Zeppelin is a web based notebook for interactive data analytics with Spark, SQL and Scala.
The Hadoop chart can be used to create a YARN cluster where Spark jobs are executed.
- For Hadoop helm installations, visit Hadoop
This chart requires the following charts before install Zeppelin
Create Namespace
$ kubectl create ns hadoop-cls
$ helm install my-release . -f values.yaml -n hadoop-cls
$ helm upgrade --install my-release . -f values.yaml -n hadoop-cls
$ helm delete -n hadoop-cls my-release
The following table lists the configurable parameters of the Hadoop chart and their default values.
| Parameter | Description | Default |
|---|---|---|
zeppelin.image |
Zeppelin image | apache/zeppelin:0.10.1 |
zeppelin.resources |
Resource limits and requests | limits.memory=4096Mi,limits.cpu=2000m |
spark.driverMemory |
Memory used by Spark driver (Java notation) | 1g |
spark.executorMemory |
Memory used by Spark executors (Javanotation) | 1g |
spark.numExecutors |
Number of Spark executors | 2 |
hadoop.useConfigMap |
Use external Hadoop configuration for Spark executors | false |
hadoop.configPath |
Path in the Zeppelin image where the Hadoop config is mounted | /usr/hadoop-3.3.4/etc/hadoop |
ingress.enabled |
Enable ingress | false |
ingress.annotations |
Ingress annotations | {} |
ingress.hosts |
Ingress Hostnames | ["zeppelin.local"] |
ingress.path |
Path within the URL structure | / |
ingress.tls |
Ingress TLS configuration | [] |
nodeSelecor |
Node selector for the Zeppelin deployment | {} |
- Chart can use the hadoop config for the hadoop cluster and use the YARN executor: https://github.com/kubernetes/charts/tree/master/stable/zeppelin