1.准备工作
安装minikube, 安装kubectl(不再赘述, 自行查找)
启动minikube
minikube start image-mirror-country='cn' --registry-mirror=https://registry.docker-cn.com --memory=4096 --image-repository=registry.cn-hangzhou.aliyuncs.com/google_containers新建k8s namespace, 用于逻辑隔离整套elk服务
kubectl create namespace elkspace
2.编写k8s yaml文件(###标记都是注释, copy请删除)
elasticsearch.yaml
- apiVersion: apps/v1 kind: Deployment ###启动es容器, 级别可以是Pod等,有不同定义的 metadata: name: es-single namespace: elkspace ###第一步已定义, 可以全部不使用, 走默认default namespace labels: app: es-single spec: replicas: 1 selector: matchLabels: app: es-single ###定义k:v格式标签, 用于Service通过标签查询并连接 template: metadata: name: es-single labels: app: es-single spec: containers: - image: elasticsearch:7.9.1 ###docker镜像, 注意ELK三者的版本号请统一 imagePullPolicy: IfNotPresent name: es resources: limits: cpu: 1 memory: 2Gi requests: cpu: 0.5 memory: 500Mi env: ###环境变量, 相当于docker run中的-e参数 - name: "discovery.type" ###单节点es value: "single-node" - name: ES_JAVA_OPTS value: "-Xms512m -Xmx2g" ports: ###设置容器暴露端口, 相当于设置Pod端口9200,9300 tcp可用 - containerPort: 9200 protocol: TCP - containerPort: 9300 protocol: TCP - apiVersion: v1 kind: Service ###启动服务, 用于对外暴露Pod信息, 因为k8s外无法直连Pod, 必须通过Service暴露出去端口 metadata: name: es-single-service namespace: elkspace spec: type: NodePort ###相当于Pod端口映射Node端口, 外部访问通过NodeIP+NodePort就可以连到服务 ports: - name: esport port: 9200 ###Service连接Pod端口 targetPort: 9200 ###上面暴露的es Pod端口 protocol: TCP nodePort: 30011 ###k8s外部访问端口, 自定义, 但是有范围 - name: esportlink port: 9300 targetPort: 9300 protocol: TCP nodePort: 30021 selector: ###很重要, 服务绑定到指定K:V形式的Pod, 上面配置好的 app: es-singlekibana.yaml
- apiVersion: apps/v1 kind: Deployment metadata: name: kb-single namespace: elkspace labels: app: kb-single spec: replicas: 1 selector: matchLabels: app: kb-single template: metadata: name: kb-single labels: app: kb-single spec: containers: - name: kb image: kibana:7.9.1 imagePullPolicy: IfNotPresent env: ###非常重要!!!服务之间ip动态变化, 配置连接最好用域名, Service域名规则是:service.namespace - name: ELASTICSEARCH_HOSTS ###非常重要!!!注意是ELASTICSEARCH_HOSTS还是ELASTICSEARCH_URL, 得和当前版本kibana中/usr/share/kibana/config/kibana.yml的参数名称一样, 可以docker启动镜像进去看看 value: "http://es-single-service.elkspace:9200" - name: XPACK_SECURITY_ENABLED value: "true" ports: - name: ui containerPort: 5601 protocol: TCP resources: limits: cpu: 1 memory: 2Gi requests: cpu: 0.5 memory: 500Mi - apiVersion: v1 kind: Service metadata: name: kb-single-service namespace: elkspace spec: type: NodePort ports: - port: 5601 protocol: TCP targetPort: 5601 nodePort: 30031 selector: app: kb-singlelogstash.yaml
- apiVersion: apps/v1 kind: Deployment metadata: name: log-single namespace: elkspace labels: app: log-single spec: replicas: 1 selector: matchLabels: app: log-single template: metadata: name: log-single labels: app: log-single spec: containers: - name: log image: logstash:7.9.1 imagePullPolicy: IfNotPresent ports: - containerPort: 5044 protocol: TCP volumeMounts: ###该容器需要挂载的地址,也就是logstash.conf,因为logstash日志接收和转发信息都在此文件配置 - name: log-config mountPath: /usr/share/logstash/pipeline env: ###非常重要!!!和上一步的kibana此处配置一样的效果,参数名称需要和logstash/config/中配置文件一样 - name: XPACK_MONITORING_ELASTICSEARCH_HOSTS value: "http://es-single-service.elkspace:9200" securityContext: privileged: true volumes: ###指定从哪里挂载目录 - name: log-config ###对应到volumeMounts中的name hostPath: ###从当前Node本地路径加载,直接写我自己电脑绝对路径不对,此时的Node是minikube容器,所以这个地址也是minikube的目录,所以首先得把电脑本地路径挂载到minikube路径, 然后这里才能取到配置文件等信息(下一步说明) path: /data/pipeline - apiVersion: v1 kind: Service metadata: name: lg-single-service namespace: elkspace spec: type: NodePort ports: - port: 5044 protocol: TCP targetPort: 5044 nodePort: 30041 selector: app: log-single本地随便找个地方新建logstash.conf文件, 就是上一步中hostPath挂载的logstash.conf配置文件, 文件信息如下, 注意es 服务的地址, 和上面配置的一样
input { tcp { port => 5044 mode => "server" } } output { elasticsearch { hosts => ["es-single-service.elkspace:9200"] index => "apidemo" } stdout{ codec => rubydebug } }然后本地路径挂载minikube指定地址, &表示后台操作
/Users/grahamliu/elk/pipeline 电脑本地logstash.conf所在目录
/data/pipeline minikube指定地址, 也就是第3步中配置的hostpath, minikube文件目录详细信息可自行查阅
minikube mount /Users/grahamliu/elk/pipeline:/data/pipeline &
3.启动上面的各项配置文件(自行查询)
4.外部如何打开k8s中的Kibana? 如何连接写入日志到k8s logstash?
咱们部署是是单Node服务, 所有外部访问都是通过NodeIP+NodePort访问
查询minikube node的命令, InternalIP: 192.168.xx.x
pro-2:~ grahamliu$ kubectl get node NAME STATUS ROLES AGE VERSION minikube Ready master 117d v1.17.3 pro-2:~ grahamliu$ kubectl describe node minikube Name: minikube Roles: master Labels: beta.kubernetes.io/arch=amd64 beta.kubernetes.io/os=linux kubernetes.io/arch=amd64 kubernetes.io/hostname=minikube kubernetes.io/os=linux node-role.kubernetes.io/master= Annotations: 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: Sun, 03 Jan 2021 11:57:08 +0800 Taints: <none> Unschedulable: false Conditions: Type Status LastHeartbeatTime LastTransitionTime Reason Message ---- ------ ----------------- ------------------ ------ ------- MemoryPressure False Fri, 30 Apr 2021 16:45:08 +0800 Sun, 03 Jan 2021 11:57:02 +0800 KubeletHasSufficientMemory kubelet has sufficient memory available DiskPressure False Fri, 30 Apr 2021 16:45:08 +0800 Sun, 03 Jan 2021 11:57:02 +0800 KubeletHasNoDiskPressure kubelet has no disk pressure PIDPressure False Fri, 30 Apr 2021 16:45:08 +0800 Sun, 03 Jan 2021 11:57:02 +0800 KubeletHasSufficientPID kubelet has sufficient PID available Ready True Fri, 30 Apr 2021 16:45:08 +0800 Sun, 03 Jan 2021 11:57:11 +0800 KubeletReady kubelet is posting ready status Addresses: InternalIP: 192.168.xx.x Hostname: minikube Capacity:此时只要用以上ip加上之前Service中暴露的端口就能成功访问各个服务了, 可以用命令查询所有端口信息
pro-2:~ grahamliu$ kubectl get svc -n elkspace NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE es-single-service NodePort 10.100.233.92 <none> 9200:30011/TCP,9300:30021/TCP 22h kb-single-service NodePort 10.104.129.185 <none> 5601:30031/TCP 3h22m lg-single-service NodePort 10.111.174.220 <none> 5044:30041/TCP 176m
5.整个流程都通了以后, 其实elk三个yaml可以整合到一个yaml一次性启动, elk.yaml编排如下
kind: List
apiVersion: v1
items:
- apiVersion: apps/v1
kind: Deployment
metadata:
name: es-single
namespace: elkspace
labels:
app: es-single
spec:
replicas: 1
selector:
matchLabels:
app: es-single
template:
metadata:
name: es-single
labels:
app: es-single
spec:
containers:
- image: elasticsearch:7.9.1
imagePullPolicy: IfNotPresent
name: es
resources:
limits:
cpu: 1
memory: 2Gi
requests:
cpu: 0.5
memory: 500Mi
env:
- name: "discovery.type"
value: "single-node"
- name: ES_JAVA_OPTS
value: "-Xms512m -Xmx2g"
ports:
- containerPort: 9200
protocol: TCP
- containerPort: 9300
protocol: TCP
- apiVersion: v1
kind: Service
metadata:
name: es-single-service
namespace: elkspace
spec:
type: NodePort
ports:
- name: esport
port: 9200
targetPort: 9200
protocol: TCP
nodePort: 30011
- name: esportlink
port: 9300
targetPort: 9300
protocol: TCP
nodePort: 30021
selector:
app: es-single
- apiVersion: apps/v1
kind: Deployment
metadata:
name: kb-single
namespace: elkspace
labels:
app: kb-single
spec:
replicas: 1
selector:
matchLabels:
app: kb-single
template:
metadata:
name: kb-single
labels:
app: kb-single
spec:
containers:
- name: kb
image: kibana:7.9.1
imagePullPolicy: IfNotPresent
env:
- name: ELASTICSEARCH_HOSTS
value: "http://es-single-service.elkspace:9200"
- name: XPACK_SECURITY_ENABLED
value: "true"
ports:
- name: ui
containerPort: 5601
protocol: TCP
resources:
limits:
cpu: 1
memory: 2Gi
requests:
cpu: 0.5
memory: 500Mi
- apiVersion: v1
kind: Service
metadata:
name: kb-single-service
namespace: elkspace
spec:
type: NodePort
ports:
- port: 5601
protocol: TCP
targetPort: 5601
nodePort: 30031
selector:
app: kb-single
- apiVersion: apps/v1
kind: Deployment
metadata:
name: log-single
namespace: elkspace
labels:
app: log-single
spec:
replicas: 1
selector:
matchLabels:
app: log-single
template:
metadata:
name: log-single
labels:
app: log-single
spec:
containers:
- name: log
image: logstash:7.9.1
imagePullPolicy: IfNotPresent
ports:
- containerPort: 5044
protocol: TCP
volumeMounts:
- name: log-config
mountPath: /usr/share/logstash/pipeline
env:
- name: XPACK_MONITORING_ELASTICSEARCH_HOSTS
value: "http://es-single-service.elkspace:9200"
securityContext:
privileged: true
volumes:
- name: log-config
hostPath:
path: /data/pipeline
- apiVersion: v1
kind: Service
metadata:
name: lg-single-service
namespace: elkspace
spec:
type: NodePort
ports:
- port: 5044
protocol: TCP
targetPort: 5044
nodePort: 30041
selector:
app: log-single