Openshift部署Zookeeper和Kafka

部署Zookeeper

github网址

https://github.com/ericnie2015/zookeeper-k8s-openshift

1.在openshift目录中，首先构建images

oc create -f buildconfig.yaml

oc new-app zk-builder -p  IMAGE_STREAM_VERSION="3.4.13"

buildconfig.yaml中主要是启动github中的docker类型的构建，并把结果push到imagestream中

- kind: BuildConfig
  apiVersion: v1
  metadata:
    name: zk-builder
  spec:
    runPolicy: Serial
    triggers:
      - type: GitHub
        github:
          secret: ${GITHUB_HOOK_SECRET}
      - type: ConfigChange
    source:
      git:
        uri: ${GITHUB_REPOSITORY}
        ref: ${GITHUB_REF}
    strategy:
      type: Docker
    output:
      to:
        kind: ImageStreamTag
        name: "${IMAGE_STREAM_NAME}:${IMAGE_STREAM_VERSION}"

而Dockerfile更多的是一个从无到有的构建过程，为解决启动无法访问/opt/zookeeper/data目录，把user设置成root

FROM openjdk:-jre-alpine

MAINTAINER Enrique Garcia <engapa@gmail.com>

ARG ZOO_HOME=/opt/zookeeper
ARG ZOO_USER=zookeeper
ARG ZOO_GROUP=zookeeper
ARG ZOO_VERSION="3.4.13"

ENV ZOO_HOME=$ZOO_HOME \
    ZOO_VERSION=$ZOO_VERSION \
    ZOO_CONF_DIR=$ZOO_HOME/conf \
    ZOO_REPLICAS=

# Required packages
RUN set -ex; \
    apk add --update --no-cache \
      bash tar wget curl gnupg openssl ca-certificates

# Download zookeeper distribution under ZOO_HOME /zookeeper-3.4./
ADD zk_download.sh /tmp/

RUN set -ex; \
    mkdir -p /opt/zookeeper/bin; \
    mkdir -p /opt/zookeeper/conf; \
    chmod a+x /tmp/zk_download.sh;

RUN /tmp/zk_download.sh

RUN set -ex \
    rm -rf /tmp/zk_download.sh; \
    apk del wget gnupg

# Add custom scripts and configure user
ADD zk_env.sh zk_setup.sh zk_status.sh /opt/zookeeper/bin/

RUN set -ex; \
    chmod a+x $ZOO_HOME/bin/zk_*.sh; \
    addgroup $ZOO_GROUP; \
    addgroup sudo; \
    adduser -h $ZOO_HOME -g "Zookeeper user" -s /sbin/nologin -D -G $ZOO_GROUP -G sudo $ZOO_USER; \
    chown -R $ZOO_USER:$ZOO_GROUP $ZOO_HOME; \
    ln -s $ZOO_HOME/bin/zk_*.sh /usr/bin

USER root
#USER $ZOO_USER
WORKDIR $ZOO_HOME/bin/
# 

EXPOSE ${ZK_clientPort:-} ${ZOO_SERVER_PORT:-} ${ZOO_ELECTION_PORT:-}

ENTRYPOINT ["./zk_env.sh"]

#RUN echo "aaa" > /usr/alog

#CMD ["tail","-f","/usr/alog"]
CMD zk_setup.sh && ./zkServer.sh start-foreground

如果因为虚拟机网络问题无法访问外网，可以先在login到registry,然后直接在本地构建

ericdeMacBook-Pro:zookeeper-k8s-openshift ericnie$ docker build -t 172.30.1.1:/myproject/zookeeper:3.4. .
Sending build context to Docker daemon  .78kB
Step / : FROM openjdk:-jre-alpine
Trying to pull repository registry.access.redhat.com/openjdk ...
Trying to pull repository docker.io/library/openjdk ...
sha256:e3168174d367db9928bb70e33b4750457092e61815d577e368f53efb29fea48b: Pulling from docker.io/library/openjdk
4fe2ade4980c: Pull complete
6fc58a8d4ae4: Pull complete
d3e6d7e9702a: Pull complete
Digest: sha256:e3168174d367db9928bb70e33b4750457092e61815d577e368f53efb29fea48b
Status: Downloaded newer image for docker.io/openjdk:-jre-alpine
 ---> 0fe3f0d1ee48

docker images

然后推送到镜像库

ericdeMacBook-Pro:zookeeper-k8s-openshift ericnie$ docker push 172.30.1.1:/myproject/zookeeper:3.4.
The push refers to a repository [172.30.1.1:/myproject/zookeeper]
5fe222836c76: Pushed
55e1a1171f7a: Pushed
347a06ac9233: Pushed
03a33ce83585: Pushed
94058c4e233d: Pushed
984d85b76d76: Pushed
cd4b8e8a8238: Pushed
12c374f8270a: Pushed
0c3170905795: Pushed
df64d3292fd6: Pushed
3.4.: digest: sha256:87bf78acf297bc2144d77ce4465294fec519fd50a4c197a1663cc4304c8040c9 size: 

完成后在console上看到imagestream

2.基于模版部署

创建部署模版

oc create -f zk-persistent.yaml

ericdeMacBook-Pro:openshift ericnie$ cat zk-persistent.yaml
kind: Template
apiVersion: v1
metadata:
  name: zk-persistent
  annotations:
    openshift.io/display-name: Zookeeper (Persistent)
    description: Create a replicated Zookeeper server with persistent storage
    iconClass: icon-database
    tags: database,zookeeper
labels:
  template: zk-persistent
  component: zk
parameters:
- name: NAME
  value: zk-persistent
  required: true
- name: SOURCE_IMAGE
  description: Container image
  value: 172.30.1.1:/myproject/zookeeper
  required: true
- name: ZOO_VERSION
  description: Version
  value: "3.4.13"
  required: true
- name: ZOO_REPLICAS
  description: Number of nodes
  value: ""
  required: true
- name: VOLUME_DATA_CAPACITY
  description: Persistent volume capacity for zookeeper dataDir directory (e.g. 512Mi, 2Gi)
  value: 1Gi
  required: true
- name: VOLUME_DATALOG_CAPACITY
  description: Persistent volume capacity for zookeeper dataLogDir directory (e.g. 512Mi, 2Gi)
  value: 1Gi
  required: true
- name: ZOO_TICK_TIME
  description: The number of milliseconds of each tick
  value: ""
  required: true
- name: ZOO_INIT_LIMIT
  description: The number of ticks that the initial synchronization phase can take
  value: ""
  required: true
- name: ZOO_SYNC_LIMIT
  description: The number of ticks that can pass between sending a request and getting an acknowledgement
  value: ""
  required: true
- name: ZOO_CLIENT_PORT
  description: The port at which the clients will connect
  value: ""
  required: true
- name: ZOO_SERVER_PORT
  description: Server port
  value: ""
  required: true
- name: ZOO_ELECTION_PORT
  description: Election port
  value: ""
  required: true
- name: ZOO_MAX_CLIENT_CNXNS
  description: The maximum number of client connections
  value: ""
  required: true
- name: ZOO_SNAP_RETAIN_COUNT
  description: The number of snapshots to retain in dataDir
  value: ""
  required: true
- name: ZOO_PURGE_INTERVAL
  description: Purge task interval in hours. Set to  to disable auto purge feature
  value: ""
  required: true
- name: ZOO_HEAP_SIZE
  description: JVM heap size
  value: "-Xmx960M -Xms960M"
  required: true
- name: RESOURCE_MEMORY_REQ
  description: The memory resource request.
  value: "1Gi"
  required: true
- name: RESOURCE_MEMORY_LIMIT
  description: The limits for memory resource.
  value: "1Gi"
  required: true
- name: RESOURCE_CPU_REQ
  description: The CPU resource request.
  value: ""
  required: true
- name: RESOURCE_CPU_LIMIT
  description: The limits for CPU resource.
  value: ""
  required: true

objects:
- apiVersion: v1
  kind: Service
  metadata:
    name: ${NAME}
    labels:
      zk-name: ${NAME}
    annotations:
      service.alpha.kubernetes.io/tolerate-unready-endpoints: "true"
  spec:
    ports:
    - port: ${ZOO_CLIENT_PORT}
      name: client
    - port: ${ZOO_SERVER_PORT}
      name: server
    - port: ${ZOO_ELECTION_PORT}
      name: election
    clusterIP: None
    selector:
      zk-name: ${NAME}
- apiVersion: apps/v1beta1
  kind: StatefulSet
  metadata:
    name: ${NAME}
    labels:
      zk-name: ${NAME}
  spec:
    podManagementPolicy: "Parallel"
    serviceName: ${NAME}
    replicas: ${ZOO_REPLICAS}
    template:
      metadata:
        labels:
          zk-name: ${NAME}
          template: zk-persistent
          component: zk
        annotations:
          scheduler.alpha.kubernetes.io/affinity: >
            {
              "podAntiAffinity": {
                "requiredDuringSchedulingIgnoredDuringExecution": [{
                  "labelSelector": {
                    "matchExpressions": [{
                      "key": "zk-name",
                      "operator": "In",
                      "values": ["${NAME}"]
                    }]
                  },
                  "topologyKey": "kubernetes.io/hostname"
                }]
              }
            }
      spec:
        containers:
        - name: ${NAME}
          imagePullPolicy: IfNotPresent
          image: ${SOURCE_IMAGE}:${ZOO_VERSION}
          resources:
            requests:
              memory: ${RESOURCE_MEMORY_REQ}
              cpu: ${RESOURCE_CPU_REQ}
            limits:
              memory: ${RESOURCE_MEMORY_LIMIT}
              cpu: ${RESOURCE_CPU_LIMIT}
          ports:
          - containerPort: ${ZOO_CLIENT_PORT}
            name: client
          - containerPort: ${ZOO_SERVER_PORT}
            name: server
          - containerPort: ${ZOO_ELECTION_PORT}
            name: election
          env:
          - name : SETUP_DEBUG
            value: "true"
          - name : ZOO_REPLICAS
            value: ${ZOO_REPLICAS}
          - name : ZK_HEAP_SIZE
            value: ${ZOO_HEAP_SIZE}
          - name : ZK_tickTime
            value: ${ZOO_TICK_TIME}
          - name : ZK_initLimit
            value: ${ZOO_INIT_LIMIT}
          - name : ZK_syncLimit
            value: ${ZOO_SYNC_LIMIT}
          - name : ZK_maxClientCnxns
            value: ${ZOO_MAX_CLIENT_CNXNS}
          - name : ZK_autopurge_snapRetainCount
            value: ${ZOO_SNAP_RETAIN_COUNT}
          - name : ZK_autopurge_purgeInterval
            value: ${ZOO_PURGE_INTERVAL}
          - name : ZK_clientPort
            value: ${ZOO_CLIENT_PORT}
          - name : ZOO_SERVER_PORT
            value: ${ZOO_SERVER_PORT}
          - name : ZOO_ELECTION_PORT
            value: ${ZOO_ELECTION_PORT}
          - name : JAVA_ZK_JVMFLAG
            value: "\"${ZOO_HEAP_SIZE}\""
          readinessProbe:
            exec:
              command:
              - zk_status.sh
            initialDelaySeconds:
            timeoutSeconds:
          livenessProbe:
            exec:
              command:
              - zk_status.sh
            initialDelaySeconds:
            timeoutSeconds:
          volumeMounts:
          - name: datadir
            mountPath: /opt/zookeeper/data
          - name: datalogdir
            mountPath: /opt/zookeeper/data-log
    volumeClaimTemplates:
    - metadata:
        name: datadir
        annotations:
          volume.alpha.kubernetes.io/storage-class: anything
      spec:
        accessModes: [ "ReadWriteOnce" ]
        resources:
          requests:
            storage: ${VOLUME_DATA_CAPACITY}
    - metadata:
        name: datalogdir
        annotations:
          volume.alpha.kubernetes.io/storage-class: anything
      spec:
        accessModes: [ "ReadWriteOnce" ]
        resources:
          requests:
            storage: ${VOLUME_DATALOG_CAPACITY}

oc new-app zk-persistent -p NAME=myzk
--> Deploying template "test/zk-persistent" to project test

     Zookeeper (Persistent)
     ---------
     Create a replicated Zookeeper server with persistent storage

     * With parameters:
        * NAME=myzk
        * SOURCE_IMAGE=bbvalabs/zookeeper
        * ZOO_VERSION=3.4.
        * ZOO_REPLICAS=
        * VOLUME_DATA_CAPACITY=1Gi
        * VOLUME_DATALOG_CAPACITY=1Gi
        * ZOO_TICK_TIME=
        * ZOO_INIT_LIMIT=
        * ZOO_SYNC_LIMIT=
        * ZOO_CLIENT_PORT=
        * ZOO_SERVER_PORT=
        * ZOO_ELECTION_PORT=
        * ZOO_MAX_CLIENT_CNXNS=
        * ZOO_SNAP_RETAIN_COUNT=
        * ZOO_PURGE_INTERVAL=
        * ZOO_HEAP_SIZE=-Xmx960M -Xms960M
        * RESOURCE_MEMORY_REQ=1Gi
        * RESOURCE_MEMORY_LIMIT=1Gi
        * RESOURCE_CPU_REQ=
        * RESOURCE_CPU_LIMIT=

--> Creating resources ...
    service "myzk" created
    statefulset "myzk" created
--> Success
    Run 'oc status' to view your app.

$ oc get all,pvc,statefulset -l zk-name=myzk
NAME       CLUSTER-IP   EXTERNAL-IP   PORT(S)                      AGE
svc/myzk   None         <none>        /TCP,/TCP,/TCP   11m

NAME                DESIRED   CURRENT   AGE
statefulsets/myzk                     11m

NAME        READY     STATUS    RESTARTS   AGE
po/myzk-   /       Running             2m
po/myzk-   /       Running             1m
po/myzk-   /       Running             46s

NAME                    STATUS    VOLUME                                     CAPACITY   ACCESSMODES   AGE
pvc/datadir-myzk-      Bound     pvc-a654d055-6dfa-11e7-abe1-42010a840002   1Gi        RWO           11m
pvc/datadir-myzk-      Bound     pvc-a6601148-6dfa-11e7-abe1-42010a840002   1Gi        RWO           11m
pvc/datadir-myzk-      Bound     pvc-a667fa41-6dfa-11e7-abe1-42010a840002   1Gi        RWO           11m
pvc/datalogdir-myzk-   Bound     pvc-a657ff77-6dfa-11e7-abe1-42010a840002   1Gi        RWO           11m
pvc/datalogdir-myzk-   Bound     pvc-a664407a-6dfa-11e7-abe1-42010a840002   1Gi        RWO           11m
pvc/datalogdir-myzk-   Bound     pvc-a66b85f7-6dfa-11e7-abe1-42010a840002   1Gi        RWO           11m

NAME                DESIRED   CURRENT   AGE
statefulsets/myzk                     11m

如果在cdk或者minishift上，因为只有一个节点，所以只会启动一个myzk

ericdeMacBook-Pro:openshift ericnie$ oc get pods
NAME      READY     STATUS    RESTARTS   AGE
myzk-    /       Running             1m
myzk-    /       Pending             1m
myzk-    /       Pending             1m

进行验证

ericdeMacBook-Pro:openshift ericnie$ for i in   ; do oc  exec myzk-$i -- hostname; done
myzk-
myzk-
myzk-

ericdeMacBook-Pro:openshift ericnie$ for i in   ; do echo "myid myzk-$i";oc exec myzk-$i -- cat /opt/zookeeper/data/myid;  done
myid myzk-

myid myzk-

myid myzk-

ericdeMacBook-Pro:openshift ericnie$ for i in   ; do oc exec myzk-$i -- hostname -f; done
myzk-.myzk.myproject.svc.cluster.local
myzk-.myzk.myproject.svc.cluster.local
myzk-.myzk.myproject.svc.cluster.local

3.删除实例

oc delete all,statefulset,pvc -l zk-name=myzk

中间遇到的问题

• pod启动以后处于crash状态，看日志是zk_status.sh没找到，后来调了半天的Dockerfile,发现部署模版调用的是zookeeper:3.4.13下载的版本，并非本地版本，所以强制修改为172.30.1.1:5000/myproject/zookeeper
• pod无法启动，报错没有访问/opt/zookeeper/data目录的权限，去掉SecurityContext的Runas语句后，以root启动避免

部署kafka

过程和zookeeper类似

1.clone代码

ericdeMacBook-Pro:minishift ericnie$ git clone https://github.com/ericnie2015/kafka-k8s-openshift.git
Cloning into 'kafka-k8s-openshift'...
remote: Enumerating objects: , done.
remote: Total  (delta ), reused  (delta ), pack-reused
Receiving objects: % (/), 102.01 KiB | 24.00 KiB/s, done.
Resolving deltas: % (/), done.

2.本地构建并push镜像仓库

ericdeMacBook-Pro:kafka-k8s-openshift ericnie$ docker build -t 172.30.1.1:/myproject/kafka:2.12-2.0. .
Sending build context to Docker daemon  .53kB
Step / : FROM openjdk:-jre-alpine
 ---> 0fe3f0d1ee48
Step / : MAINTAINER Enrique Garcia <engapa@gmail.com>
 ---> Using cache
 ---> e51b1e313e0c
Step / : ARG KAFKA_HOME=/opt/kafka
 ---> Running in 0a464e9d1781
 ---> abadcf5d52d5
Removing intermediate container 0a464e9d1781
Step / : ARG KAFKA_USER=kafka
 ---> Running in b2e50be2d35b
 ---> e3f1455c4aca

ericdeMacBook-Pro:kafka-k8s-openshift ericnie$ docker push 172.30.1.1:/myproject/kafka:2.12-2.0.
The push refers to a repository [172.30.1.1:/myproject/kafka]
84cb97552ea5: Pushed
681963d6c624: Pushed
47afbbc52b62: Pushed
81d8600a6e97: Pushed
8457712c19b8: Pushed
6286fd332b87: Pushed
c2f9d211658b: Pushed
12c374f8270a: Mounted from myproject/zookeeper
0c3170905795: Mounted from myproject/zookeeper
df64d3292fd6: Mounted from myproject/zookeeper
2.12-2.0.: digest: sha256:9ed95c9c7682b49f76d4b5454a704db5ba9561127fe86fe6ca52bd673c279ee5 size: 

3.基于模版部署

ericdeMacBook-Pro:openshift ericnie$ cat kafka-persistent.yaml
kind: Template
apiVersion: v1
metadata:
  name: kafka-persistent
  annotations:
    openshift.io/display-name: Kafka (Persistent)
    description: Create a Kafka cluster, with persistent storage.
    iconClass: icon-database
    tags: messaging,kafka
labels:
  template: kafka-persistent
  component: kafka
parameters:
- name: NAME
  description: Name.
  required: true
  value: kafka
- name: KAFKA_VERSION
  description: Kafka Version (Scala and kafka version).
  required: true
  value: "2.12-2.0.0"
- name: SOURCE_IMAGE
  description: Container image source.
  value: 172.30.1.1:/myproject/kafka
  required: true
- name: REPLICAS
  description: Number of replicas.
  required: true
  value: ""
- name: KAFKA_HEAP_OPTS
  description: Kafka JVM Heap options. Consider value of params RESOURCE_MEMORY_REQ and RESOURCE_MEMORY_LIMIT.
  required: true
  value: "-Xmx1960M -Xms1960M"
- name: SERVER_NUM_PARTITIONS
  description: >
    The default number of log partitions per topic.
    More partitions allow greater
    parallelism for consumption, but this will also result in more files across
    the brokers.
  required: true
  value: ""
- name: SERVER_DELETE_TOPIC_ENABLE
  description: >
    Topic deletion enabled.
    Switch to enable topic deletion or not, default value is 'true'
  value: "true"
- name: SERVER_LOG_RETENTION_HOURS
  description: >
    Log retention hours.
    The minimum age of a log file to be eligible for deletion.
  value: ""
- name: SERVER_ZOOKEEPER_CONNECT
  description: >
    Zookeeper conection string, a list as URL with nodes separated by ','.
  value: "zk-persistent-0.zk-persistent:2181,zk-persistent-1.zk-persistent:2181,zk-persistent-2.zk-persistent:2181"
  required: true
- name: SERVER_ZOOKEEPER_CONNECT_TIMEOUT
  description: >
    The max time that the client waits to establish a connection to zookeeper (ms).
  value: ""
  required: true
- name: VOLUME_KAFKA_CAPACITY
  description: Kafka logs capacity.
  required: true
  value: "10Gi"
- name: RESOURCE_MEMORY_REQ
  description: The memory resource request.
  value: "2Gi"
- name: RESOURCE_MEMORY_LIMIT
  description: The limits for memory resource.
  value: "2Gi"
- name: RESOURCE_CPU_REQ
  description: The CPU resource request.
  value: ""
- name: RESOURCE_CPU_LIMIT
  description: The limits for CPU resource.
  value: ""
- name: LP_INITIAL_DELAY
  description: >
    LivenessProbe initial delay in seconds.
  value: ""

objects:
- apiVersion: v1
  kind: Service
  metadata:
    name: ${NAME}
    labels:
      app: ${NAME}
    annotations:
      service.alpha.kubernetes.io/tolerate-unready-endpoints: "true"
  spec:
    ports:
    - port:
      name: server
    - port:
      name: zkclient
    - port:
      name: zkserver
    - port:
      name: zkleader
    clusterIP: None
    selector:
      app: ${NAME}
- apiVersion: apps/v1beta1
  kind: StatefulSet
  metadata:
    name: ${NAME}
    labels:
      app: ${NAME}
  spec:
    podManagementPolicy: "Parallel"
    serviceName: ${NAME}
    replicas: ${REPLICAS}
    template:
      metadata:
        labels:
          app: ${NAME}
          template: kafka-persistent
          component: kafka
        annotations:
          # Use this annotation if you want allocate each pod on different node
          # Note the number of nodes must be upper than REPLICAS parameter.
          scheduler.alpha.kubernetes.io/affinity: >
            {
              "podAntiAffinity": {
                "requiredDuringSchedulingIgnoredDuringExecution": [{
                  "labelSelector": {
                    "matchExpressions": [{
                      "key": "app",
                      "operator": "In",
                      "values": ["${NAME}"]
                    }]
                  },
                  "topologyKey": "kubernetes.io/hostname"
                }]
              }
            }
      spec:
        containers:
        - name: ${NAME}
          imagePullPolicy: IfNotPresent
          image: ${SOURCE_IMAGE}:${KAFKA_VERSION}
          resources:
            requests:
              memory: ${RESOURCE_MEMORY_REQ}
              cpu: ${RESOURCE_CPU_REQ}
            limits:
              memory: ${RESOURCE_MEMORY_LIMIT}
              cpu: ${RESOURCE_CPU_LIMIT}
          ports:
          - containerPort:
            name: server
          - containerPort:
            name: zkclient
          - containerPort:
            name: zkserver
          - containerPort:
            name: zkleader
          env:
          - name : KAFKA_REPLICAS
            value: ${REPLICAS}
          - name:  KAFKA_ZK_LOCAL
            value: "false"
          - name:  KAFKA_HEAP_OPTS
            value: ${KAFKA_HEAP_OPTS}
          - name:  SERVER_num_partitions
            value: ${SERVER_NUM_PARTITIONS}
          - name:  SERVER_delete_topic_enable
            value: ${SERVER_DELETE_TOPIC_ENABLE}
          - name:  SERVER_log_retention_hours
            value: ${SERVER_LOG_RETENTION_HOURS}
          - name:  SERVER_zookeeper_connect
            value: ${SERVER_ZOOKEEPER_CONNECT}
          - name:  SERVER_log_dirs
            value: "/opt/kafka/data/logs"
          - name:  SERVER_zookeeper_connection_timeout_ms
            value: ${SERVER_ZOOKEEPER_CONNECT_TIMEOUT}
          livenessProbe:
            exec:
              command:
              - kafka_server_status.sh
            initialDelaySeconds: ${LP_INITIAL_DELAY}
            timeoutSeconds:
          volumeMounts:
          - name: kafka-data
            mountPath: /opt/kafka/data
    volumeClaimTemplates:
    - metadata:
        name: kafka-data
        annotations:
          volume.alpha.kubernetes.io/storage-class: anything
      spec:
        accessModes: [ "ReadWriteOnce" ]
        resources:
          requests:
            storage: ${VOLUME_KAFKA_CAPACITY}

修改模版的镜像库和zookeeper地址，我只有一个所以只修改成一个了

ericdeMacBook-Pro:openshift ericnie$ oc create -f kafka-persistent.yaml
template "kafka-persistent" created
ericdeMacBook-Pro:openshift ericnie$ oc new-app kafka-persistent
--> Deploying template "myproject/kafka-persistent" to project myproject

     Kafka (Persistent)
     ---------
     Create a Kafka cluster, with persistent storage.

     * With parameters:
        * NAME=kafka
        * KAFKA_VERSION=2.12-2.0.
        * SOURCE_IMAGE=172.30.1.1:/myproject/kafka
        * REPLICAS=
        * KAFKA_HEAP_OPTS=-Xmx1960M -Xms1960M
        * SERVER_NUM_PARTITIONS=
        * SERVER_DELETE_TOPIC_ENABLE=true
        * SERVER_LOG_RETENTION_HOURS=
        * SERVER_ZOOKEEPER_CONNECT=myzk-.myzk.myproject.svc.cluster.local:,myzk-.myzk.myproject.svc.cluster.local:,myzk-.myzk.myproject.svc.cluster.local:
        * SERVER_ZOOKEEPER_CONNECT_TIMEOUT=
        * VOLUME_KAFKA_CAPACITY=10Gi
        * RESOURCE_MEMORY_REQ=.2Gi
        * RESOURCE_MEMORY_LIMIT=2Gi
        * RESOURCE_CPU_REQ=0.2
        * RESOURCE_CPU_LIMIT=
        * LP_INITIAL_DELAY=

--> Creating resources ...
    service "kafka" created
    statefulset "kafka" created
--> Success
    Application is not exposed. You can expose services to the outside world by executing one or more of the commands below:
     'oc expose svc/kafka'
    Run 'oc status' to view your app.

因为内存不够，无法跑起来。

4.删除

ericdeMacBook-Pro:openshift ericnie$ oc delete all,statefulset,pvc -l app=kafka
statefulset "kafka" deleted
service "kafka" deleted
persistentvolumeclaim "kafka-data-kafka-0" deleted
persistentvolumeclaim "kafka-data-kafka-1" deleted
persistentvolumeclaim "kafka-data-kafka-2" deleted