Kafka Kubernetes tutorial: How to Run HA Kafka on Google Kubernetes Engine

 

This post is part of our ongoing series on how to run Apache Kafka on Kubernetes.  We’ve published a number of articles about running Kafka on Kubernetes for specific platforms and for specific use cases.  If you are looking for a specific Kubernetes platform, check out these related articles.

Running HA Kafka on Amazon Elastic Container Service for Kubernetes (EKS)

Running HA Kafka on Azure Kubernetes Service (AKS)

Running HA Kafka on Red Hat OpenShift

Running HA Kafka on Amazon Elastic Container Service (ECS)

Running HA Kafka on IBM Kubernetes Service (IKS)

Running HA Kafka with Rancher Kubernetes Engine (RKE)

Running HA Kafka on IBM Cloud Private

And now, onto the post…

 

Google Kubernetes Engine (GKE) is a managed, production-ready environment for deploying containerized applications in the Google Cloud Platform. Launched in 2015, GKE is one of the first hosted container platforms which is built on the learnings from Google’s experience of running services like Gmail and YouTube in containers for over 12 years. GKE allows customers to quickly get up and running with Kubernetes by completely eliminating the need to install, manage, and operate Kubernetes clusters.

Portworx is a cloud native storage platform to run persistent workloads deployed on a variety of orchestration engines including Kubernetes. With Portworx, customers can manage the database of their choice on any infrastructure using any container scheduler. It provides a single data management layer for all stateful services, no matter where they run.

This Kafka Kubernetes tutorial is a walk-through of the steps involved in deploying and managing a highly available Kafka cluster on GKE as a Kubernetes StatefulSet. Kafka is a popular open source streaming platform that can be used with scenarios such as streaming clickstream data from web applications and sensor data from IoT devices. Portworx customer NIO, for example, uses Kafka to stream data off of self-driving cars. As a stateful application, Kafka itself needs a solution for persistence but so does one of its primary dependencies, ZooKeeper.

In summary, to run an HA Kafka cluster on GKE you need to:

  1. Install a GKE cluster by following instructions in the GCP docs
  2. Install a cloud native storage solution like Portworx as a daemon set on GKE
  3. Create a storage class defining your storage requirements like replication factor, snapshot policy, and performance profile
  4. Deploy ZooKeeper as a StatefulSet on Kubernetes
  5. Deploy Kafka as a StatefulSet on Kubernetes
  6. Test failover by killing or cordoning nodes in your cluster
  7. Perform backup and restore Kafka nodes

 

How to set up a GKE cluster

When launching a GKE cluster to run Portworx, you need to ensure that the cluster is based on Ubuntu. Due to certain restrictions with GKE clusters based on Container-Optimized OS (COS), Portworx requires Ubuntu as the base image for the GKE Nodes.

The following command configures a 3-node GKE Cluster in zone ap-south-1-a. You can modify the parameters accordingly.

$ gcloud container clusters create "gke-px" \
--zone "asia-south1-a" \
--username "admin" \
--machine-type "n1-standard-4" \
--image-type "UBUNTU" \
--disk-type "pd-ssd" \
--disk-size "100" \
--num-nodes "3" \
--enable-cloud-logging \
--enable-cloud-monitoring \
--network "default" \
--addons HorizontalPodAutoscaling,HttpLoadBalancing,KubernetesDashboard

Once the cluster is ready, configure kubectl CLI with the following command:

$ gcloud container clusters get-credentials gke-px --zone asia-south1-a

Portworx requires a ClusterRoleBinding for your user. Without this configuration, the command fails with an error clusterroles.rbac.authorization.k8s.io "portworx-pvc-controller-role" is forbidden.

Let’s create a ClusterRoleBinding with the following command:

$ kubectl create clusterrolebinding cluster-admin-binding \
--clusterrole cluster-admin \
--user $(gcloud config get-value account)

You should now have a three node Kubernetes cluster deployed in the Google Cloud Platform.

$ kubectl get nodes
NAME                                    STATUS   ROLES    AGE   VERSION
gke-gke-px-default-pool-2be69cfe-8189   Ready    <none>   42m   v1.10.9-gke.5
gke-gke-px-default-pool-2be69cfe-vbgq   Ready    <none>   43m   v1.10.9-gke.5
gke-gke-px-default-pool-2be69cfe-wj4z   Ready    <none>   43m   v1.10.9-gke.5

$ kubectl get nodes

 

Installing Portworx on GKE

Installing Portworx on GKE is not very different from installing it on any other Kubernetes cluster. Portworx GKE documentation has the steps involved in running the Portworx cluster in a Kubernetes environment deployed in GCP.

Once the GKE cluster is up and running, and Portworx is installed and configured, we will deploy a highly available Kafka and ZooKeeper StatefulSet.

The Portworx cluster needs to be up and running on GKE before proceeding to the next step. The kube-system namespace should have the Portworx pods in running state.

$ kubectl get pods -n=kube-system -l name=portworx
NAME             READY   STATUS    RESTARTS   AGE
portworx-pxfk8   1/1     Running   1          39m
portworx-zh4vf   1/1     Running   1          39m
portworx-zz5fn   1/1     Running   1          39m

$ kubectl get pods -n=kube-system -l name=portwor

 

Create a storage class for ZooKeeper and Kafka

Once the GKE cluster is up and running, and Portworx is installed and configured, we will deploy a highly available Kafka cluster in Kubernetes.

Through storage class objects, an admin can define different classes of Portworx volumes that are offered in a cluster. These classes will be used during the dynamic provisioning of volumes. The storage class defines the replication factor, I/O profile (e.g., for a database or a CMS), and priority (e.g., SSD or HDD). These parameters impact the availability and throughput of workloads and can be specified for each volume. This is important because a production database will have different requirements than a development Jenkins cluster.

In this step, we will create two storage classes for ZooKeeper and Kafka clusters.

Notice that we the replication factor for ZooKeeper storage class is set to 1. That’s because ZooKeeper servers keep their entire state machine in memory, and write every mutation to a durable WAL (Write Ahead Log) on storage media. When a server crashes, it can recover its previous state by replaying the WAL. To prevent the WAL from growing without bound, ZooKeeper servers will periodically snapshot their in-memory state to storage media. These snapshots can be loaded directly into memory, and all WAL entries that preceded the snapshot may be discarded.

Since ZooKeeper keeps the data in memory with an in-built recovery mechanism, we don’t need to configure Portworx volumes for replication.

We also defined a separate group for ZooKeeper and Kafka. This is helpful if we take 3DSnaps that are consistent across the whole cluster.

$ cat > px-ha-sc.yaml << EOF
kind: StorageClass
apiVersion: storage.k8s.io/v1
metadata:
  name: portworx-sc
provisioner: kubernetes.io/portworx-volume
parameters:
  repl: "1"
  io_priority: "high"
  group: "zk_vg"
---
kind: StorageClass
apiVersion: storage.k8s.io/v1
metadata:
  name: portworx-sc-rep3
provisioner: kubernetes.io/portworx-volume
parameters:
  repl: "3"
  io_priority: "high"
  group: "kafka_vg"
EOF

Create the storage classes and verify their availability in the default namespace.

$ kubectl create -f px-ha-sc.yaml
storageclass.storage.k8s.io/portworx-sc created
storageclass.storage.k8s.io/portworx-sc-rep3 created

$ kubectl get sc
NAME                 PROVISIONER                     AGE
portworx-sc          kubernetes.io/portworx-volume   23s
portworx-sc-rep3     kubernetes.io/portworx-volume   23s
standard (default)   kubernetes.io/gce-pd            52m
stork-snapshot-sc    stork-snapshot                  44m

 

Deploying ZooKeeper StatefulSet on GKE

As of v0.8, Kafka uses ZooKeeper for storing a variety of configurations as a key/value pair in the ZooKeeper data tree and uses them across the cluster in a distributed fashion. Our first task will be to deploy a 3 node ZooKeeper cluster using a StatefulSet backed by a Portworx volume.

We will then deploy a Kafka StatefulSet which uses our ZooKeeper cluster and also has Portworx volumes with 3 replicas. Using these replicas, we can have fast failover of the Kafka nodes and eliminate the I/O load during the rebuild. Additionally, because Portworx provides HA for Kafka, a typical customer can run fewer Kafka brokers for the same level of reliability, significantly reducing compute costs. Running 3 brokers instead of 5 is a 40% cost savings.

We will start by creating three objects.

1- A ConfigMap to inject configuration data into our ZooKeeper containers
2- A PodDisruptionBudget to limit the number of concurrent disruptions that ZooKeeper application experiences when we do maintenance operations on Kubernetes nodes
3- and finally, a Service for ZooKeeper so that Kafka can connect to the ZooKeeper cluster.

$ cat > zk-config.yaml << EOF
apiVersion: v1
kind: ConfigMap
metadata:
  name: zk-config
data:
  ensemble: "zk-0;zk-1;zk-2"
  jvm.heap: "512M"
  tick: "2000"
  init: "10"
  sync: "5"
  client.cnxns: "60"
  snap.retain: "3"
  purge.interval: "1"
---
apiVersion: policy/v1beta1
kind: PodDisruptionBudget
metadata:
  name: zk-budget
spec:
  selector:
    matchLabels:
      app: zk
  minAvailable: 2
---  
apiVersion: v1
kind: Service
metadata:
  name: zk-headless
  labels:
    app: zk-headless
spec:
  ports:
  - port: 2888
    name: server
  - port: 3888
    name: leader-election
  clusterIP: None
  selector:
    app: zk
EOF
$ kubectl create -f zk-config.yaml
configmap/zk-config created
poddisruptionbudget.policy/zk-budget created
service/zk-headless created

Now, deploy the ZooKeeper StatefulSet.

$ cat > zk-config.yaml << EOF
apiVersion: apps/v1beta1
kind: StatefulSet
metadata:
  name: zk
spec:
  serviceName: zk-headless
  replicas: 3
  template:
    metadata:
      labels:
        app: zk
      annotations:
        pod.alpha.kubernetes.io/initialized: "true"
    spec:
      # Use the stork scheduler to enable more efficient placement of the pods
      schedulerName: stork
      affinity:
        nodeAffinity:
          requiredDuringSchedulingIgnoredDuringExecution:
            nodeSelectorTerms:
            - matchExpressions:
              - key: px/running
                operator: NotIn
                values:
                - "false"
              - key: px/enabled
                operator: NotIn
                values:
                - "false"
        podAntiAffinity:
          requiredDuringSchedulingIgnoredDuringExecution:
            - labelSelector:
                matchExpressions:
                  - key: "app"
                    operator: In
                    values: 
                    - zk-headless
              topologyKey: "kubernetes.io/hostname"
      containers:
      - name: k8szk
        imagePullPolicy: Always
        image: gcr.io/google_samples/k8szk:v1
        ports:
        - containerPort: 2181
          name: client
        - containerPort: 2888
          name: server
        - containerPort: 3888
          name: leader-election
        env:
        - name : ZK_ENSEMBLE
          valueFrom:
            configMapKeyRef:
              name: zk-config
              key: ensemble
        - name : ZK_HEAP_SIZE
          valueFrom:
            configMapKeyRef:
                name: zk-config
                key: jvm.heap
        - name : ZK_TICK_TIME
          valueFrom:
            configMapKeyRef:
                name: zk-config
                key: tick
        - name : ZK_INIT_LIMIT
          valueFrom:
            configMapKeyRef:
                name: zk-config
                key: init
        - name : ZK_SYNC_LIMIT
          valueFrom:
            configMapKeyRef:
                name: zk-config
                key: tick
        - name : ZK_MAX_CLIENT_CNXNS
          valueFrom:
            configMapKeyRef:
                name: zk-config
                key: client.cnxns
        - name: ZK_SNAP_RETAIN_COUNT
          valueFrom:
            configMapKeyRef:
                name: zk-config
                key: snap.retain
        - name: ZK_PURGE_INTERVAL
          valueFrom:
            configMapKeyRef:
                name: zk-config
                key: purge.interval
        - name: ZK_CLIENT_PORT
          value: "2181"
        - name: ZK_SERVER_PORT
          value: "2888"
        - name: ZK_ELECTION_PORT
          value: "3888"
        command:
        - sh
        - -c
        - zkGenConfig.sh && zkServer.sh start-foreground
        readinessProbe:
          exec:
            command:
            - "zkOk.sh"
          initialDelaySeconds: 15
          timeoutSeconds: 5
        livenessProbe:
          exec:
            command:
            - "zkOk.sh"
          initialDelaySeconds: 15
          timeoutSeconds: 5
        volumeMounts:
        - name: datadir
          mountPath: /var/lib/zookeeper
      securityContext:
        runAsUser: 1000
        fsGroup: 1000
  volumeClaimTemplates:
  - metadata:
      name: datadir
    spec:
      storageClassName: portworx-sc
      accessModes: [ "ReadWriteOnce" ]
      resources:
        requests:
          storage: 2Gi
EOF
$ kubectl create -f zk-ss.yaml
statefulset.apps/zk created

Verify that all the pods are in the running state before proceeding further.

$ kubectl get pods
NAME   READY   STATUS    RESTARTS   AGE
zk-0   1/1     Running   0          5m
zk-1   1/1     Running   0          4m
zk-2   1/1     Running   0          3m

$ kubectl get pods

Let’s ensure that the ZooKeeper cluster is fully functional by creating and retrieving values from different nodes.

We create a key/value pair in node zk-0.

$ kubectl exec zk-0 -- /opt/zookeeper/bin/zkCli.sh create /hello world
Connecting to localhost:2181
2019-01-02 06:46:23,163 [myid:] - INFO  [main:Environment@100] - Client environment:zookeeper.version=3.4.9-1757313, built on 08/23/2016 06:50 GMT
2019-01-02 06:46:23,166 [myid:] - INFO  [main:Environment@100] - Client environment:host.name=zk-0.zk-headless.default.svc.cluster.local
2019-01-02 06:46:23,167 [myid:] - INFO  [main:Environment@100] - Client environment:java.version=1.8.0_111
2019-01-02 06:46:23,168 [myid:] - INFO  [main:Environment@100] - Client environment:java.vendor=Oracle Corporation
2019-01-02 06:46:23,168 [myid:] - INFO  [main:Environment@100] - Client environment:java.home=/usr/lib/jvm/java-8-openjdk-amd64/jre
2019-01-02 06:46:23,169 [myid:] - INFO  [main:Environment@100] - Client environment:java.class.path=/opt/zookeeper/bin/../build/classes:/opt/zookeeper/bin/../build/lib/*.jar:/opt/zookeeper/bin/../lib/slf4j-log4j12-1.6.1.jar:/opt/zookeeper/bin/../lib/slf4j-api-1.6.1.jar:/opt/zookeeper/bin/../lib/netty-3.10.5.Final.jar:/opt/zookeeper/bin/../lib/log4j-1.2.16.jar:/opt/zookeeper/bin/../lib/jline-0.9.94.jar:/opt/zookeeper/bin/../zookeeper-3.4.9.jar:/opt/zookeeper/bin/../src/java/lib/*.jar:/opt/zookeeper/bin/../conf:
2019-01-02 06:46:23,169 [myid:] - INFO  [main:Environment@100] - Client environment:java.library.path=/usr/java/packages/lib/amd64:/usr/lib/x86_64-linux-gnu/jni:/lib/x86_64-linux-gnu:/usr/lib/x86_64-linux-gnu:/usr/lib/jni:/lib:/usr/lib
2019-01-02 06:46:23,169 [myid:] - INFO  [main:Environment@100] - Client environment:java.io.tmpdir=/tmp
2019-01-02 06:46:23,169 [myid:] - INFO  [main:Environment@100] - Client environment:java.compiler=
2019-01-02 06:46:23,169 [myid:] - INFO  [main:Environment@100] - Client environment:os.name=Linux
2019-01-02 06:46:23,169 [myid:] - INFO  [main:Environment@100] - Client environment:os.arch=amd64
2019-01-02 06:46:23,169 [myid:] - INFO  [main:Environment@100] - Client environment:os.version=4.15.0-1017-gcp
2019-01-02 06:46:23,169 [myid:] - INFO  [main:Environment@100] - Client environment:user.name=zookeeper
2019-01-02 06:46:23,169 [myid:] - INFO  [main:Environment@100] - Client environment:user.home=/home/zookeeper
2019-01-02 06:46:23,169 [myid:] - INFO  [main:Environment@100] - Client environment:user.dir=/
2019-01-02 06:46:23,170 [myid:] - INFO  [main:ZooKeeper@438] - Initiating client connection, connectString=localhost:2181 sessionTimeout=30000 watcher=org.apache.zookeeper.ZooKeeperMain$MyWatcher@1de0aca6
2019-01-02 06:46:23,192 [myid:] - INFO  [main-SendThread(localhost:2181):ClientCnxn$SendThread@1032] - Opening socket connection to server localhost/127.0.0.1:2181. Will not attempt to authenticate using SASL (unknown error)
2019-01-02 06:46:23,242 [myid:] - INFO  [main-SendThread(localhost:2181):ClientCnxn$SendThread@876] - Socket connection established to localhost/127.0.0.1:2181, initiating session
2019-01-02 06:46:23,275 [myid:] - INFO  [main-SendThread(localhost:2181):ClientCnxn$SendThread@1299] - Session establishment complete on server localhost/127.0.0.1:2181, sessionid = 0x1680d4b4f400000, negotiated timeout = 30000

WATCHER::

WatchedEvent state:SyncConnected type:None path:null
Created /hello

$ kubectl exec zk-0 -- /opt/zookeeper/bin/zkCli.sh create /hello w

Let’s now retrieve the value from node zk-2.

$ kubectl exec zk-2 -- /opt/zookeeper/bin/zkCli.sh get /hello
Connecting to localhost:2181
2019-01-02 06:49:53,232 [myid:] - INFO  [main:Environment@100] - Client environment:zookeeper.version=3.4.9-1757313, built on 08/23/2016 06:50 GMT
2019-01-02 06:49:53,236 [myid:] - INFO  [main:Environment@100] - Client environment:host.name=zk-2.zk-headless.default.svc.cluster.local
2019-01-02 06:49:53,236 [myid:] - INFO  [main:Environment@100] - Client environment:java.version=1.8.0_111
2019-01-02 06:49:53,238 [myid:] - INFO  [main:Environment@100] - Client environment:java.vendor=Oracle Corporation
2019-01-02 06:49:53,239 [myid:] - INFO  [main:Environment@100] - Client environment:java.home=/usr/lib/jvm/java-8-openjdk-amd64/jre
2019-01-02 06:49:53,239 [myid:] - INFO  [main:Environment@100] - Client environment:java.class.path=/opt/zookeeper/bin/../build/classes:/opt/zookeeper/bin/../build/lib/*.jar:/opt/zookeeper/bin/../lib/slf4j-log4j12-1.6.1.jar:/opt/zookeeper/bin/../lib/slf4j-api-1.6.1.jar:/opt/zookeeper/bin/../lib/netty-3.10.5.Final.jar:/opt/zookeeper/bin/../lib/log4j-1.2.16.jar:/opt/zookeeper/bin/../lib/jline-0.9.94.jar:/opt/zookeeper/bin/../zookeeper-3.4.9.jar:/opt/zookeeper/bin/../src/java/lib/*.jar:/opt/zookeeper/bin/../conf:
2019-01-02 06:49:53,239 [myid:] - INFO  [main:Environment@100] - Client environment:java.library.path=/usr/java/packages/lib/amd64:/usr/lib/x86_64-linux-gnu/jni:/lib/x86_64-linux-gnu:/usr/lib/x86_64-linux-gnu:/usr/lib/jni:/lib:/usr/lib
2019-01-02 06:49:53,239 [myid:] - INFO  [main:Environment@100] - Client environment:java.io.tmpdir=/tmp
2019-01-02 06:49:53,239 [myid:] - INFO  [main:Environment@100] - Client environment:java.compiler=
2019-01-02 06:49:53,239 [myid:] - INFO  [main:Environment@100] - Client environment:os.name=Linux
2019-01-02 06:49:53,239 [myid:] - INFO  [main:Environment@100] - Client environment:os.arch=amd64
2019-01-02 06:49:53,240 [myid:] - INFO  [main:Environment@100] - Client environment:os.version=4.15.0-1017-gcp
2019-01-02 06:49:53,240 [myid:] - INFO  [main:Environment@100] - Client environment:user.name=zookeeper
2019-01-02 06:49:53,240 [myid:] - INFO  [main:Environment@100] - Client environment:user.home=/home/zookeeper
2019-01-02 06:49:53,240 [myid:] - INFO  [main:Environment@100] - Client environment:user.dir=/
2019-01-02 06:49:53,241 [myid:] - INFO  [main:ZooKeeper@438] - Initiating client connection, connectString=localhost:2181 sessionTimeout=30000 watcher=org.apache.zookeeper.ZooKeeperMain$MyWatcher@1de0aca6
2019-01-02 06:49:53,265 [myid:] - INFO  [main-SendThread(localhost:2181):ClientCnxn$SendThread@1032] - Opening socket connection to server localhost/127.0.0.1:2181. Will not attempt to authenticate using SASL (unknown error)
2019-01-02 06:49:53,333 [myid:] - INFO  [main-SendThread(localhost:2181):ClientCnxn$SendThread@876] - Socket connection established to localhost/127.0.0.1:2181, initiating session
2019-01-02 06:49:53,353 [myid:] - INFO  [main-SendThread(localhost:2181):ClientCnxn$SendThread@1299] - Session establishment complete on server localhost/127.0.0.1:2181, sessionid = 0x3680d4c049c0000, negotiated timeout = 30000

WATCHER::

WatchedEvent state:SyncConnected type:None path:null
cZxid = 0x100000002
world
ctime = Wed Jan 02 06:46:23 UTC 2019
mZxid = 0x100000002
mtime = Wed Jan 02 06:46:23 UTC 2019
pZxid = 0x100000002
cversion = 0
dataVersion = 0
aclVersion = 0
ephemeralOwner = 0x0
dataLength = 5
numChildren = 0

$ kubectl exec zk-0 -- /opt/zookeeper/bin/zkCli.sh create /hello w

 

Deploying Kafka StatefulSet on GKE

With the ZooKeeper cluster in place, it’s time for us to deploy our Kafka cluster. Let’s start by creating the ConfigMap and a headless service required by Kafka.

$ cat > kafka-config.yaml << EOF kind: ConfigMap metadata: name: broker-config namespace: default apiVersion: v1 data: init.sh: |- #!/bin/bash set -x KAFKA_BROKER_ID=${HOSTNAME##*-} cp -Lur /etc/kafka-configmap/* /etc/kafka/ sed -i "s/#init#broker.id=#init#/broker.id=$KAFKA_BROKER_ID/" /etc/kafka/server.properties hash kubectl 2>/dev/null || {
      sed -i "s/#init#broker.rack=#init#/#init#broker.rack=# kubectl not found in path/" /etc/kafka/server.properties
    } && {
      ZONE=$(kubectl get node "$NODE_NAME" -o=go-template='{{index .metadata.labels "failure-domain.beta.kubernetes.io/zone"}}')
      if [ $? -ne 0 ]; then
        sed -i "s/#init#broker.rack=#init#/#init#broker.rack=# zone lookup failed, see -c init-config logs/" /etc/kafka/server.properties
      elif [ "x$ZONE" == "x" ]; then
        sed -i "s/#init#broker.rack=#init#/#init#broker.rack=# zone label not found for node $NODE_NAME/" /etc/kafka/server.properties
      else
        sed -i "s/#init#broker.rack=#init#/broker.rack=$ZONE/" /etc/kafka/server.properties
      fi
    }

  server.properties: |-
    delete.topic.enable=true
    num.network.threads=3
    num.io.threads=8
    socket.send.buffer.bytes=102400
    socket.receive.buffer.bytes=102400
    socket.request.max.bytes=104857600
    log.dirs=/tmp/kafka-logs
    num.partitions=1
    num.recovery.threads.per.data.dir=1
    offsets.topic.replication.factor=1
    transaction.state.log.replication.factor=1
    transaction.state.log.min.isr=1
    log.retention.hours=168
    log.segment.bytes=1073741824
    log.retention.check.interval.ms=300000
    zookeeper.connect=zk-0.zk-headless.default.svc.cluster.local:2181,zk-1.zk-headless.default.svc.cluster.local:2181,zk-2.zk-headless.default.svc.cluster.local:2181
    zookeeper.connection.timeout.ms=6000
    group.initial.rebalance.delay.ms=0

  log4j.properties: |-
    log4j.rootLogger=INFO, stdout

    log4j.appender.stdout=org.apache.log4j.ConsoleAppender
    log4j.appender.stdout.layout=org.apache.log4j.PatternLayout
    log4j.appender.stdout.layout.ConversionPattern=[%d] %p %m (%c)%n

    log4j.appender.kafkaAppender=org.apache.log4j.DailyRollingFileAppender
    log4j.appender.kafkaAppender.DatePattern='.'yyyy-MM-dd-HH
    log4j.appender.kafkaAppender.File=${kafka.logs.dir}/server.log
    log4j.appender.kafkaAppender.layout=org.apache.log4j.PatternLayout
    log4j.appender.kafkaAppender.layout.ConversionPattern=[%d] %p %m (%c)%n

    log4j.appender.stateChangeAppender=org.apache.log4j.DailyRollingFileAppender
    log4j.appender.stateChangeAppender.DatePattern='.'yyyy-MM-dd-HH
    log4j.appender.stateChangeAppender.File=${kafka.logs.dir}/state-change.log
    log4j.appender.stateChangeAppender.layout=org.apache.log4j.PatternLayout
    log4j.appender.stateChangeAppender.layout.ConversionPattern=[%d] %p %m (%c)%n

    log4j.appender.requestAppender=org.apache.log4j.DailyRollingFileAppender
    log4j.appender.requestAppender.DatePattern='.'yyyy-MM-dd-HH
    log4j.appender.requestAppender.File=${kafka.logs.dir}/kafka-request.log
    log4j.appender.requestAppender.layout=org.apache.log4j.PatternLayout
    log4j.appender.requestAppender.layout.ConversionPattern=[%d] %p %m (%c)%n

    log4j.appender.cleanerAppender=org.apache.log4j.DailyRollingFileAppender
    log4j.appender.cleanerAppender.DatePattern='.'yyyy-MM-dd-HH
    log4j.appender.cleanerAppender.File=${kafka.logs.dir}/log-cleaner.log
    log4j.appender.cleanerAppender.layout=org.apache.log4j.PatternLayout
    log4j.appender.cleanerAppender.layout.ConversionPattern=[%d] %p %m (%c)%n

    log4j.appender.controllerAppender=org.apache.log4j.DailyRollingFileAppender
    log4j.appender.controllerAppender.DatePattern='.'yyyy-MM-dd-HH
    log4j.appender.controllerAppender.File=${kafka.logs.dir}/controller.log
    log4j.appender.controllerAppender.layout=org.apache.log4j.PatternLayout
    log4j.appender.controllerAppender.layout.ConversionPattern=[%d] %p %m (%c)%n

    log4j.appender.authorizerAppender=org.apache.log4j.DailyRollingFileAppender
    log4j.appender.authorizerAppender.DatePattern='.'yyyy-MM-dd-HH
    log4j.appender.authorizerAppender.File=${kafka.logs.dir}/kafka-authorizer.log
    log4j.appender.authorizerAppender.layout=org.apache.log4j.PatternLayout
    log4j.appender.authorizerAppender.layout.ConversionPattern=[%d] %p %m (%c)%n

    # Change the two lines below to adjust ZK client logging
    log4j.logger.org.I0Itec.zkclient.ZkClient=INFO
    log4j.logger.org.apache.zookeeper=INFO

    # Change the two lines below to adjust the general broker logging level (output to server.log and stdout)
    log4j.logger.kafka=INFO
    log4j.logger.org.apache.kafka=INFO

    # Change to DEBUG or TRACE to enable request logging
    log4j.logger.kafka.request.logger=WARN, requestAppender
    log4j.additivity.kafka.request.logger=false

    log4j.logger.kafka.network.RequestChannel$=WARN, requestAppender
    log4j.additivity.kafka.network.RequestChannel$=false

    log4j.logger.kafka.controller=TRACE, controllerAppender
    log4j.additivity.kafka.controller=false

    log4j.logger.kafka.log.LogCleaner=INFO, cleanerAppender
    log4j.additivity.kafka.log.LogCleaner=false

    log4j.logger.state.change.logger=TRACE, stateChangeAppender
    log4j.additivity.state.change.logger=false

    log4j.logger.kafka.authorizer.logger=WARN, authorizerAppender
    log4j.additivity.kafka.authorizer.logger=false

---
apiVersion: v1
kind: Service
metadata:
  name: kafka-broker
  namespace: default
spec:
  ports:
  - port: 9092
  # [podname].broker.kafka.svc.cluster.local
  clusterIP: None
  selector:
    app: kafka
---
EOF
$ kubectl create -f kafka-config.yaml
configmap/broker-config created
service/kafka-broker created

Deploy Kafka cluster with a single node with the below manifest:

$ cat > kafka-ss.yaml << EOF
apiVersion: apps/v1beta1
kind: StatefulSet
metadata:
  name: kafka
  namespace: default
spec:
  serviceName: "kafka-broker"
  replicas: 1
  template:
    metadata:
      labels:
        app: kafka
      annotations:
    spec:
      # Use the stork scheduler to enable more efficient placement of the pods
      schedulerName: stork
      terminationGracePeriodSeconds: 30
      affinity:
        nodeAffinity:
          requiredDuringSchedulingIgnoredDuringExecution:
            nodeSelectorTerms:
            - matchExpressions:
              - key: px/running
                operator: NotIn
                values:
                - "false"
              - key: px/enabled
                operator: NotIn
                values:
                - "false"
      initContainers:
      - name: init-config
        image: solsson/kafka-initutils@sha256:c275d681019a0d8f01295dbd4a5bae3cfa945c8d0f7f685ae1f00f2579f08c7d
        env:
        - name: NODE_NAME
          valueFrom:
            fieldRef:
              fieldPath: spec.nodeName
        command: ['/bin/bash', '/etc/kafka-configmap/init.sh']
        volumeMounts:
        - name: configmap
          mountPath: /etc/kafka-configmap
        - name: config
          mountPath: /etc/kafka
      containers:
      - name: broker
        image: solsson/kafka:0.11.0.0@sha256:b27560de08d30ebf96d12e74f80afcaca503ad4ca3103e63b1fd43a2e4c976ce
        env:
        - name: KAFKA_LOG4J_OPTS
          value: -Dlog4j.configuration=file:/etc/kafka/log4j.properties
        ports:
        - containerPort: 9092
        command:
        - ./bin/kafka-server-start.sh
        - /etc/kafka/server.properties
        - --override
        -   zookeeper.connect=zk-0.zk-headless.default.svc.cluster.local:2181,zk-1.zk-headless.default.svc.cluster.local:2181,zk-2.zk-headless.default.svc.cluster.local:2181
        - --override
        -   log.retention.hours=-1
        - --override
        -   log.dirs=/var/lib/kafka/data/topics
        - --override
        -   auto.create.topics.enable=false
        resources:
          requests:
            cpu: 100m
            memory: 512Mi
        readinessProbe:
          exec:
            command:
            - /bin/sh
            - -c
            - 'echo "" | nc -w 1 127.0.0.1 9092'
        volumeMounts:
        - name: config
          mountPath: /etc/kafka
        - name: data
          mountPath: /var/lib/kafka/data
      volumes:
      - name: configmap
        configMap:
          name: broker-config
      - name: config
        emptyDir: {}
  volumeClaimTemplates:
  - metadata:
      name: data
    spec:
      storageClassName: portworx-sc-rep3
      accessModes: [ "ReadWriteOnce" ]
      resources:
        requests:
          storage: 3Gi
EOF
$ kubectl create -f kafka-ss.yaml
statefulset.apps/kafka created 

To make it easy to communicate with the Kafka cluster, let’s create a Pod with the Kafka CLI.

$ cat > kafka-cli.yaml << EOF
apiVersion: v1
kind: Pod
metadata:
  name: kafka-cli
spec:
  containers:
  - name: kafka
    image: solsson/kafka:0.11.0.0
    command:
      - sh
      - -c
      - "exec tail -f /dev/null"
EOF
$ kubectl create -f kafka-cli.yaml
pod/kafka-cli created

Verify that the Kafka StatefulSet is up and running.

$ kubectl get pods -l app=kafka
NAME      READY   STATUS    RESTARTS   AGE
kafka-0   1/1     Running   0          3m

We can now inspect the Portworx volume associated with the Kafka pod by accessing the pxctl tool.

$ VOL=`kubectl get pvc | grep kafka | awk '{print $3}'`
$ PX_POD=$(kubectl get pods -l name=portworx -n kube-system -o jsonpath='{.items[0].metadata.name}')
$ kubectl exec -it $PX_POD -n kube-system -- /opt/pwx/bin/pxctl volume inspect ${VOL}
Volume	:  1081594002605189656
	Name            	 :  pvc-82b55251-0e5c-11e9-a3a6-42010aa0009b
	Group            	 :  kafka_vg
	Size            	 :  3.0 GiB
	Format          	 :  ext4
	HA              	 :  3
	IO Priority     	 :  MEDIUM
	Creation time   	 :  Jan 2 07:03:34 UTC 2019
	Shared          	 :  no
	Status          	 :  up
	State           	 :  Attached: 691a2da6-b9e5-4d24-81a3-6b52fd80facb (10.160.0.3)
	Device Path     	 :  /dev/pxd/pxd1081594002605189656
	Labels          	 :  namespace=default,pvc=data-kafka-0
	Reads           	 :  36
	Reads MS        	 :  72
	Bytes Read      	 :  372736
	Writes          	 :  113
	Writes MS       	 :  204
	Bytes Written   	 :  50499584
	IOs in progress 	 :  0
	Bytes used      	 :  113 MiB
	Replica sets on nodes:
		Set 0
		  Node 		 : 10.160.0.4 (Pool 0)
		  Node 		 : 10.160.0.3 (Pool 0)
		  Node 		 : 10.160.0.2 (Pool 0)
	Replication Status	 :  Up
	Volume consumers	 :
		- Name           : kafka-0 (82b61fdc-0e5c-11e9-a3a6-42010aa0009b) (Pod)
		  Namespace      : default
		  Running on     : gke-gke-px-default-pool-2be69cfe-vbgq
		  Controlled by  : kafka (StatefulSet)

$ VOL=`kubectl get pvc | grep kafka | awk '{print $3}

 

Failing over a Kafka pod on Kubernetes

Let’s start by ingesting sample messages into a topic through the CLI pod. Hit CTRL+C after entering the last message.

$ kubectl exec -it kafka-cli bash
# ./bin/kafka-topics.sh --create --zookeeper zk-headless:2181 --replication-factor 1 --partitions 1 --topic test
# ./bin/kafka-console-producer.sh --broker-list kafka-broker:9092 --topic test

>message 1
>message 2
>message 3

# ./bin/kafka-console-consumer.sh --bootstrap-server kafka-broker:9092 --topic test --partition 0 --from-beginning
message 1
message 2
message 3
Processed a total of 3 messages
# exit

We will now simulate the failover by cordoning off one of the nodes and deleting the Kafka Pod deployed on it. When the new Pod is created it has the same data as the original Pod.

$ NODE=`kubectl get pods -o wide | grep kafka-0 | awk '{print $7}'`
$ kubectl cordon ${NODE}

$ kubectl get nodes

$ kubectl delete pod kafka-0
pod "kafka-0" deleted

Kubernetes controller now tries to create the Pod on a different node. Wait for the Kafka Pod to be in Running state on the node. Don’t forget to uncordon the node before proceeding further.

$ kubectl get pods -l app=kafka -o wide
NAME      READY   STATUS    RESTARTS   AGE   IP           NODE
kafka-0   1/1     Running   0          38s   10.48.2.13   gke-gke-px-default-pool-2be69cfe-8189

$ kubectl get pod

$ kubectl uncordon ${NODE}
node/gke-gke-px-default-pool-2be69cfe-vbgq uncordoned

Finally, let’s verify that the messages are still available under the test topic.

$ kubectl exec -it kafka-cli bash
# ./bin/kafka-console-consumer.sh --bootstrap-server kafka-broker:9092 --topic test --partition 0 --from-beginning
message 1
message 2
message 3
Processed a total of 3 messages

 

Backing up and restoring a Kafka node through snapshots

Portworx supports creating Snapshots for Kubernetes PVCs. Since there is only one Kafka node, we can use regular, local snapshots to backup and restore.

For production scenarios where there is more than one Kafka node in the cluster, it is highly recommended that customers use 3DSnap for application consistent backup and recovery.

Let’s create a snapshot for the Kubernetes PVC we created for MySQL.

cat >  kafka-snap.yaml << EOF
apiVersion: volumesnapshot.external-storage.k8s.io/v1
kind: VolumeSnapshot
metadata:
  name: px-kafka-snapshot
  namespace: default
spec:
  persistentVolumeClaimName: data-kafka-0
EOF
$ kubectl create -f kafka-snap.yaml
volumesnapshot.volumesnapshot.external-storage.k8s.io "px-kafka-snapshot" created

Verify the creation of the volume snapshot.

$ kubectl get volumesnapshot
NAME                AGE
px-kafka-snapshot   30s
$ kubectl get volumesnapshotdatas
NAME                                                       AGE
k8s-volume-snapshot-b1c06e67-1feb-11e9-8f35-0a580a30020a   34s

With the snapshot in place, let’s go ahead and delete the Kafka StatefulSet and the associated PVC.

$ kubectl delete sts/kafka
statefulset.apps "kafka" deleted
$ kubectl delete sts/kafka
persistentvolumeclaim "data-kafka-0" deleted

Since snapshots are just like volumes, we can use it to create a new PVC. Since we follow the standard naming convention used by Kubernetes StatefulSets, the Kafka node will be automatically associated with this PVC.

Create a new PVC definition from the snapshot. Notice how the new PVC is named. It’s the same as the original PVC that we deleted after taking the snapshot.

cat >  kafka-snap.yaml << EOF
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: data-kafka-0
  annotations:
    snapshot.alpha.kubernetes.io/snapshot: px-kafka-snapshot
spec:
  accessModes:
     - ReadWriteOnce
  storageClassName: stork-snapshot-sc
  resources:
    requests:
      storage: 5Gi
EOF
$ kubectl create -f kafka-snap-pvc.yaml
persistentvolumeclaim/data-kafka-0 created

Launch a new StatefulSet backed by the restored PVC.

$ kubectl create -f kafka-ss.yaml
statefulset.apps/kafka created

Once the Pod is ready, let’s access it through the CLI Pod to check the availability of messages sent by the producer.

$ kubectl exec -it kafka-cli bash
root@kafka-cli:/opt/kafka# ./bin/kafka-console-consumer.sh --bootstrap-server kafka-broker:9092 --topic test --partition 0 --from-beginning
msg 1
msg 2
msg 3
Processed a total of 3 messages
root@kafka-cli:/opt/kafka#

Congratulations! You have successfully restored a snapshot for Kafka.

 

Summary

Portworx can easily be deployed on Google GKE to run stateful workloads in production. It integrates well with Kubernetes StatefulSets by providing dynamic provisioning. Additional operations such as expanding the volumes and performing backups stored as snapshots on object storage can be performed while managing production workloads.

Janakiram MSV

Contributor | Certified Kubernetes Administrator (CKA) and Developer (CKAD)

Share Share on Facebook Tweet about this on Twitter Share on LinkedIn



Back to Blog