Skip to content

Latest commit

 

History

History

zeppelin-spark

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
 
 
 
 
 
 
 
 

Zeppelin-Spark Chart

This is a composite chart deploying the following sub-charts:

Chart Details

Installing the Chart

To install the chart:

$ helm install banzaicloud-stable/zeppelin

Configuration

Helm let's you override all parameters in values.yaml of every chart and sub-chart, however in this chart you don't need anything special you can install this as it is, with default values. If you want to use Spark History Server, you need to enable it and set some cloud storage for event logs file, depending on your cloud provider. The below table lists the required parameters you need to set if you want to enable Spark History Server.

Parameter Required Description Example
zeppelin.username no Admin username, by default is admin
zeppelin.password no Salted password of admin user, by default is zeppelin You can salt your own password using shiro cli tool java -jar ~/dev/tools/shiro-tools-hasher-1.3.2-cli.jar -p
zeppelin.userCredentialSecretName no Credentials above are set in a K8s secret. Instead of specifying username & password directly you can provide the name of this K8s secret containing these fields
zeppelin.sparkEventLogStorage.logDirectory yes the URL to the directory containing application event logs to load yourBucketName/eventLogFoloder
zeppelin.sparkEventLogStorage.cloudProvider yes the cloud provider where the objectstore/bucket located amazon
google
azure
oracle
alibaba
zeppelin.sparkEventLogStorage.secretName no the name of K8s secret containing credentials for selected cloud provider. If no secretName is passed then there will be a secret created with the same structure populated from values. Checkout the required secret properties for each provider in next section below. see below
historyServer.enabled false by default Enable deploying Spark History Server true / false
spark.spark-hs.sparkEventLogStorage.logDirectory yes the URL to the directory containing application event logs to load yourBucketName/eventLogFoloder
spark.spark-hs.sparkEventLogStorage.cloudProvider yes the cloud provider where the objectstore/bucket located amazon
google
azure
oracle
alibaba
spark.spark-hs.sparkEventLogStorage.secretName no the name of K8s secret containing credentials for selected cloud provider. If no secretName is passed then there will be a secret created with the same structure populated from values. Checkout the required secret properties for each provider in next section below. see below

Structure of credential secret for each supported cloud provider

Amazon

AWS_ACCESS_KEY_ID: {{ .Values.sparkEventLogStorage.awsAccessKeyId | b64enc | quote }}
AWS_SECRET_ACCESS_KEY: {{ .Values.sparkEventLogStorage.awsSecretAccessKey | b64enc | quote }}

Azure

storageAccount: {{ .Values.sparkEventLogStorage.azureStorageAccountName | b64enc | quote }}
accessKey: {{ .Values.sparkEventLogStorage.azureStorageAccessKey | b64enc | quote }}

Alibaba

ALIBABA_ACCESS_KEY_ID: {{ .Values.sparkEventLogStorage.aliAccessKeyId | b64enc | quote }}
ALIBABA_ACCESS_KEY_SECRET: {{ .Values.sparkEventLogStorage.aliSecretAccessKey | b64enc | quote }}

Google

google.json: {{ .Values.sparkEventLogStorage.googleJson | quote }}

Oracle

api_key: {{ .Values.sparkEventLogStorage.apiKey | b64enc | quote }}
tenancy_ocid: {{ .Values.sparkEventLogStorage.oracleTenancyId | b64enc | quote }}
user_ocid:  {{ .Values.sparkEventLogStorage.oracleUserId | b64enc | quote }}
api_key_fingerprint:  {{ .Values.sparkEventLogStorage.oracleApiKeyFingerprint | b64enc | quote }}

NOTE: values must be prefixed with subchart name, see the example below how to specify AWS credential properties for zepplin and spark-hs chart sparkEventLogStorage:

zeppelin.sparkEventLogStorage.awsAccessKeyId=XXXXXXXXX
zeppelin.sparkEventLogStorage.awsSecretAccessKey=XXXXXXXXXXXXXXXX
spark.spark-hs.sparkEventLogStorage.awsAccessKeyId=XXXXXXXXX
spark.spark-hs.sparkEventLogStorage.awsSecretAccessKey=XXXXXXXXXXXXXXXX