UPH24AX (1 references)
target prot opt source destination
KUBE-MARK-MASQ all -- 10.244.1.2 0.0.0.0/0
DNAT tcp -- 0.0.0.0/0 0.0.0.0/0 tcp to:10.244.1.2:80
[root@node-1 ~]# iptables -t nat -L KUBE-SEP-56SLMGHHOILJT36K -n
Chain KUBE-SEP-56SLMGHHOILJT36K (1 references)
target prot opt source destination
KUBE-MARK-MASQ all -- 10.244.1.3 0.0.0.0/0
DNAT tcp -- 0.0.0.0/0 0.0.0.0/0 tcp to:10.244.1.3:80
[root@node-1 ~]# iptables -t nat -L KUBE-SEP-K6G4Z74HQYF6X7SI -n
Chain KUBE-SEP-K6G4Z74HQYF6X7SI (1 references)
target prot opt source destination
KUBE-MARK-MASQ all -- 10.244.2.4 0.0.0.0/0
DNAT tcp -- 0.0.0.0/0 0.0.0.0/0 tcp to:10.244.2.4:80當應用程序的負載比較高無法滿足應用請求時，一般我們會通過擴展RS的數量來實現，在kubernetes中，擴展RS實際上通過擴展副本數replicas來實現，擴展RS非常便利，快速實現彈性伸縮。kubernets能提供兩種方式的伸縮能力：1. 手動伸縮能力scale up和scale down，2. 動態的彈性伸縮horizontalpodautoscalers,基于CPU的利用率實現自動的彈性伸縮，需要依賴與監控組件如metrics server，當前未實現，后續再做深入探討，本文以手動的scale的方式擴展應用的副本數。
1、手動擴展副本數
[root@node-1 ~]# kubectl scale --replicas=4 deployment nginx-app-demo
deployment.extensions/nginx-app-demo scaled2、查看副本擴展情況,deployments自動部署一個應用
[root@node-1 ~]# kubectl get deployments
NAME READY UP-TO-DATE AVAILABLE AGE
nginx-app-demo 4/4 4 4 133m3、此時service的情況會怎樣呢？查看service詳情,新擴展的pod會自動更新到service的endpoints中，自動服務發現
查看service詳情
[root@node-1 ~]# kubectl describe services nginx-service-demo
Name: nginx-service-demo
Namespace: default
Labels: run=nginx-app-demo
Annotations: <none>
Selector: run=nginx-app-demo
Type: NodePort
IP: 10.102.1.1
Port: <unset> 80/TCP
TargetPort: 80/TCP
NodePort: <unset> 32416/TCP
Endpoints: 10.244.1.2:80,10.244.1.3:80,10.244.2.4:80 + 1 more...#地址已自動加入
Session Affinity: None
External Traffic Policy: Cluster
Events: <none>
查看endpioints詳情
[root@node-1 ~]# kubectl describe endpoints nginx-service-demo
Name: nginx-service-demo
Namespace: default
Labels: run=nginx-app-demo
Annotations: endpoints.kubernetes.io/last-change-trigger-time: 2019-08-11T16:04:56Z
Subsets:
Addresses: 10.244.1.2,10.244.1.3,10.244.2.4,10.244.2.5
NotReadyAddresses: <none>
Ports:
Name Port Protocol
---- ---- --------
<unset> 80 TCP
Events: <none>4、測試，將新加入的pod站點內容設置為web4，參考前面的設置方法，測試service的ip，查看負載均衡效果
[root@node-1 ~]# curl http://10.102.1.1
web4
[root@node-1 ~]# curl http://10.102.1.1
web4
[root@node-1 ~]# curl http://10.102.1.1
web2
[root@node-1 ~]# curl http://10.102.1.1
web3
[root@node-1 ~]# curl http://10.102.1.1
web1
[root@node-1 ~]# curl http://10.102.1.1
web2
[root@node-1 ~]# curl http://10.102.1.1
web1由此可知，彈性伸縮會自動自動加入到service中實現服務自動發現和負載均衡，應用的擴展相比于傳統應用快速非常多。
在kubernetes中更新應用程序時可以將應用程序打包到鏡像中，然后更新應用程序的鏡像以實現升級。默認Deployments的升級策略為RollingUpdate，其每次會更新應用中的25%的pod，新建新的pod逐個替換，防止應用程序在升級過程中不可用。同時，如果應用程序升級過程中失敗，還可以通過回滾的方式將應用程序回滾到之前的狀態，回滾時通過replicasets的方式實現。
1、更換nginx的鏡像，將應用升級至最新版本,打開另外一個窗口使用kubectl get pods -w觀察升級過程
[root@node-1 ~]# kubectl set image deployments/nginx-app-demo nginx-app-demo=nginx:latest
deployment.extensions/nginx-app-demo image updated2、觀察升級過程，通過查看可知，升級過程中是通過新建+刪除的方式逐個替換pod的方式
[root@node-1 ~]# kubectl get pods -w
NAME READY STATUS RESTARTS AGE
nginx-app-demo-7bdfd97dcd-7t72x 1/1 Running 0 145m
nginx-app-demo-7bdfd97dcd-hsrft 1/1 Running 0 145m
nginx-app-demo-7bdfd97dcd-j6lgd 1/1 Running 0 12m
nginx-app-demo-7bdfd97dcd-qtbzd 1/1 Running 0 145m
nginx-app-demo-5cc8746f96-xsxz4 0/1 Pending 0 0s #新建一個pod
nginx-app-demo-5cc8746f96-xsxz4 0/1 Pending 0 0s
nginx-app-demo-7bdfd97dcd-j6lgd 1/1 Terminating 0 14m #刪除舊的pod，替換
nginx-app-demo-5cc8746f96-xsxz4 0/1 ContainerCreating 0 0s
nginx-app-demo-5cc8746f96-s49nv 0/1 Pending 0 0s #新建第二個pod
nginx-app-demo-5cc8746f96-s49nv 0/1 Pending 0 0s
nginx-app-demo-5cc8746f96-s49nv 0/1 ContainerCreating 0 0s
nginx-app-demo-7bdfd97dcd-j6lgd 0/1 Terminating 0 14m #更換第二個pod
nginx-app-demo-5cc8746f96-s49nv 1/1 Running 0 7s
nginx-app-demo-7bdfd97dcd-qtbzd 1/1 Terminating 0 146m
nginx-app-demo-5cc8746f96-txjqh 0/1 Pending 0 0s
nginx-app-demo-5cc8746f96-txjqh 0/1 Pending 0 0s
nginx-app-demo-5cc8746f96-txjqh 0/1 ContainerCreating 0 0s
nginx-app-demo-7bdfd97dcd-j6lgd 0/1 Terminating 0 14m
nginx-app-demo-7bdfd97dcd-j6lgd 0/1 Terminating 0 14m
nginx-app-demo-5cc8746f96-xsxz4 1/1 Running 0 9s
nginx-app-demo-5cc8746f96-txjqh 1/1 Running 0 1s
nginx-app-demo-7bdfd97dcd-hsrft 1/1 Terminating 0 146m
nginx-app-demo-7bdfd97dcd-qtbzd 0/1 Terminating 0 146m
nginx-app-demo-5cc8746f96-rcpmw 0/1 Pending 0 0s
nginx-app-demo-5cc8746f96-rcpmw 0/1 Pending 0 0s
nginx-app-demo-5cc8746f96-rcpmw 0/1 ContainerCreating 0 0s
nginx-app-demo-7bdfd97dcd-7t72x 1/1 Terminating 0 146m
nginx-app-demo-7bdfd97dcd-7t72x 0/1 Terminating 0 147m
nginx-app-demo-7bdfd97dcd-hsrft 0/1 Terminating 0 147m
nginx-app-demo-7bdfd97dcd-hsrft 0/1 Terminating 0 147m
nginx-app-demo-5cc8746f96-rcpmw 1/1 Running 0 2s
nginx-app-demo-7bdfd97dcd-7t72x 0/1 Terminating 0 147m
nginx-app-demo-7bdfd97dcd-7t72x 0/1 Terminating 0 147m
nginx-app-demo-7bdfd97dcd-hsrft 0/1 Terminating 0 147m
nginx-app-demo-7bdfd97dcd-hsrft 0/1 Terminating 0 147m
nginx-app-demo-7bdfd97dcd-qtbzd 0/1 Terminating 0 147m
nginx-app-demo-7bdfd97dcd-qtbzd 0/1 Terminating 0 147m3、再次查看deployments的詳情可知道，deployments已經更換了新的replicasets，原來的replicasets的版本為1，可用于回滾。
[root@node-1 ~]# kubectl describe deployments nginx-app-demo
Name: nginx-app-demo
Namespace: default
CreationTimestamp: Sun, 11 Aug 2019 21:52:32 +0800
Labels: run=nginx-app-demo
Annotations: deployment.kubernetes.io/revision: 2 #新的版本號，用于回滾
Selector: run=nginx-app-demo
Replicas: 4 desired | 4 updated | 4 total | 4 available | 0 unavailable
StrategyType: RollingUpdate
MinReadySeconds: 0
RollingUpdateStrategy: 25% max unavailable, 25% max surge
Pod Template:
Labels: run=nginx-app-demo
Containers:
nginx-app-demo:
Image: nginx:latest
Port: 80/TCP
Host Port: 0/TCP
Environment: <none>
Mounts: <none>
Volumes: <none>
Conditions:
Type Status Reason
---- ------ ------
Available True MinimumReplicasAvailable
Progressing True NewReplicaSetAvailable
OldReplicaSets: <none>
NewReplicaSet: nginx-app-demo-5cc8746f96 (4/4 replicas created) #新的replicaset，實際是替換新的replicasets
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal ScalingReplicaSet 19m deployment-controller Scaled up replica set nginx-app-demo-7bdfd97dcd to 4
Normal ScalingReplicaSet 4m51s deployment-controller Scaled up replica set nginx-app-demo-5cc8746f96 to 1
Normal ScalingReplicaSet 4m51s deployment-controller Scaled down replica set nginx-app-demo-7bdfd97dcd to 3
Normal ScalingReplicaSet 4m51s deployment-controller Scaled up replica set nginx-app-demo-5cc8746f96 to 2
Normal ScalingReplicaSet 4m43s deployment-controller Scaled down replica set nginx-app-demo-7bdfd97dcd to 2
Normal ScalingReplicaSet 4m43s deployment-controller Scaled up replica set nginx-app-demo-5cc8746f96 to 3
Normal ScalingReplicaSet 4m42s deployment-controller Scaled down replica set nginx-app-demo-7bdfd97dcd to 1
Normal ScalingReplicaSet 4m42s deployment-controller Scaled up replica set nginx-app-demo-5cc8746f96 to 4
Normal ScalingReplicaSet 4m42s deployment-controller Scaled down replica set nginx-app-demo-7bdfd97dcd to 04、查看滾動升級的版本，可以看到有兩個版本，分別對應的兩個不同的replicasets
[root@node-1 ~]# kubectl rollout history deployment nginx-app-demo
deployment.extensions/nginx-app-demo
REVISION CHANGE-CAUSE
1 <none>
2 <none>
查看replicasets列表,舊的包含pod為0
[root@node-1 ~]# kubectl get replicasets
NAME DESIRED CURRENT READY AGE
nginx-app-demo-5cc8746f96 4 4 4 9m2s
nginx-app-demo-7bdfd97dcd 0 0 0 155m5、測試應用的升級情況,發現nginx已經升級到最新nginx/1.17.2版本
[root@node-1 ~]# curl -I http://10.102.1.1
HTTP/1.1 200 OK
Server: nginx/1.17.2 #nginx版本信息
Date: Sun, 11 Aug 2019 16:30:03 GMT
Content-Type: text/html
Content-Length: 612
Last-Modified: Tue, 23 Jul 2019 11:45:37 GMT
Connection: keep-alive
ETag: "5d36f361-264"
Accept-Ranges: bytes6、回滾到舊的版本
[root@node-1 ~]# kubectl rollout undo deployment nginx-app-demo --to-revision=1
deployment.extensions/nginx-app-demo rolled back
再次測應用,已經回滾到舊版本。
[root@node-1 ~]# curl -I http://10.102.1.1
HTTP/1.1 200 OK
Server: nginx/1.7.9
Date: Sun, 11 Aug 2019 16:34:33 GMT
Content-Type: text/html
Content-Length: 612
Last-Modified: Tue, 23 Dec 2014 16:25:09 GMT
Connection: keep-alive
ETag: "54999765-264"
Accept-Ranges: bytes寫在最后：本文以命令行的方式實踐探索kubernetes中涉及的最重要的幾個概念：應用部署，負載均衡，彈性伸縮和滾動升級，并以命令行的形式實際操作，讀者可以參照文檔實現快速入門，后續會大部分以yaml文件的形式部署和kubernets交互。
參考文檔
基礎概念：https://kubernetes.io/docs/tutorials/kubernetes-basics/
部署應用：https://kubernetes.io/docs/tutorials/kubernetes-basics/deploy-app/deploy-intro/
訪問應用：https://kubernetes.io/docs/tutorials/kubernetes-basics/explore/explore-intro/
外部訪問：https://kubernetes.io/docs/tutorials/kubernetes-basics/expose/expose-intro/
訪問應用：https://kubernetes.io/docs/tutorials/kubernetes-basics/scale/scale-intro/
滾動升級：https://kubernetes.io/docs/tutorials/kubernetes-basics/update/update-intro/
當你的才華撐不起你的野心時，你就應該靜下心來學習
返回kubernetes系列教程目錄
免責聲明：本站發布的內容（圖片、視頻和文字）以原創、轉載和分享為主，文章觀點不代表本網站立場，如果涉及侵權請聯系站長郵箱：is@yisu.com進行舉報，并提供相關證據，一經查實，將立刻刪除涉嫌侵權內容。
