Hive / Helm / On Minikube |

This page shows how to use Helm and a pre-built Docker image available at DockerHub in order to operate Hive on MR3 on Minikube. All components (Metastore, HiveServer2, MR3 DAGAppMaster) will be running inside Minikube. For Metastore, we will run a MySQL database as a Pod inside Minikube. By following the instruction, the user will learn:

how to start Metastore using Helm
how to use Helm to run Hive on MR3 on Minikube
how to create Beeline connections and send queries to HiveServer2 running inside Minikube

This scenario has the following prerequisites:

A running Minikube cluster is available.
The user should be able to execute: 1) command kubectl; 2) command helm to use Helm; 3) command docker if no MySQL database is available.

This scenario should take less than 30 minutes to complete, not including the time for downloading a pre-built Docker image. All commands are executed by user gla. We use Helm 2.17.0.

This page has been tested with MR3 release 1.5 on CentOS 7.5 running Minikube v1.15.0.

For asking questions, please visit MR3 Google Group.

Installation

Download an MR3 release containing the executable scripts.

$ git clone https://github.com/mr3project/mr3-run-k8s.git
$ cd mr3-run-k8s/kubernetes/
# git reset --hard 9cf574369d8c9f21cc6ebf0fc06606f24f0a858e

Starting a MySQL database

For simplicity, we will run a MySQL database for Metastore as a Docker container.

$ docker run -d --name mysql-server -p 3306:3306 -e MYSQL_ROOT_PASSWORD=passwd mysql:5.6
$ mysql --user=root --password=passwd --host=127.0.0.1 -e 'show databases;'
+--------------------+
| Database           |
+--------------------+
| information_schema |
| mysql              |
| performance_schema |
+--------------------+

Linking configuration files

We will reuse the configuration files in conf/ (and keys in key if Kerberos is used for authentication). Create symbolic links.

$ mkdir -p key
$ ln -s $(pwd)/conf/ helm/hive/conf
$ ln -s $(pwd)/key/ helm/hive/key

Now any change to the configuration files in conf/ is honored when running Hive on MR3.

Creating local directories

We need to create two new local directories:

for a PersistentVolume to be shared by Pods;
for a hostPath volume for ContainerWorker Pods.

Create a local directory for the PersistentVolume. In our example, we use /home/gla/workdir for the directory for the PersistentVolume.

$ mkdir -p /home/gla/workdir
$ chmod 777 /home/gla/workdir

Hive on MR3 uses local disks for writing intermediate data. In the case of running on Kubernetes, we mount hostPath volumes to mount directories of the local machine. For our example, we create a local directory /data1/gla/k8s for the hostPath volume for ContainerWorker Pods.

$ mkdir -p /data1/gla/k8s
$ chmod 777 /data1/gla/k8s

Configuring Pods

Create a new file helm/hive/values-minikube.yaml which is a collection of values to override those in helm/hive/values.yaml.

$ vi helm/hive/values-minikube.yaml

docker:
  image: mr3project/hive3:1.5
  imagePullPolicy: IfNotPresent

create:
  metastore: true
  
metastore:
  databaseHost: 192.168.10.1   # use your IP address
  warehouseDir: file:///opt/mr3-run/work-dir/warehouse
  initSchema: true
  resources:
    requests:
      cpu: 1
      memory: 4Gi
    limits:
      cpu: 1
      memory: 4Gi
  heapSize: 4096

hive:
  externalIp: 192.168.10.1    # use your IP address
  resources:
    requests:
      cpu: 1
      memory: 8Gi
    limits:
      cpu: 1
      memory: 8Gi
  heapSize: 8192

workDir:
  isNfs: false
  volumeStr: "hostPath:\n  path: /home/gla/workdir"

docker.image is set to a pre-built Docker image mr3project/hive3:1.5 available at DockerHub.
docker.imagePullPolicy is set to IfNotPresent because we download the Docker image from DockerHub.
create.metastore is set to true because we will create a Metastore Pod.
metastore.databaseHost is set to the address of the MySQL database.
metastore.initSchema is set to true because it is the first time to run Metastore. For subsequent runs, the user may set it to false.
hive.externalIp is set to the public IP address of the local machine.
workDir.volumeStr is set to the path to the local directory for the PersistentVolume.

Update helm/hive/templates/metastore.yaml to remove nodeAffinity as we do not use node affinity rules.

$ vi helm/hive/templates/metastore.yaml

      affinity:
      # nodeAffinity:
      #   requiredDuringSchedulingIgnoredDuringExecution:
      #     nodeSelectorTerms:
      #     - matchExpressions:
      #       - key: roles
      #         operator: In
      #         values:
      #         - "masters"

Update conf/mr3-site.xml.

$ vi conf/mr3-site.xml

<property>
  <name>mr3.k8s.pod.image.pull.policy</name>
  <value>IfNotPresent</value>
</property>

<property>
  <name>mr3.k8s.pod.worker.hostpaths</name>
  <value>/data1/gla/k8s/</value>
</property>

mr3.k8s.pod.image.pull.policy is set to IfNotPresent because we download the Docker image from DockerHub.
mr3.k8s.pod.worker.hostpaths is set to the path to the local directory for the hostPath volume.

Update conf/hive-site.xml.

$ vi conf/hive-site.xml

<property>
  <name>javax.jdo.option.ConnectionUserName</name>
  <value>root</value>
</property>

<property>
  <name>javax.jdo.option.ConnectionPassword</name>
  <value>passwd</value>
</property>

<property>
  <name>hive.metastore.pre.event.listeners</name>
  <value></value>
</property>
<property>
  <name>metastore.pre.event.listeners</name>
  <value></value>
</property>

<property>
  <name>hive.security.authorization.manager</name>
  <value>org.apache.hadoop.hive.ql.security.authorization.plugin.sqlstd.SQLStdHiveAuthorizerFactory</value> 
</property>

javax.jdo.option.ConnectionPassword is set to the password of the MySQL database.
hive.metastore.pre.event.listeners and metastore.pre.event.listeners are set to empty because we do not enable security on the Metastore side.
hive.security.authorization.manager is set to use SQLStdHiveAuthorizerFactory.

Starting Hive on MR3

Before running HiveServer2, the user should remove the label node-role.kubernetes.io/master from minikube node. This is because Hive on MR3 does not count the resources of master nodes when estimating the resources for ContainerWorker Pods. Since minikube node, the only node in a Minikube cluster, is a master node, we should demote it to an ordinary node in order to secure resources for ContainerWorker Pods. Thus, in order to be able to create ContainerWorker Pods in minikube node, the user should execute the following command:

$ kubectl label node minikube node-role.kubernetes.io/master-

Before running HiveServer2, the user should also make sure that no ConfigMaps and Services exist in the namespace hivemr3. For example, the user may see ConfigMaps and Services left over from a previous run.

$ kubectl get configmaps -n hivemr3
NAME                       DATA   AGE
mr3conf-configmap-master   1      16m
mr3conf-configmap-worker   1      16m

$ kubectl get svc -n hivemr3
NAME                    TYPE        CLUSTER-IP      EXTERNAL-IP   PORT(S)   AGE
service-master-1237-0   ClusterIP   10.105.238.21   <none>        80/TCP    11m
service-worker          ClusterIP   None            <none>        <none>    11m

In such a case, manually delete these ConfigMaps and Services.

$ kubectl delete configmap -n hivemr3 mr3conf-configmap-master mr3conf-configmap-worker
$ kubectl delete svc -n hivemr3 service-master-1237-0 service-worker

Install Helm chart for Hive on MR3 with values-minikube.yaml. We use hivemr3 for the namespace. Metastore automatically downloads a MySQL connector from https://cdn.mysql.com/Downloads/Connector-J/mysql-connector-java-8.0.28.tar.gz.

$ helm install --namespace hivemr3 helm/hive -f helm/hive/values-minikube.yaml
2022/07/30 23:25:20 found symbolic link in path: /data1/gla/mr3-run-k8s/kubernetes/helm/hive/conf resolves to /data1/gla/mr3-run-k8s/kubernetes/conf
2022/07/30 23:25:20 found symbolic link in path: /data1/gla/mr3-run-k8s/kubernetes/helm/hive/key resolves to /data1/gla/mr3-run-k8s/kubernetes/key
NAME:   jaundiced-lightningbug
LAST DEPLOYED: Sat Jul 30 23:25:20 2022
NAMESPACE: hivemr3
STATUS: DEPLOYED
...

==> v1/ConfigMap
NAME                    DATA  AGE
client-am-config        4     0s
env-configmap           1     0s
hivemr3-conf-configmap  15    0s
...

Check if all ConfigMaps are non-empty. If the DATA column for hivemr3-conf-configmap is 0, try to remove unnecessary files in the directory conf or helm/hive/conf. This usually happens when a temporary file (e.g., .hive-site.xml.swp) is kept at the time of installing Helm chart.

The user can find three Pods running in the Minikube cluster: 1) Metastore; 2) HiveServer2; 3) MR3 DAGAppMaster.

$ kubectl get pods -n hivemr3
NAME                                  READY   STATUS    RESTARTS   AGE
hivemr3-hiveserver2-b489d4d7f-s77b8   1/1     Running   0          111s
hivemr3-metastore-0                   1/1     Running   0          111s
mr3master-4609-0-dfff6fc7f-g8hbv      1/1     Running   0          79s

HiveServer2 Pod becomes ready after a readiness probe contacts it.

Running Beeline

Download a sample dataset and copy it to the directory for the PersistentVolume.

$ wget https://github.com/mr3project/mr3-release/releases/download/v1.0/pokemon.csv
$ cp pokemon.csv /home/gla/workdir
$ chmod 777 /home/gla/workdir/pokemon.csv

The user can verify that the sample dataset is accessible inside the HiveServer2 Pod.

$ kubectl exec -n hivemr3 -it hivemr3-hiveserver2-b489d4d7f-s77b8 -- /bin/bash
hive@hivemr3-hiveserver2-b489d4d7f-s77b8:/opt/mr3-run/hive$ ls /opt/mr3-run/work-dir/
91c13320-b051-41b1-a0d9-d6c4c7d218a3_resources	hive
_resultscache_					lib
db7da8c1-d360-43d4-a72a-cbad18e64832_resources	pokemon.csv
hive@hivemr3-hiveserver2-b489d4d7f-s77b8:/opt/mr3-run/hive$ exit

The user may use any client program (such as beeline) to connect to HiveServer2 which is running at port 9852. Alternatively the user can run Beeline inside the Hiveserver2 Pod.

$ kubectl exec -n hivemr3 -it hivemr3-hiveserver2-b489d4d7f-s77b8 -- /bin/bash
hive@hivemr3-hiveserver2-b489d4d7f-s77b8:/opt/mr3-run/hive$ export USER=root
hive@hivemr3-hiveserver2-b489d4d7f-s77b8:/opt/mr3-run/hive$ /opt/mr3-run/hive/run-beeline.sh
...
Connecting to jdbc:hive2://hivemr3-hiveserver2-b489d4d7f-s77b8:9852/;;;
Connected to: Apache Hive (version 3.1.3)
Driver: Hive JDBC (version 3.1.3)
Transaction isolation: TRANSACTION_REPEATABLE_READ
Beeline version 3.1.3 by Apache Hive
0: jdbc:hive2://hivemr3-hiveserver2-b489d4d7f>

Execute queries.

0: jdbc:hive2://hivemr3-hiveserver2-b489d4d7f> show databases;
...
+----------------+
| database_name  |
+----------------+
| default        |
+----------------+
1 row selected (2.129 seconds)

0: jdbc:hive2://hivemr3-hiveserver2-b489d4d7f> use default;
...
No rows affected (0.054 seconds)

0: jdbc:hive2://hivemr3-hiveserver2-b489d4d7f> CREATE TABLE pokemon (Number Int,Name String,Type1 String,Type2 String,Total Int,HP Int,Attack Int,Defense Int,Sp_Atk Int,Sp_Def Int,Speed Int) row format delimited fields terminated BY ',' lines terminated BY '\n' tblproperties("skip.header.line.count"="1");
...
No rows affected (1.416 seconds)

0: jdbc:hive2://hivemr3-hiveserver2-b489d4d7f> load data local inpath '/opt/mr3-run/work-dir/pokemon.csv' INTO table pokemon;
...
No rows affected (0.609 seconds)

0: jdbc:hive2://hivemr3-hiveserver2-b489d4d7f> select avg(HP) from pokemon;
...
+---------------------+
|         _c0         |
+---------------------+
| 144.84882280049567  |
+---------------------+
1 row selected (12.004 seconds)

0: jdbc:hive2://hivemr3-hiveserver2-b489d4d7f> create table pokemon1 as select *, IF(HP>160.0,'strong',IF(HP>140.0,'moderate','weak')) AS power_rate from pokemon;
...
No rows affected (1.246 seconds)

0: jdbc:hive2://hivemr3-hiveserver2-b489d4d7f> select COUNT(name), power_rate from pokemon1 group by power_rate;
...
+------+-------------+
| _c0  | power_rate  |
+------+-------------+
| 108  | moderate    |
| 363  | strong      |
| 336  | weak        |
+------+-------------+
3 rows selected (1.328 seconds)

The user can see that ContainerWorker Pods have been created.

$ kubectl get pods -n hivemr3
NAME                                  READY   STATUS    RESTARTS   AGE
hivemr3-hiveserver2-b489d4d7f-s77b8   1/1     Running   0          6m42s
hivemr3-metastore-0                   1/1     Running   0          6m42s
mr3master-4609-0-dfff6fc7f-g8hbv      1/1     Running   0          6m10s
mr3worker-2590-1                      1/1     Running   0          71s
mr3worker-2590-2                      1/1     Running   0          19s

The user can find the warehouse directory /home/gla/workdir/warehouse/.

$ ls /home/gla/workdir/warehouse
pokemon  pokemon1

Stopping Hive on MR3

In order to terminate Hive on MR3, the user should first delete the DAGAppMaster Pod and then delete Helm chart, not the other way. This is because deleting Helm chart revokes the ServiceAccount object which DAGAppMaster uses to delete ContainerWorker Pods. Hence, if the user deletes Helm chart first, all remaining Pods should be deleted manually.

Delete Deployment for DAGAppMaster which in turn deletes all ContainerWorker Pods automatically.

$ kubectl get deployment -n hivemr3
NAME                  READY   UP-TO-DATE   AVAILABLE   AGE
hivemr3-hiveserver2   1/1     1            1           7m8s
mr3master-4609-0      1/1     1            1           6m36s

$ kubectl -n hivemr3 delete deployment mr3master-4609-0
deployment.extensions "mr3master-4609-0" deleted

Delete Helm chart.

$ helm delete jaundiced-lightningbug
release "jaundiced-lightningbug" deleted

As the last step, the user will find that the following objects belonging to the namespace hivemr3 are still alive:

two ConfigMaps mr3conf-configmap-master and mr3conf-configmap-worker
Service for DAGAppMaster, e.g., service-master-4609-0
Service service-worker

$ kubectl get configmaps -n hivemr3
NAME                       DATA   AGE
mr3conf-configmap-master   1      7m36s
mr3conf-configmap-worker   1      7m32s

$ kubectl get svc -n hivemr3
NAME                    TYPE        CLUSTER-IP    EXTERNAL-IP   PORT(S)           AGE
service-master-4609-0   ClusterIP   10.103.4.82   <none>        80/TCP,9890/TCP   7m50s
service-worker          ClusterIP   None          <none>        <none>            7m47s

These ConfigMaps and Services are not deleted by the command helm delete because they are created not by Helm but by HiveServer2 and DAGAppMaster. Hence the user should delete these ConfigMaps and Services manually.

$ kubectl delete configmap -n hivemr3 mr3conf-configmap-master mr3conf-configmap-worker
configmap "mr3conf-configmap-master" deleted
configmap "mr3conf-configmap-worker" deleted

$ kubectl delete svc -n hivemr3 service-master-4609-0 service-worker
service "service-master-4609-0" deleted
service "service-worker" deleted